In this follow-up article I am once again looking at market movement between the bookmaker’s opening show and the returned Starting Price, writes Dave Renham. The data I've used has been taken from 2018 to 2022 for all UK flat and all-weather racing. Bookmaker data is taken from William Hill.
In the first piece, which you can read here, the overall stats showed that a bigger percentage of horses drift (lengthen) in price from their opening price than shorten, when comparing market movement from opening show to SP. The biggest difference can be seen in horses that open 10/1 or bigger; of these over 42% drift in price compared with 29% that shorten.
Up to now in this research, I have not considered how much the price changed. Clearly all price movement is not the same; for example, a horse can drift from 4/1 to 9/2, but another 4/1 shot could drift far more dramatically out to 10/1 as an extreme, but perfectly credible, example. Likewise, horses that shorten in price can vary markedly in terms of how much their price contracts.
At this point I want to discuss what I mean by "significant price changes" as we need to be clear that looking at the difference in two prices does not necessarily tell the whole story. To help explain what I mean, let me give you some context. Let us consider two horses:
Horse A – whose opening price is 40/1 and whose SP is 20/1
Horse B – whose opening price is 6/4 and whose SP is Evens
If we focus solely on the prices, Horse A has shortened more because 40/1 to 20/1 is a 20-point truncation, whereas for Horse B moving from 6/4 to Evens is only a 0.5-point move. However, in terms of the chance of winning, Horse B has actually improved its chances more. In order to appreciate this, we need to understand the percentages behind the odds – in other words the percentage chance of winning according to the odds. The table below illustrates this:
As the table shows, Horse A has improved its chances of winning by 2.4% (4.8% minus 2.4%). However, Horse B has improved its chance of winning by 10% (50% minus 40%). Knowing and understanding betting odds in terms of the percentage chance of winning is very important. When I wrote my article on trying to create an odds line, the percentage chance for each horse was something I touched upon. Without it, you cannot easily create an accurate odds line.
At the end of this article, I have produced a table with fractional odds, decimal odds and the implied probability (percentage chance of winning) for betting prices. Readers, if required, may use it as a ready reckoner to convert odds into percentages.
It is now time to compare horses that shorten considerably from opening show to SP with those that lengthen/drift considerably. I am going to only consider horses whose prices have changed, in win percentage terms, by 10% or more. Hence the 6/4 to Even money horses mentioned earlier will count, but not the 40/1 into 20/1 ones. (N.B. The word ‘steamer’ is often used about a horse whose price starts to drop dramatically so for the rest of the article I will use ‘steamer’ for horses that shorten in price).
Let's look at strike rates first – for horses whose win chance changes by 10% or more due to their odds move:
These figures correlate with the findings of the first article and the general data set. Steamers have won 5.5% more often than drifters. However, this does not mean that steamers have produced better returns. In fact, it is quite the contrary as the graph below shows:
These strong steamers actually lost over 13 pence in the £ betting to Betfair SP; drifters lost just 1p in the £. The full breakdown of the results is as follows:
A 10%+ change in win percentage between opening odds and SP only occurs roughly once a day on average. Hence, this type of percentage movement can be considered a major price change.
I thought it might be interesting to breakdown the results further by splitting them by the odds at which they opened. Here are the stats for the steamers:
I have tried to group them so that the number of runs for each group is similar. What seems to be clear is that the worst performing steamers from a profit/returns perspective were those that opened at 11/2 or bigger. The shorter prices of 6/4 or shorter also fared relatively poorly.
Onto driftersnow. There are not many big prices here because a 10%-win probability movement is impossible for horses priced 9/1 or bigger. For an extreme example to illustrate this, a horse drifting from 9/1 (10% chance) to 1000/1 (0.1% chance) has a win probability price movement of only 9.9%:
The results for drifters are more consistent in terms of profits/returns – there is very little difference between the worst performing price band and the best one.
It is clear from this data that if betting at BSP, significant drifters are better value than significant steamers.
Trainers
It is not worth looking at individual trainer data for the strength of steamer/drifter (+ / - 10%), as the sample sizes are going to be too small. Therefore, I am going to focus on trainer data for those horses whose prices changed in win probability terms by 5% or more, rather than 10%. This would include, for example, a horse shortening from 8/1 to 5/1, or one drifting from 6/1 to 10/1. I also restricted the runners a little by looking at horses that opened in price between Evens and 50/1. I ignored odds-on runners and those priced 66/1 or bigger for two reasons: firstly, because the majority of punters do not bet in these price brackets and, secondly, the individual trainer sample sizes are extremely small.
Steamers - As noted above I have broken down trainer performance where the win probability implied from opening odds to SP has improved by 5% or more. Trainers with at least 100 qualifying runners are listed. They are ordered by strike rate/win percentage:
The strike rates are decent as you might expect for steamers, but only a handful of trainers have made a profit. The Johnston and Channon stables have both done well from a profit perspective, as has Mick Appleby. Of course, for steamers, any value is gradually diminishing as the price is contracting.
Ten trainers have produced losses of more than 20 pence in the £, which is quite significant especially considering we are using BSP. The Crisford stable has the worst returns, edging above 30% for their losses.
Of course, if you could predict the steamer before it started shortening in price, and bet before it moves markedly, then the likelihood is that most of the trainers in the list would become profitable. Unfortunately, none of us has a crystal ball, which makes those types of predictions somewhat tricky! What would be interesting to find out is what percentage of horses that initially shorten from opening show, continue to shorten. Likewise, it would be really useful to know what percentage of horses that initially drift from opening show continue to drift. However, I do not have that information and guess the only way to find out would be by doing a live day by day data gathering exercise. That’s for another lifetime!
Drifters – Switching to drifters now below is a table of trainer performance where the win probability change from opening odds to SP has decreased by 5% or more. Trainers with at least 100 qualifying runners are once again listed.
There are six trainers in profit here with Stoute, Haggas and Balding drifters performing particularly well from a returns and strike rate perspective. Indeed, all trainers in the above table have combined to secure a small profit of £21.52.
The two trainer tables I have shared indicate that at this 5%+ level of steam/drift, bettors might be better off backing drifters rather than steamers. In fact, if we look ALL trainers as a whole for both groups, we get the following overall figures:
Steamers, as we have come to expect, have the better strike rate, but drifters are offering punters much better value. So, the question is, do you want more chance of winners, or more chance of a profit?
Going back to the stats for Sir Michael Stoute in the second table above, his non-handicap drifters have totally outperformed his handicap drifters as the chart below shows:
Strike rates are similar but there is a big difference in the profits. Non-handicap drifters of this strength (5%+) would have secured Stoute followers impressive returns of 70 pence in the £ to BSP.
I must admit this has been a very interesting area to research. It is something I have looked at before but not in as much detail. It may be worthwhile comparing early morning odds to SP odds in a future piece, and also at some point I should see if the patterns I have found in these two articles correlate with National Hunt data. Anyway, as ever, I hope you have found the research enlightening, and all comments are appreciated as it helps me with my future work.
- DR
p.s. here is the table I mentioned earlier with fractional odds, decimal odds and the implied probability (percentage chance of winning) for betting prices.
Fractional Odds
Decimal Odds
Implied probability (% chance)
1/100
1.01
99%
1/5
1.2
83.3%
2/9
1.22
81.8%
1/4
1.25
80%
2/7
1.29
77.8%
3/10
1.3
76.9%
1/3
1.33
75%
4/11
1.36
73.3%
2/5
1.4
71.4%
4/9
1.44
69.2%
1/2
1.5
66.7%
8/15
1.53
65.2%
4/7
1.57
63.6%
8/13
1.62
61.9%
4/6
1.67
60%
8/11
1.73
57.9%
4/5
1.8
55.6%
5/6
1.83
54.5%
10/11
1.91
52.4%
Evens
2
50%
21/20
2.05
48.8%
11/10
2.1
47.6%
23/20
2.15
46.5%
6/5
2.2
45.5%
5/4
2.25
44.4%
11/8
2.38
42.1%
7/5
2.4
41.7%
6/4
2.5
40%
8/5
2.6
38.5%
13/8
2.62
38.1%
7/4
2.75
36.4%
9/5
2.8
35.7%
15/8
2.88
34.8%
2/1
3
33.3%
11/5
3.2
31.2%
9/4
3.25
30.8%
12/5
3.4
29.4%
5/2
3.5
28.6%
13/5
3.6
27.8%
11/4
3.75
26.7%
3/1
4
25%
16/5
4.2
23.8%
10/3
4.33
23.1%
7/2
4.5
22.2%
4/1
5
20%
9/2
5.5
18.2%
5/1
6
16.7%
11/2
6.5
15.4%
6/1
7
14.3%
13/2
7.5
13.3%
7/1
8
12.5%
15/2
8.5
11.8%
8/1
9
11.1%
9/1
10
10%
10/1
11
9.1%
11/1
12
8.3%
12/1
13
7.7%
13/1
14
7.1%
14/1
15
6.7%
15/1
16
6.2%
16/1
17
5.9%
18/1
19
5.3%
20/1
21
4.8%
25/1
26
3.8%
33/1
34
2.9%
50/1
51
2%
66/1
67
1.5%
100/1
101
1%
1000/1
1001
0.1%
https://www.geegeez.co.uk/wp-content/uploads/2018/10/DavidProbert_newGeegeezLogo-e1538582067746.jpg313830Dave Renhamhttps://www.geegeez.co.uk/wp-content/uploads/2022/10/geegeez_banner_new_170x78.pngDave Renham2023-10-09 11:03:152023-10-09 11:03:15Steamers and Drifters: Part 2
In this article I am going to look at market movement between the bookmaker’s opening show and the final Starting Price and unearth some truth about steamers and drifters, writes Dave Renham.
In what follows I will be focusing on flat and all-weather racing in the UK spanning five years from 2018 to 2022. Bookmaker data is taken from William Hill.
Typically, the opening show tends to be around ten minutes before the off, and these are the initial prices the bookmakers set. Backing a horse at ‘opening show’ and seeing it shorten in price means you have probably gained a decent edge and potentially some value. Conversely, if you take the opening show price and the horse drifts (lengthens) in price, then you may have lost some value.
However, it is important to note that more horses will lengthen in price than shorten. Here is a graphic to illustrate this by looking at the percentage of all runners that either shorten in price, stay the same, or lengthen in price:
As can be seen nearly 42% of all horses drift, compared with 34% who shorten. Roughly a quarter of all runners see their price remain steady. It is interesting to note that the percentages are very similar when comparing handicaps with non-handicaps (within 1%), hence the market behaves in a very similar way from opening show until the off regardless of race type.
My starting point for researching this article is very simple – look at opening show versus Starting Price and seeing what effect the differential has on strike rate and profit/loss. In terms of profit/loss I am going to calculate returns to Betfair Starting Price. For the sake of simplicity, I am going to split the runners into three groups:
- horses that shorten in price from opening show
- horses that stay the same price as opening show
- horses that lengthen in price from opening show.
Using numerical examples:
Let me start by looking at all races.
As we can see, horses that shorten in price have comfortably the best strike rate, therefore, unsurprisingly, the market does tend to get it right most of the time. In terms of returns to Betfair Starting Price horses that have remained the same price have proved the best value by a couple of pence in the £. It may be interesting to note that there is little in it between horses which shortened in price compared with those which lengthened.
So, we have a very even looking starting point in terms of returns / value, now it is time to dig deeper.
Horses whose opening show price was 4/1 or shorter
I thought it made sense to look at different price brackets so let’s start with the better fancied runners. It also seemed logical to use the opening show price for this rather than SP as the opening show price is known pre-race. Here are the splits:
There is the same sliding scale in terms of strike rate, but it is the horses that lengthened / drifted in price that have been better value this time. Horses that shortened in price proved the worst value.
Horses whose opening show price was 4/1 or shorter and then lengthened in price
Focusing on this subset of drifters, it is interesting when you compare the results on Grade 1 tracks compared with other tracks. The Grade 1 tracks on the flat are Ascot, Doncaster, Epsom, Goodwood, Newbury, Newmarket, Sandown and York. Firstly, let’s review the win and each-way (win + placed) strike rates:
There is a difference of 1.6% in the win strike rates; 4.3% in terms of combining win and placed percentages (each way). These differences may look quite modest, but they are significant.
The graph below shows the return on investment (ROI%) to BSP for each group to highlight the significance:
Now we can see the significance of a 1.6% difference in win strike rates – the returns are over 7p in the £ worse at Grade 1 tracks for these shorter priced runners compared with the non-Grade 1 tracks. Indeed, away from the top tracks we see a situation where one would have virtually broken even backing every single drifter to BSP when it opened at 4/1 or shorter.
Whatever is occurring to create these differences between Grade 1 tracks and non-Grade 1 tracks for drifters, I am not sure. It may be connected with average field size; it may be connected with the quality of racing. It could be a combination of those, or neither.
Sometimes it is not worth speculating, especially as in this case it is nigh on impossible to isolate why. I’m happy on this one that it makes sense to just go with the data.
Before moving on I have checked the 2023 data (up to 27th Sept) and the same pattern for horses that drift/lengthen in price having opened 4/1 or shorter has continued:
The message according to all the information at my disposal is clear: horses which open at 4/1 or shorter and drift look to be POOR value when racing at Grade 1 tracks; away from these top tracks, such horses seem much better value – taking 2023 into account, going back to 2018 these runners would have lost you just 5p for every £100 bet.
Horses whose opening show price was 2/1 or shorter
Going back to data from ALL courses, if we focus on a shorter opening odds criterion of 2/1 or lower, and only look at drifters, we almost get to a break-even scenario. There were 7696 qualifiers of which 2801 won (SR 36.4%). Backing all 7696 runners at £1 stakes to BSP would have lost a meagre £49.13 (ROI -0.6%).
Horses whose opening show price was 2/1 or shorter that lengthened in price
If we once again look at the Grade 1 track data compared with other tracks for this subset of drifters, we see the following:
A similar, if stronger, pattern than with the 4/1 or shorter opening show cohort of drifters. Here, we are looking at a nearly 3% difference in win strike rate which equates to a difference of over 11p in the £ in terms of BSP returns.
As per the table above, drifters at non-Grade 1 tracks opening 2/1 or shorter have edged into profit. For this to happen across such a large sample – over 6500 runners – is interesting and impressive.
Horses whose opening show price was between 9/2 and 9/1
Time to look at the data for a bigger odds bracket. Here are the splits for each subset of the cohort whose opening show price was between 9/2 and 9/1:
It should have been no surprise to see those shortening in price winning more in percentage terms once again. In terms of returns, as with the shorter priced runners, horses that have shortened have been the worst value, albeit by just over 1p in the £.
Horses whose opening show price was between 9/2 and 9/1 that lengthened in price
I wanted to continue the comparison between drifters in this price bracket at Grade 1 and non-Grade 1 tracks to see if we get a similar differential as before. I assumed we would, as my expectations were that it would only start to reverse with longer-priced runners:
As before the non-Grade 1 track data is notably better, both in strike rate terms and BSP returns. However, the gap is starting to narrow, though the difference between the two is still clear. This dynamic has to switch around for bigger-priced runners and we will see whether this is the case shortly.
Horses whose opening show price was 10/1 or bigger
Onto double figure priced runners on the opening show now. Here are the results:
Horses that remained the same price have provided the best returns, while those shortening have marginally out-performed drifters. All groups though show poorer returns than the shorter priced runners we reviewed earlier. Both the shorteners and the drifters offered poor value for the punter.
Horses whose opening show price was 10/1 or bigger that lengthened in price
Finally for this section, I wanted to investigate whether this subset of drifters produced converse results whereby horses that raced at Grade 1 tracks performed better than those which did not.
Here, necessarily, we have the big switch around: horses racing at Grade 1 tracks have the better strike rate for the first time and their returns are much better than their non-Grade 1 track counterparts.
When comparing the results of horses running at a different level of racetrack, splitting up the drifters’ data into opening show price bands has been an eye opener for me. In the future, if my plan is to place my bet late at a Grade 1 track and my horse opens up 9/1 or shorter, I would think twice about backing it if it started to drift. Conversely if I was planning the same at a non-Grade 1 track then I would want to see it drift!
As it stands, the research shared so far has been very general – hence the huge sample sizes. I have not yet considered how much the price has changed, because clearly a horse can drift from 6/1 to 13/2, but another 6/1 shot could drift dramatically out to 10/1 and beyond. In a follow-up article I will be digging deeper into the size of the change in price.
Back to this article and, having looked at the splits in terms of strike rate and returns for different price bands, I thought it would be interesting to go back to where this all started and look at the percentage of all runners that either shorten in price, stay the same or lengthen/drift within each of these bands. Here are my findings:
This is very enlightening as we can see that the percentage of horses that drifted compared to horses that shortened is similar in the 4/1 or less group, and also in the 9/2 to 9/1 group (green/yellow bars). However, in the 10/1+ group, 42% of all runners drifted, compared to a much lower 28.9% of runners who shortened. These stats are implying that we should delay wagering longer-priced runners by either waiting to the last minute or simply using Betfair SP. For horses that open at prices of 9/1 or shorter, the timing of bet placement seems generally less crucial.
Trainers
Changing tack, a quick look at some raw trainer data now. I have chosen 25 high volume flat trainers, and I am simply comparing win strike rates and A/E indices for all of their runners within the three cohorts we’ve used throughout this piece – horses that shortened, horses that stayed the same price and horses that drifted. I have highlighted A/E indices of 0.95 or higher (in green) – these are essentially positive. A/E indices of 0.79 or lower (in red) are essentially negative:
There are more greens than reds and, as a rule, the strike rates increase as you read across the columns left to right. This is what we would expect based on the overall data presented earlier. However, George Boughey is interesting as his three strike rates sit very close together, between 17.25 and 17.72%. Horses of his that have drifted have proved much better value than those that have shortened.
Clive Cox has a poor record with horses that drift in price as does James Tate, perhaps suggesting these yards know when to bet! Meanwhile, David Menuisier has done extremely well with horses that have shortened in price. They would have provided you with returns of over 17p in the £ to BSP which is impressive: another yard to follow when they’re fancied maybe?
I wanted to delve a little more deeply into trainer statistics and analyse the percentage of runners for individual trainers that either shorten, stay the same price or lengthen.
Below is a list of trainers whose runners drift far more than they. I have ordered them by highest percentage of runners that lengthen in price:
It was surprising to see George Boughey in the list and even more surprising to see him at the top. I also had not expected to see Sir Mark Prescott or Gary Moore appear either. It may be that these horses are often put in at defensively short prices on the opening show, bookmakers fearful of shrewd trainers/connections landing a gamble. Elsewhere, some less well-known trainers are arguably more predictable entries in the table.
There are not many trainers where this scenario is reversed with the percentage of horses that shorten in price being higher than the percentage that drift. However, five well known handlers have this profile and are shown below:
There are some big guns in this list. Punters are aware of the skills of these trainers and hence their runners are usually going to be strong in the betting market. It may be that these yards are, generally speaking, less inclined to gamble their horses, though in the case of Aidan O’Brien that’s not typically the case.
*
It is time to wind up this first article into market movement. There are plenty of stats to chew over and hopefully for punters who bet near or around ‘the off’ it has given some useful data to potentially improve your profit/loss bottom line. You can read part two of this article here.
- DR
https://www.geegeez.co.uk/wp-content/uploads/2023/10/steamerdrifter.png320830Dave Renhamhttps://www.geegeez.co.uk/wp-content/uploads/2022/10/geegeez_banner_new_170x78.pngDave Renham2023-10-02 13:57:492023-10-10 09:36:09The TRUTH About Steamers and Drifters
This is the third in a short series of articles connected with betting on horse races in running, writes Dave Renham. In the first piece, I discussed the idea of DOBBING which essentially means ‘double or bust’. You either double your money or lose your stake. Just to recap, here is a worked example:
Imagine you back a horse at 20.0 for £5; in order to create a potential DOB you try and lay the horse at half the odds for double the stake – so you set a lay at 10.0 for £10. If the horse hits 10.0 or lower in running, your lay bet will be matched and regardless of the result you will win £5 (less commission). If the horse loses but doesn’t hit 10.0 or lower then you lose your £5 stake.
In the second piece I looked at some in-play horse racing data on the flat and this time my attention switches to National Hunt racing. One could argue that National Hunt racing is easier to trade as the races are longer which generally allows the trader more time to make informed decisions. But does the data support that contention?
As before I am going to look at 20 months’ worth of recent UK data which is a very decent sample in terms of size. In fact, I started my research by splitting the data into two and looked at the overall NH DOBBING percentages for each group. One came out at 44.4%, the other at 44.3%. Hence, I feel we can be very confident that this data set will provide us with extremely accurate figures.
Dobbing Percentages by NH Race Type
My first port of call is to split the data into three race tyoe brackets: steeplechases, hurdle races and National Hunt Flat races (bumpers). Here are the findings:
There is not much difference between the three, but NH Flat races have offered the best chance of DOBBING, followed by chases and finally hurdle races. Most National Hunt Flat races are around two miles so it should be no surprise to see the DOB percentage at 45.9%. If you remember from the previous article the flat results for races of 17f+ saw the DOBBING percentage standing at 45.2%. Hence the NH Flat figure correlates positively with that.
Dobbing Percentages by Handicaps / Non-handicaps
Onto handicaps versus non-handicaps next and, for this data, I am excluding the NH Flat results (which are all non-handicaps) as we have that figure already. In the graph below I have shown the overall handicap versus non-handicaps splits, as well as then splitting this by chase races and hurdle races:
As can be seen in the chart above, horses are more likely to DOB in handicap races as opposed to non-handicaps. Meanwhile, handicap chases are the most successful DOBBING-wise although handicap hurdlers DOB only 1% below this figure. The lowest figure goes to non-handicap hurdle races where just 41.3% of horses successfully DOB.
Dobbing Percentages by Distance
When we looked at the distance splits on the flat in the second part of this series we saw that, as the distance increased, so did the DOBBING percentages. Here now is the National Hunt racing breakdown:
As we can see, the same pattern occurs here with the longer the race, the better the chance of a horse DOBBING. In the first article I had alluded to the fact that this might be the case. Races of two miles or less give us the lowest overall percentage (41.9%), whereas the longest distances of beyond three miles have seen horses DOB 46.9% of the time. There is a very strong linearity of improved DOB percentage as race distance increases.
It is interesting to note that races beyond three miles have seen a slightly higher percentage of DOBBERS in non-handicaps compared to handicaps, which is a surprise given the handicap/non-handicap stats I shared earlier. However, for the record, non-handicap chases at further than three miles have seen 48% of runners DOB. This is the highest figure based on distance parameters I have found to date.
Dobbing Percentages by Market Odds
Let's now look at the data in terms of Betfair Starting Price (BSP). For the flat data I used market rank rather than price and the flat DOB% stats were quite even although favourites (that were not odds on) had comfortably the highest figure in those findings. I felt it was worth changing it up a bit and using actual market prices, which is arguably a more accurate measure:
As with the market data from the flat there is no discernible pattern here. It does, however, look best to avoid the essential ‘no hopers’ priced over 500/1. However, it is surprising, to me at least, to see both the 50.01 to 100 price bracket and the 100.01 to 500 price bracket both hitting over 47%. I can’t explain that one, I’m afraid!
Dobbing Percentages by Courses
A look at courses now. Here are the DOB%s for each course, ordered highest to lowest:
There is quite a difference between the highest figure, at Newbury, 50.6%, and the lowest, Fakenham, 41.2%. What immediately strikes me is the difference in the configuration of these two tracks. Newbury is one of the biggest in terms of circumference being 1m 7f; Fakenham is at the other end of the scale at just one mile all the way round. Cartmel (one mile circumference) and Plumpton (1m 1f circumference) are other tight/sharp tracks that appear down the bottom of the DOBBING percentage list. Meanwhile near the top you have Ascot and Lingfield whose tracks measure 1m5f round, Cheltenham at 1m4f and Donny at 2m. Kelso which is in 4th spot does buck the trend though being just 1m1f in circumference.
Digging a little bit deeper, there could be something in this theory as I decided to find the average circumference of the top ten DOBBING% courses and compare that with the average circumference of the bottom ten. The top ten courses averaged out at 12.6 furlongs, while the bottom ten averaged out at over two furlongs shorter at 10.4 furlongs. Of course, a theory is simply a theory, but the numbers I have uncovered at least seem offer some support.
I need to add a proviso here that these stats come from only 20 months' worth of data. It is still a decent chunk of data for each course but, ideally, I would like four or five years’ worth.
National Hunt Horses with good past DOB%s
To finish this piece, I have tried to find a handful of horses that have, in the past, had a high DOB%. My hope is therefore that this will be replicated over the coming season. So here goes – there are nine in total and I have listed them in alphabetical order:
Ahoy Senor – Ahoy Senor is one of the top 3 mile chasers in the country. He has DOBBED in 11 of his 15 races, but with one of those races seeing him start odds on, this improves to 11 out of 14 (78.6%). In those remaining 14 races he has won five of them, but in six of the other nine he has still halved in price or better in running. I am guessing one of the key reasons for this DOBBING success is that he is habitual front runner. 14 of his 15 starts have seen him take the early lead.
Ashtown Lad – Ashtown Lad, trained by Dan Skelton, is a versatile runner who last season switched between hurdling and chasing. He has DOBBED 76.5% of the time (13 races out of 17). He has DOBBED in six of his seven chases, while in hurdle races it stands at seven out of ten. He has shown a mix of running styles with equal DOBBING success.
Before Midnight – Before Midnight is a 10-year-old gelding trained by Sam Thomas. His career DOB% stands at 75% with 18 DOBs from the 24 races when he has priced 2.02 or bigger. The slight concern is that his DOB percentage has been nearer the 50% mark when looking at the last 18 months or so, which is mainly down to a drop in form. However, he could now be well handicapped so it will be interesting to see what the 2023-24 season brings.
Brewinupastorm – Trained by Olly Murphy, Brewinupastorm has achieved 14 DOBs from 25 races, but this becomes 14 from 22 (63.6%) when you ignore his odds on runs (remember, odds on runners cannot dob because they cannot halve in price on Betfair). He has raced mainly over hurdles and his DOBBING percentage in hurdle races stands at a very impressive 75% (12 of 16 qualifying runs). He has raced nine times in the last two NH seasons DOBBING six times (66.7%).
French Dynamite – French Dynamite is an 8yo Irish chaser. He has achieved 13 DOBs from his 20 runs (65%), including ten in his last 12 (83.3%). He races up with the pace (led seven times, raced prominently 11 times, and held up twice) which may be a factor. Six wins helps, but overall, this horse seems to have very solid potential for keeping up a good DOB percentage this season.
Gatsby Grey – Gatsby Grey is a 7yo trained in Ireland by Oliver Kiernan with just 14 NH races under his belt. Of those 14 he has DOBBED in nine (64.3%). It is interesting to note, too, that he has never started shorter than 5.0 BSP. He has three wins to his name and his recent DOBBING record (since Nov 2022) stands at five from his last seven (71.4%).
Guy – Guy is an 8-year-old gelding trained by Nigel Twiston-Davies. He has raced in chases but was switched back to hurdles at the end of the last National Hunt season. He has raced 21 times in his career and has DOBBED an impressive 15 times, equating to 71.4% of races. He has only won twice in these 21 starts, which was what initially caught my eye, as that would account for only two of the 15 DOBs. Digging deeper he has finished second nine times of which he DOBBED on eight of those occasions.
Hatcher – Hatcher is another trained by Dan Skelton. His DOBBING percentage for his career stands at 65%. He has won six races when odds on, so that 65% is based on his other 40 NH runs where he has DOBBED 26 times. Of those 40 he has won eight. What makes his overall record more impressive is that most of his racing has come at around two miles. As we know from the distance data shared earlier, this trip produces the lowest average of all the NH distance DOB%s. However, there is one caveat: as with Before Midnight his more recent DOB% record stands around the 50% mark so this does need to be taken into account.
Le Tueur – Le Tueur is an 8-year-old gelding who has been racing over fences since November 2021. He has DOBBED 14 times in 23 races (60.8%), but again is not a serial winner with just three wins. It is interesting to see that in two of the three races where he was pulled up, he still DOBBED!
So it’s time to wind up this third piece on DOBBING. I am in the process of starting to crunch some run style data for NH races, but it takes a long time - several weeks in fact. Once done, I will report back on that I’m sure in the future.
- DR
https://www.geegeez.co.uk/wp-content/uploads/2023/09/AhoySenor.png319830Dave Renhamhttps://www.geegeez.co.uk/wp-content/uploads/2022/10/geegeez_banner_new_170x78.pngDave Renham2023-09-25 16:48:232023-09-25 16:50:31Introduction to Dobbing: Part 3
This is the second in a short series of articles connected with betting on horse races in running, writes Dave Renham. In the first piece, which you can read here, I discussed the idea of DOBBING, which essentially means ‘double or bust’: you either double your money or lose your stake. Just to recap, here is a worked example:
Imagine you back a horse at 6.0 for £10; in order to create a potential DOB you try and lay the horse at half the odds for double the stake – so you set a lay at 3.0 (half of 6.0) for £20 (double £10). If the horse hits 3.0 or lower in running, your lay bet will be matched and regardless of the result you will win £10 (less commission). If required, there is a little more detail in the first article.
This second article will dig into the numbers in an attempt to see whether we can improve our chances of finding DOBBERS. I will look at 20 months’ worth of recent UK flat data, which equates to over 12,000 races, covering over 100,000 runners. So let’s get cracking!
Dobbing Percentages by Distance
My first port of call is the distance of the race. In terms of dobbing percentage there is a clear pattern when it comes to distance:
As we can see, the longer the race, the better the chance a horse has of DOBBING. In the first article I had alluded to the fact that this might be the case, and it is always good to see the numbers support the theory. The minimum distance of 5 furlongs gives us the lowest overall percentage (38.8%), whereas the longer distances of 13 to 16 furlongs have seen horses DOB over 44% of the time, while 17f+ races hit 45.2%.
I can think of three logical reasons why there is such a discrepancy when we compare 5f races to races of 17f or more.
Firstly, 5f sprints only take around a minute; races of 17f or more take three or four times as long. These longer races give more time for traders to spot horses that are making eye-catching progress / travelling well.
Secondly, we know that 5f events really disadvantage horses that take up a position near, or at, the rear of the field early. Hence a good proportion of slow starters / held up runners in sprints are not going to get close enough to the action at the front in time. So the chances of these horses DOBBING is relatively low. In contrast, horses that start slowly or are held up in 17f+ races have plenty of time to recover and get into a more competitive position.
And thirdly, keeping on the run style theme, front runners win around 2.5 times as often over 5f as they do over 17f+. It is easy to imagine that, in races where horses that make most or all the running over 5f, very few other runners get in a dangerous enough position to shorten markedly in price and hence DOB.
In fact, having checked that last theory, and although I only looked at 30 5f races where the winner led from start to finish, only 25% of runners dobbed in these races. That’s well below the 38.8% overall percentage for 5f races. I would not expect that 25% DOBBING figure to change too much even if I back-checked 300 races rather than 30. Unfortunately it is not something I can research quickly, therefore the modest sample.
Dobbing Percentages by Market Rank
Moving on to the position the horse holds in the betting market, and as a reminder, odds on runners have been ignored in the figures as they cannot DOB (see first article for full explanation).
Favourites DOB the most; close to 45% of the time – this is probably because a good proportion of market leaders win and, of those who don’t, most run well. I am slightly surprised to see the other market ranks relatively uniform and not sliding downwards left to right. I thought that would be the case, but there is no clear cut pattern.
Dobbing Percentages by Course
In the first article I mused on whether course configurations can make a difference to DOB percentages. Camera angles are different at certain tracks, for example, and as we know courses in this country vary massively in terms of layout. If we look at Chester’s tight bullring track…
…we can see the course is roughly a mile in circumference with short straights. Compare this now to York’s expansive gallopers’ paradise:
The circumference of the track at York is roughly double that of Chester and the finishing straight is nearly five furlongs in length.
Every track in the UK is unique – some are undulating, some have downhill stretches, uphill stretches, long or short straights, sharp bends, etc. Therefore, I would expect to see some variance across the different courses in terms of DOB%. To begin with, let’s look at the DOBBING percentages for each course.
There is a 7% differential in terms of percentages of runners successfully dobbing between Sandown at the top (46.3%) and Newcastle at the bottom (39.2%). It is interesting to note that three of the four lowest DOBBING courses are all-weather ones (Chelmsford, Wolves and Newcastle). Indeed, the other three all-weather courses also reside in the bottom half of the table. Is this down to the level of competition on the all-weather being slightly below that of the turf? Possibly.
Now we know from earlier findings that the distance of races makes a difference DOBBING wise, so maybe courses that have a lower percentage of sprint races enjoy better DOBBING percentages. Likewise do courses with a higher percentage of sprint races have poorer DOBBING percentages?
To try and test out this theory, I ordered all courses in the previous table from 1 to 37 starting with Newcastle who had the lowest DOB%. Hence I put Newcastle in position 1, Wolves in position 2 and so on up to Sandown in position 37. I wanted to use these ‘positions’ to help make the comparison.
I then calculated the percentage of sprint races (6f or shorter) held at each course during the same time frame, ordering the courses from 1 to 37, thus:
Bath top the list with over 50% of all their races being sprints, while Epsom has the lowest figure with just 15.3% of their races being at trips no further than 6f.
Having given the tracks a rank in terms of percentage of races at the course that were sprints, I could compare this with their DOB% rank. For the course DOB% to be strongly affected by race distance then the course ranks for the two variables should be similar.
In some cases they were – Thirsk for example was in position 3 in both: the North Yorkshire track had the third lowest DOB% matching perfectly with the third highest percentage of races that were sprints. In other cases, though, they didn’t match. Wolverhampton, for instance, was in position 2 for DOB%, but position 25 for percentage of sprint races. To try and show the comparison for all the courses I have created line graphs comparing their ranks. I have split it into two so that it fits on the screen:
For perfect correlation we would need to see the blue and orange lines almost follow the same path. That has not happened here taking all the courses as a whole, so we need to look to see how many courses have their orange and blue dots close to each other. 15 of the 37 courses have their two ranks varying by five or less. Meanwhile, nine of the 37 courses have their two ranks varying by 15 or more.
Thus the jury is still out in terms of saying that the course DOB percentages are affected by distance considerations. My guess is that it is a factor at some courses, but there are other factors also making a difference.
Dobbing Percentages by Run Style
To conclude this second article I want to look at possibly my most favoured area of analysis: run style. It should be noted for the run style research for this piece, I have not been able to use such a big data set, due to the time-consuming nature of this type of research. However, I have been able to analyse 4000 runners looking at how run style impacts the chances of DOBBING.
I mentioned in the first article that horses that lead for the majority of the race, or are leading at the furlong pole while looking like a potential winner, are occasions when the leader’s price is likely to shorten considerably. Obviously, if the price drops enough then the horse will DOB. Hence it would logically follow that front runners should have the highest DOB percentage. This is indeed the case as the chart below clearly illustrates. I have used the run style categories on geegeez.co.uk, and the following stats are pulled from all flat race distances from 5f to 2m 6f:
Clearly run style is important from a DOBBING perspective. Front runners DOB over 60% of the time across the test sample, and the chart clearly shows the downward trend from front of the field early to back of the field early. 4000 runners across all distances should be a big enough sample for these figures to be accurate. If I was able to look at 100,000 runners, I would be surprised if the percentages for each group changed by more than two or three percentage points. Moreover, I personally researched run style DOBBING percentages back around 2011/2012 and the percentage splits then correlate well with this newer sample.
It should be noted, however, that the distance of the race will cause slight changes to the run style DOB figures. 5-6f races will see the DOB% for ‘Led’ increase slightly to around the 64-65% mark; conversely the DOB% for ‘Held Up’ drops to under the 30% mark. In longer races of 1m4f or more the reverse happens, with the DOB% for ‘Led’ dropping to 55-56% while the ‘Held Up’ DOB% increases to 38-39%. This makes perfect sense as front runners win such a high percentage of sprint races compared to longer races and, as we know, winners will DOB except for that very small proportion that are priced under 2.02.
-----
That’s the end of this review of flat race DOBBING. The Run Style figures should give readers who may be thinking about employing a DOBBING strategy a possible way in. Next time, I’ll perform a similar analysis of National Hunt racing. Until then…
DR
https://www.geegeez.co.uk/wp-content/uploads/2019/05/telecaster_toodarnhot_Dante_830x320.jpg320830Dave Renhamhttps://www.geegeez.co.uk/wp-content/uploads/2022/10/geegeez_banner_new_170x78.pngDave Renham2023-09-19 13:06:462023-09-19 13:06:46Introduction to Dobbing: Part 2
DOBBING is a word I came across around ten years ago in connection with in play/in running betting, writes Dave Renham. DOBBING is usually shortened to ‘DOB’ which means ‘double or bust’. Essentially it is an in play trading strategy. If the trade/DOB is successful, we double our original stake, if the trade/DOB is unsuccessful we ‘bust’ or lose our stake.
What is DOBBING?
For people who have not heard of DOBBING before I will give you a worked example which hopefully will help:
Let us imagine you back a horse at 10.0 for £10; in order to create a potential DOB you try and lay the horse at half the odds for double the stake – so you set a lay at 5.0 for £20. If the horse hits 5.0 or lower in running, your lay bet will be matched and regardless of the result you will win £10 (less commission).
Here are the basic mathematics behind the two potential winning outcomes:
- If the horse goes onto win the race, you get £90 returned from the ‘back’ part of the bet while you lose £80 on the ‘lay’ part of the bet. This gives you a £10 profit.
- If the horse does not go onto win but reaches 5.0 or lower in running, then you lose your £10 stake from the ‘back’ bet, but gain £20 from the lay part of wager – again giving a £10 profit.
- If the lay part of the bet is not matched with a horse that does not win the race, you lose your original £10 stake.
The table below is another way to look at it, showing the three possible outcomes:
For dobbing to be profitable long term, we are probably looking to have a success rate of around 54% or more. This figure has to be a bit higher than a 50.1% baseline as we need to take commission into account.
As Russell Clarke mentioned in the first of his excellent in running articles (which you can read here), only about 20% of all money traded on a horse race occurs in running. Hence, there are far fewer people that trade in running compared with those that don’t. I am sure there are plenty of you out there who have thought about betting in running, but have decided against as it is not for you. There will be some of you who trade and are very successful. Personally, I do dabble in running from time to time, but despite using the market leading trading software, I know I am up against seasoned trading pros. I might be able to produce pre-race plans that are as good as most, but my decision making / speed under pressure is definitely not at the expert level.
An advantage of using this dobbing idea for some punters/traders is that you can place both parts of the bet/trade before the race starts. Therefore there is no need for trading software and you do not have to make quick decisions in running because you have made them already. Hence if you are like me, this is potentially a big plus. However, as the saying goes ‘for every positive there is a negative’. I find I say this phrase regularly in my life away from racing.
I tutor maths and chess online, and my chess students regularly hear this positive/negative quote pertaining to certain moves they make. It is the same here: let us imagine you set your ‘DOB’ pre-race and leave it to run its course. What happens if say you back it at 10.0 and set the lay at 5.0, but by the time the race starts the horse has drifted 15.5? OK the horse might still hit 5.0 in running, but this is going to happen far less often than it would if the starting price was 10.0. Of course the horse could shorten before the off as well, but as a general rule more horses drift than shorten. I know this because I have written about this before, and I have double checked recent data too. As an example, if we look at opening show prices compared to SP in 2023 (UK flat racing), we get the following figures:
Some pretty strong evidence to back up the general rule I mentioned above.
Of course there are ways round this potential issue by placing your back bet as late as possible; literally as the last horse loads in the stalls. The later you place it, the closer the price will be to its eventual Betfair Starting Price. That will mean however, that you will have to calculate and place the lay part of the bet immediately afterwards, and if you have literally placed your bet at the last second, you will be setting your lay after the race has got underway. Having said that, you should be able to put the lay price and stake in the Betfair machine before the horses have reached the end of the first furlong. This manual approach, though, clearly requires you to be around at the start of every race.
An alternative to avoid either the price fluctuating or needing to be tied to your trading screen at the start of each race, is to use some trading software. It is not too complicated to automate the software to back a horse at Betfair SP and once the Betfair SP is established, a lay at half those odds will be automatically placed. The lay will be calculated immediately the Betfair SP has been established (a few seconds after the off) to create the potential ‘DOB’.
DOB Examples
It is time to look at some races to see what can happen to Betfair prices in running. How many horses tend to DOB in a race, how many do not? Initially let me look at four races run on the same day (August 29th 2023). They are all 10 runner events – I chose those races simply to make the ‘dobbing calculation’ easy to understand.
Race 1 – 2.15 Ripon 1m2f handicap (4yo+)
The result is shown below with the Betfair SPs (BSP) and the lowest price matched in running (IP LOW). The penultimate column (BSP/IPL) is the result of dividing the BSP by the IP LOW. For a successful DOB the BSP needs to have at least halved in price; hence showing a figure of 2 or more. Successful DOBBERS are highlighted in green:
In this example, despite nine of the ten horses shortening in price, only two (Bollin Margaret and Cedar Rapids) DOBBED. There was one near miss (Tele Red 1.90). If you watch the race back, or even just look at the in running comments, you will probably understand why there were so few DOBBERS. The early leader, King Titan, led for less than two furlongs and hence was never going to shorten in price enough leading for less than a fifth of the race, especially when the lead was a narrow one.
Cedar Rapids took up the running after 2f leading for the next six furlongs and, considering his starting price of 83.06 and that he was still leading 2f from home, it is fairly easy to appreciate why he shortened to 22 and hence DOBBED. Bollin Margaret then took over the lead having just passed the 2f pole and led to the finish. Hence, as a winner at a BSP of 13.36, Bollin Margaret was always going to DOB.
In addition, once Bollin Margaret took over, the nearest challengers never really looked like getting to her. We could have found that out by watching the race replay or by looking at the in running comments. The comments for Bollin Margaret were ‘took keen hold, prominent, switched right over 2f out, ridden to lead over 1f out, kept on well final 110yds, always doing enough’. Hence with none of her closest pursuers really looking like winning this helps explains why they did not DOB.
Obviously each race is different and the number of horses that DOB will not be the same proportion of runners in each race (as we will see).
IMPORTANT NOTE: Before moving onto the second race, it should be noted that race winners do not always DOB, because the BSP has to be at least 2.02 for a horse to halve in decimal price (to 1.01, the lowest value on Betfair). Hence odds on winners cannot DOB.
Race 2 – 4.00 Ripon 6f handicap (3yo+)
Onto a sprint handicap a bit later on the same card:
This time, three of the ten runners DOBBED, one more than the first race. In this race three horses led at various points: the winner Twelfth Knight, as well as Abate and Russco. Twelfth Knight and Russco both DOBBED, while Abate was a very near miss with a BSP to IP LOW ratio of 1.98. Horses that lead at some point in the race are usually going to shorten in price, sometimes considerably so. That is the type of pattern I would generally expect to see, and the first two races have conformed to that pattern.
Race 3 – 4.15 Newbury 1m 4f handicap (3yo only)
Over to Newbury now for a handicap over 12 furlongs:
This time we see over half the field (six of the ten runners) DOBBING, despite only two horses leading during the race. The high number of DOBBERS is almost certainly due to the fact that the winner Graham, who had been clear 4f out, started to experience that lead steadily eroding. Hence, many in play traders observing the pack close on the leader would have thought / hoped / expected that one or more of those challengers would potentially win. This almost certainly explains why two horses traded so low; Medieval Gold (2nd) traded at 1.5, and Gordon Grey (4th) traded at 2.06.
Race 4 – 6.30 Musselburgh 5f handicap (3yo+)
Here are the facts and figures for this Musselburgh sprint:
An even split here with five horses DOBBING and five not. Three different horses led, of which two DOBBED (Sixcor, the winner, and the runner up, The Grey Lass). Two of the other three that DOBBED come as no real surprise if you watch the race back. Beneficiary made good headway mid race and as he was a big price, he would have caught the eye of enough traders to see his price contract sufficiently to DOB. Favourite Aconcagua Mountain travelled strongly and a furlong out looked the most likely winner. He faded in the final furlong but not before trading as low as 1.56.
What these four races tell us is that every race is going to be different from a DOBBING perspective. Just like every race is different if you are simply backing a horse or indeed laying one.
DOB Anomalies
Readers should note, that there are occasions when only the winner DOBS. An example was the two-mile Goodwood Cup this year run on 1st August. In this race, Quickthorn was well clear of the field after four furlongs, and a mile later with just half a mile to go he was still 20 lengths clear. The opposition assumed that Tom Marquand, the jockey of Quickthorn, had gone out too quick but they were sadly mistaken, and no other horse really stood a chance. Knowing how the race panned out explains why it was no surprise that no horse was really that close to halving in price in running.
Here is the result with the accompanying in running data:
This scenario of a single DOBBER in a race will occur from time to time especially in very one-sided events. However, it is extremely unusual for every horse in a race to DOB. In previous research from 2018 I had a dim recollection that there was a race at Nottingham where all the runners DOBBED. After doing some digging I found the race in question. It was the 7.25 on 7th August 2018. It was a 10f handicap with six runners. Here is the result:
Not only did all six DOB, but they all DOBBED fairly comfortably. One reason for this may be that four of the six led the race at some point, while the two who didn’t, were close to the lead making headway at different points.
DOB Success Rates
Going back to the four 10-runner races I shared earlier, of the 40 runners, 16 DOBBED. This equates to 40% of the runners. Previous flat racing DOBBING research I have done, over different time frames, showed the overall DOBBING percentage average out at around this 40% mark. Considering we need a success rate of roughly 14% higher than this, there is clearly a job on to make DOBBING profitable. So, how can we improve upon this base figure of 40%? Here are some thoughts / questions, which I will aim to expand upon in subsequent articles:
How does run style affect DOB success? I have already discussed the fact that horses that take the lead at some point in a race are likely to shorten in price. Horses that lead for the majority of the race, or are leading at the furlong pole while looking like a potential winner, are both occasions when the leader’s price is likely to shorten considerably. Obviously if the price drops enough then the horse will DOB. Front runners, especially in sprints, are horses that are likely to have real potential to DOB; and, of course, the longer they lead the more chance of this happening. Hold up horses are not screaming out DOBBERS, unless they get into a much more forward position at the business end of the race.
You often hear commentators say that a horse is ‘travelling well’. Horses that are ‘on the bridle’ tend to shorten in price as they are not under any pressure, or so it seems. This is a potentially time consuming idea to test, but I want to put it out there.
Does race distance make a difference? Longer races mean greater elapsed time, and logic dictates that there will be more price fluctuation as a result of this. Hence, the chance of DOBBING may increase.
Do courses make a difference?Certain courses, Bath for example, have difficult camera angles in the final couple of furlongs. I remember trying to trade in running in a Bath sprint around ten years ago; never again! It was so difficult to monitor all the horses from a front on angle. Other course considerations I guess that may have a positive bearing on DOBBING percentages, such as those with uphill finishes, or with long home straights.
What effect does the price of a horse have? One sensible argument would be that shorter priced runners may DOB more often than longer prices, simply because the market suggests they will be more competitive: they are more likely to be mounting a serious challenge at some point in the race. An alternative argument would be that horses starting a long odds do not have to go ‘low’ in running to DOB. An 80.0 BSP shot only needs to reach 40.0 to DOB. Whereas a 4.0 (3/1) shot needs to hit 2.0 (even money).
My questions and thoughts don’t stop there, but it is time to wind up this introduction to DOBBING. As you can see we have a fair bit of digging and number crunching to do – or at least I do! And, at this stage I have only really discussed flat racing; I have not even mentioned National Hunt racing as yet...
- DR
https://www.geegeez.co.uk/wp-content/uploads/2019/08/Coronet_PrixJeanRomanet_830x320.jpg320830Dave Renhamhttps://www.geegeez.co.uk/wp-content/uploads/2022/10/geegeez_banner_new_170x78.pngDave Renham2023-09-12 08:39:102023-09-12 08:40:34An Introduction to ‘Dobbing’
In recent articles I have looked at a very simple ratings method for all-age handicap races which, on initial testing, seems to have shown more positives than negatives, writes Dave Renham. I hope and expect to write further about these ratings at a later date, but need more time to do some further detailed research. This will take several weeks, probably a couple of months.
In this somewhat related article, I would like to share with you the process I went through when trying to create ratings for two-year-old (2yo) races. My plan was to stick to a similar methodology which in essence was:
a) find what I thought were key factors/variables;
b) use PRB (Percentage of Rivals Beaten) data once more as my metric;
c) combine the PRB figures in the same way as the all-age handicap ratings by simply adding up the relevant scores.
There are a number of different types of 2yo races such as maidens, novice events, Group/Listed races (which are all non-handicaps) and nurseries (handicaps). My idea was to try to rate the maiden and novice races. To me these are quite similar types of race and hence I hoped that one cap could be worn by both. Of course that would not necessarily be the case, but even if the ratings worked well for one of the two then I would have achieved something.
To begin with, let me discuss factors I considered for use. Here was my ‘longlist’:
Trainer record – in 2yo maidens/novices
Sire stats – in 2yo maidens/novices
Debut course
Horse Sex – colt, gelding or filly
Horse purchase price
Most Recent form – Last time out (LTO) finishing position
Recent market data – LTO price
Fitness – days since last race
Draw
The eagle eyed among regular readers will note that the last four factors are ones I used in my original ratings for all-age handicaps.
From this starting point I felt I needed to trim the list down, for two reasons. Firstly, as I mentioned in my very first ratings article, when creating ratings I prefer not to over complicate things. Secondly, some of the above factors would cause some problems for one reason or another.
The draw was the first to be discarded. In all of the articles I have written on the draw in the past, I have mentioned that draw bias works best in handicap races. Hence, although the draw may affect some 2yo races at certain courses, I felt it was not a reliable enough factor to use here. Next to go was purchase price as I had no easy way to source it, or indeed back check it on past results. Further, many horses are home bred and therefore never go through a sales ring. I felt it had importance, which is why it made the longlist, and I wished I had some data I could ‘crunch’ to see how important it actually was, but I felt it was a no go for these ratings.
Fitness using the days since last run metric was the third factor I decided to discard. My main reasoning here is that the advantage of a quick return, that tends to happen in older age handicaps, is not replicated for 2yo runners. I looked briefly at some win and placed stats which were very even across the days ranges, so I felt it was unlikely that the more accurate PRB figures would really give a wide enough spread of figures. I felt it wasn’t worth the hours of data gathering and sorting if the figures were likely to be almost identical across the board. One makes decisions like this all the time when delving into horse racing research. Of course sometimes we make incorrect ones but, with experience, decision making improves.
That left me with six factors/variables so let’s look at each in a little more depth.
1 Trainer record – I am not someone who bets often in 2yo races. Occasionally I will if I spot what I feel is a good betting opportunity. However, my main bets that involve 2yos occur when I play the Tote Placepot. Most meetings have at least one 2yo race in their first six so I have to use some methodology to choose which juvenile runners I am going to put into my ‘pot’. Trainer information is always my first call.
Many trainers do follow a similar path year in, year out; they generally stick to the same training methods, know which races to target, etc. Now it should be noted (albeit it is fairly obvious) that each year trainers have a completely new ‘string’ of 2yos, so variances in overall performance are going to happen from year to year. However, when we think about the bigger stables they tend to keep many of the same owners, and these are likely to be purchasing similar animals to what they have done in the past. Hence, past trainer 2yo data is usually quite a good guide to future performance. The graph below offers a real life illustration through the record of Charlie Appleby in 2yo maidens/novice races over the past four full years:
These figures are very similar from season to season and, as I am writing this, his current stats for 2023 are in the same ballpark – 31% win strike rate and 55% each way strike rate.
So how best to utilise past 2yo trainer data was my main consideration as there were different stats I could potentially use. One option would be to use PRB figures calculated from all 2yo maiden and novice events for each individual trainer. However, my concern with that was that the number of runs that a 2yo has is usually extremely important. This is a graph I shared in a previous article written in April when examining 2yos on their second starts:
As we can see there is a significant difference in 2yo performance on debut compared to second starts. Such differences would be replicated when comparing the relevant PRB figures. Not only that, this graph is taking all 2yo runners into account and as you can imagine some specific trainers have even more acute differences. For example, and once again using data from 2017 to 2022, Michael Dods had a 2yo debut win SR% of 5.3%, whereas his second starters won over 16% of the time. William Haggas 2yo debutants scored less than 12% of the time, but on second start won 27% of their races. These are just two examples showing one potential pitfall of using overall 2yo trainer data to produce a trainer rating score.
It was at this point in retrospect when the alarm bells should have been ringing, about how complicated just creating the trainer part of these ratings would be. However, I thought that using previous runs would almost certainly be the way I would want to go, and trainer stats would make the final ‘cut’. However, before digging any further I wanted to look at the other five factors.
2 Sire Stats – sire stats are often an important part of the 2yo betting picture due to the limited past run data most juveniles have. In some cases, especially early season, all the runners in a 2yo race will be making their debut. Hence we have no past form to go on, so we have to look elsewhere. Sires are the fathers of the respective horses and can have a significant influence on their offspring. When we dig deeper we find that the offspring of a good proportion of sires have clear traits or preferences. These may be going/ground related, distance related, age related, experience related, etc.
Having essentially decided to use previous starts as a key factor in determining the trainer rating PRB score, it would be difficult to do the same for sire stats, as this would potentially overlap somewhat. It is not as bad as using LTO position and Beaten distance LTO as two factors in a system as they are virtually the same metric. However, the improvement from debut to second start for sires would mirror trainer improvement to some extent.
Therefore for sire stats I felt a distance metric made more sense: splitting the 2yo sire PRB data into two, obtaining figures for sprints (5-6f) and for longer 2yo races (7f or more). The majority of 2yo races are contested at a mile or less so this seemed logical to me. To give an example of a sire whose 2yo distance stats differ across these two distance ranges, let me share the non-handicap 2yo win stats for Kingman. In 5-6f races his strike rate has been 12.8%, at 7f or longer this increases to 22.2%.
Interestingly, though, when I calculated PRB figures for Kingman they were closer than I had expected. His progeny’s 2yo PRB for 7f+ was 0.64 compared with 0.60 for 5-6f. This comparison helps to highlight why I believe PRB figures are the most accurate of all the statistical metrics that compare performance. Win stats are a good barometer, but PRB figures are much better because they effective ‘grade’ each run; not just whether the horse won or didn’t, or placed or didn’t.
No Nay Never is another sire whose 2yo offspring show a distance bias. At sprint distances his 2yo non-handicappers score over 19% of the time, at 7f or longer this drops to 13.6%. The PRB figures for No Nay Never this time do underline the strength of this bias as the sprint figure stands at 0.63, while the 7f+ one is much lower at 0.53.
Sire stats using this distance metric looked a good option to use in the ratings.
3 Debut course – this is something I have researched in the past and the track at which a horse makes its debut can be a factor in how it subsequently performs. It particularly affects the second career start as we can see if we compare the PRB figures for second starting 2yo that made their debut at either Ascot, Newmarket, Redcar or Ripon.
The importance of the debut course becomes less of a factor the more runs a 2yo has, but it still can have a bearing, so I would have to separate out the number of runs since debut in some way or other. Alarm bells were ringing this time as this factor is definitely going to be time consuming from a data gathering aspect, as I would need to collect the LTO course data one at a time and then combine number of last runs to each course. That could mean anything between 100 and 200 separate data ‘dumps’ into excel as well as adding extra columns and data to it. Ouch. However, at this point I was undeterred, as there have been times in the past when I have had to perform an enormous amount of data collection to write an article or series of articles. Also, I felt this factor was really important and would improve the ratings if it was included.
Having slightly buried my head in the sand regarding the enormity of this project, the question I now considered was does factoring in debut course combined with past career runs conflict / overlap with the trainer data idea which was going to use past runs too? I guessed it would to a small extent, but I was open-minded enough not to dismiss using this metric because of that slight concern. Clearly trainers have their preferred starting points for 2yos in terms of races and courses for debut runs, but individual course debut data combines all trainers and hence any significant overlap is extremely unclear. I was fairly confident – hopeful at least! – that the two factors would not conflict enough to make the ratings biased in any way.
Before moving on, I started to think about another problem that I had known would be a real issue in terms of 2yo ratings. What to do if the horse was making its debut? They have no past race data to work with; no debut course stats. What PRB rating could be assigned to those runners? I had several things to ponder, but decided to move onto the next factor as I felt it would at least have fewer issues.
4 Horse Sex – the sex of a horse has relevance and in 2yo races there are essentially three types of runners – colts (entire males), geldings (males who have been gelded) and fillies (females). I did some initial number crunching as this data collection was easy to do and not time consuming. I compared their PRB figures based on about 25,000 2yo runs in maidens and novices. Here are the findings:
As we can see colts have the best record, followed by fillies and finally geldings. The majority of 2yo runners are colts and fillies (around 87% of all runners combined) leaving geldings that make up a much smaller 13% of the runners.
These stats look promising from a ratings perspective, and I had some data collection completed!
Onto the last two factors now, both of which I used last time.
5 Most Recent form – LTO finishing position is a good barometer of most recent form and it seemed to work well in the handicap ratings. However, I would have same issue with the course debut stats with horses making their debut. What PRB figure would I use?
6 LTO price – LTO price also seemed to work well with the handicap ratings but again the question was what to do about debutants?
*
At this point I was feeling happy that potentially I had six factors to combine to create the ratings. On the flip side, there were a myriad of issues. Perhaps the biggest was the problem of 2yos that were making their debut. These runners would not have PRB figures for three of the six factors (LTO course, LTO position, LTO price). I needed to consider the options.
Option 1 – To use just one of the three LTO factors giving debut runners a standard PRB figure based on all debut run performances.
Option 2 - Combining the three LTO factors giving debut runners a standard PRB figure based on all debut run performances, and dividing the score by three. This would mean all three factors had some relevance (in essence 1/3 of a rating factor).
Option 3 – Use the ratings only on 2yo races where all the runners had previously run at least once.
Of the three I felt the last option made the most sense as I really wanted to combine all six factors if I could. Based on a look at race data going back to 2019, 33% of 2yo races involved horses that all had run at least once previously. This would still provide around 350 races a year where the ratings could be employed. Added to that I had the facility to pull out all these races.
Having decided that was the preferred way forward, thoughts turned to the enormity of the data collection. As a researcher one is limited by the amount of data one has, or can access. We are also limited to a great extent by our computer skills. If you are able to write and use sophisticated computer programs for example, this gives you a huge advantage over those who cannot. If you have a vast database of results with every single type of variable/factor you can think of you also have a big advantage. Time is such a precious commodity and, without either of the above, my constant issue was the hours required for complicated or detailed research.
My expertise in terms of data number crunching is purely Microsoft Excel-based. I am proficient using Excel and use certain time-saving tricks such as cell formulae, pivot tables, functions like ‘VLOOK up’, and so on. However, I cannot write VBA code for macros, which impinges greatly on what I am able to do in terms of quantity and within certain time frames.
Back to the problem in hand. It was time to look at each factor again and try to work out how much work / hours would be involved with each one.
Trainer record – the advantage I have from a research perspective in terms of trainer data is that when I export thousands of results, the trainer of each horse is part of the data set. Hence as a rule trainer data collection/manipulation is not as time consuming as many others things. On the negative side I would be looking at probably three or four separate data sets which I would need to combine and sort. Once that was done I could create the necessary formulae to calculate individual PRB figures and once those are added for all runners, I could use a pivot table to help calculate each trainer’s individual PRB figure. At least I didn’t have to worry about getting the debut stats; that would save me a little time.
The ideal plan would be to have PRB trainer figures for horses that have raced once, raced twice, and then group those who have raced three or more times together.
That part of the research was not too daunting; definitely doable. It would take several hours probably, but not several days!
Sire stats – when I started thinking about how ‘easy’ it would be pulling and then crunching the sire PRB data for the two distance ranges, I suddenly realised that a trick I often use with sire data collection would not work for PRB figures. I could pull sire data relatively quickly if I was using win strike rates or each way strike rates. BUT not for PRB figures. It suddenly dawned on me that I would have to go one at a time, sire by sire. If that wasn’t bad enough from a time perspective, I also realised that even once I’d done that I’d need to find a way of ‘marrying’ the sire data with the trainer data. That would be even more time consuming and rather fiddly to do.
I thought then, OK I could ditch the sire stats part. I’ll still have five factors to use. The other ratings worked well with five, and even with four when I rated races without the draw factor.
Debut course – back to this potentially tricky factor. I no longer needed to worry about debutants and what figure I would assign to them. However, as I mentioned earlier, I would still need to collect the LTO course data one course at a time combined with the number of career runs the horse had. As with the trainer data collection plan the aim would be to have ‘course on debut’ PRB figures for horses that had raced once, twice, and three or more. Earlier I had reckoned that I would need to collect separate data around 100 to 200 times and marry it together; it was clear that this was going to be within that range, although at the lower end (roughly 110).
-------
It was at this point that, if I had a towel nearby, I would have thrown it in! I had already reached the moment where the data collection and subsequent number crunching was too much to comprehend and hence attempt. It would take several weeks – far too many hours of my time for what I was endeavouring to do. Not only that, I still had three other rating factors where I would need to gather data. That being said, data collection for those three factors would all be far less onerous than the first three. However, it would still be several hours’ worth to add on top.
I was at a crossroads: I needed to decide whether I totally shelved my idea, or adapted it in some way. It has already been established that logically I cannot back test the data over several hundred races as I’d like to, due to the vast amount of time it would take. However, an alternative would be to look to rate races one by one, in real time as it were. Find races for the remainder of the season that qualify and then number crunch each individual race. To be able to do that though, I would still need to have sourced and collated the trainer data from the last few seasons (probably going back to 2015 or thereabouts).
In addition, I would need to source and calculate the PRB figures for LTO position and LTO price. I cannot use the PRB figures I used in the all-age handicap ratings because I used past all-age handicaps to calculate them. To collate the LTO position and LTO price PRB figures for 2yos would not take too long. Again, hours rather than days. On a more positive note the horse sex figures I had already calculated so that rating factor is no problem.
Then, for the sire stats (which I could incorporate doing it this way) and the debut course stats, I would need to check each horse in the race, crunching and then collating the relevant figures. That would take some time, and rating one race would potentially take up to 20 minutes if there was a big field of runners. On the plus side, once I had calculated the individual sire PRB figure that could be added to my 2yo ratings database.
The same would apply for the course on debut/number of career runs PRB figures. Once one was calculated that, too, could be added to the database. After rating, say, 20 to 30 races, the sire PRB stats and the debut course PRB stats would be starting to build up. That would make rating subsequent races far easier as I would start to have some data to hand for some horses that I didn’t need to recalculate.
Hence this is a potential way forward for these ratings should I choose to go that route in future. It will still be a very slow process, and because of that I am undecided in terms of what to do. What is most likely to happen is that I will start to collate some stats over the coming weeks, then try and rate five or six races, and go from there. If the first few races offer some positive signs, it will be easier to plough on and look at more races. If they don’t then it possibly is back to the drawing board.
-----
I hope this article has highlighted the fact that not all horse racing research goes smoothly.
It also shows that, despite all the best intentions, some ideas, no matter how good they may turn out to be, are simply too time-consuming or difficult to research. What has happened to me here is not a one off. In the past I have started researching numerous ideas with the plan of writing about them, only to abort the process at some point. So I’m used to the disappointment!
That was going to be the end of the article, but before checking it through I decided to source and collate the trainer data. As I have now done that I feel it is only fair to share the data with you. If nothing else, you now have some 2yo trainer PRB figures that may prove useful.
Below is a table of 2yo maiden/novice PRB figures for a selection of trainers. I have chosen the 30 trainers who have saddled the most 2yo runners. The figures are grouped as I discussed earlier into horses that have run once previously, horses that have had two career starts and then horses who have run three or more times:
As you can see most trainers have similar figures in the first two columns, with the third column being the best. My next job will be to source and calculate the PRB trainer figures for horses making their debut. However, that will need to wait for another time.
So I will finish here and ponder what next as far as my attempt to produce ratings for 2yo maiden/novice races is concerned. There will be an update in the future, I promise!
DR
https://www.geegeez.co.uk/wp-content/uploads/2023/08/LTO_PRB.png320830Dave Renhamhttps://www.geegeez.co.uk/wp-content/uploads/2022/10/geegeez_banner_new_170x78.pngDave Renham2023-09-04 08:53:522023-09-04 08:53:52An Attempt at Creating 2yo Ratings
This is the third article connected with my attempt at creating simple ratings for certain horse races, writes Dave Renham. The first piece looked at how you could create a simple ratings method; the second tested this simple idea using some historical results. As the results were quite promising, I thought I would extend the number crunching to more past races, and in this third article I will report back my new findings.
To give some context, I was looking for a method for rating handicap races and, as far as weighting was concerned, I decided to give each factor a similar one. In order to do this I used the PRB (Percentage of Rivals Beaten) metric.
The rating method I came up with involved five factors or variables – these were:
Draw – splitting the draw into thirds;
Most Recent form – for this factor I used last time out (LTO) finishing position;
Recent Market data – LTO price was used for this one – so the Industry Starting Price the horse was returned in its most recent race;
Long term form – for long term form I used career placed percentages in handicap races.
Fitness –I used days since last run to create the PRB figures for this final variable.
For more ‘meat on the bones’, the first article explains in depth what the individual PRB values were within each of the above categories. Here's the link.
Essentially each horse ends up with five PRB values and therefore when rating a race, we simply add up the five PRB figures for each horse giving them a final total or rating score. The horse with the highest total becomes top rated, the second highest total becomes second rated and so on.
In the second article I did some back testing on the ratings, although for this stat-busting exercise I decided to ignore the draw factor by focusing on longer distance races. ‘Time’ was the main reason for ignoring the draw – it was something that was going to take far too long to collate the necessary information. Hence my ratings were ‘trimmed’ – now they would be created by using the four other PRB figures produced from LTO position, LTO price, career handicap placed percentages and days since last run.
For this next batch of testing, I kept the first three rules mentioned above, but changed the runners rule to 10 to 12 runner races. Before sharing the results of this second phase of testing, let me quickly share my thoughts on what I perceive to be the most important finding. For any ratings to have the ‘potential’ to be useful, they need to show strong similarities with the actual betting market. The first phase of testing did see this happening. For example, the top rated runner started favourite in over 41% of the races rated. Just 3.7% of top rated runners started 7th, 8th or 9th in the betting. Ultimately if your ratings do not mirror the betting market that well, then the chances are they are going to be dud.
Of course, as punters, we are looking for value, and the hope is also that the ratings throw up value selections. The first set of results shared in the second article offered some promise in that regard.
OK, it is time to look in detail at the 10 to 12 runner 2018 results. For the record this comprised of 362 races in total.
Firstly, let's review how the top rated horses matched up against the actual betting market.
This graph perfectly shows the type of sliding scale one wishes to see. It is similar to the one we saw when analysing the 8 to 9 runner race data.
Over 35% of the top rated runners started as the favourite, while more than 72% of the top rated runners started in the top three in the betting. Compare this to 8th or bigger in the betting which accounted for just 5.1% of all top rated runners.
Let's now look at the second top rated runners next in the same way:
Again, this graph gives the type of results that suggest the ratings are fairly accurate in terms of assessing potential chance of winning. We would expect the highest bars in the graph to be on the left hand side once more with a sliding scale going from left to right. 57.2% of the 2nd rated runners ended up in the top three of the betting compared with 2.3% ending up in the bottom three of the betting (10th or worse).
So once again the top section of the ratings are looking good. Time to take a look at the lowest rated to see how they fit against the market. Firstly let us look at the market rank percentages for the lowest rated runners. To begin with let's review the 12th rated runner in 12 runner handicap races:
This graph is effectively a mirror image of the first two we saw, which is exactly what I would hope to see. Over 66% of the bottom rated runners ended up 10th or lower in the betting.
Now a look at the 11 runner races (bottom rated):
An even better set of figures here in terms of correlation. This is probably due to the fact that the 11 runner sample size was around 33% bigger than for 12 runner races.
I don’t see the need to show the whole graph for the 10 runner races as well, but the results were similar once more. The bottom rated runner appeared 8th, 9th or 10th in the betting market over 63% of time, while only 12% of them ended up in the top three market positions.
It is very pleasing to see that the results we got for 8 to 9 runner races are being replicated here. Essentially these simple PRB based ratings are looking like providing a relatively sound framework in terms of forming our own market – if nothing else. I discussed some ideas about how to form a betting tissue/market in this article which preceded this series. These ratings could be used in conjunction with that – or even be used in a stand-alone manner.
Time to see how the ratings fared in terms of winning – their win strike rate. In the 8 to 9 runner results, the higher rated runners comfortably out-performed the lower rated ones. Obviously, I'm hoping for the same scenario here:
These results are reasonably positive – the top rated runner has done extremely well and we do get the type of sliding scale one would hope for. In truth, the 8 to 9 runner data looked stronger, but when you analyse win and placed data, the picture looks more clear-cut:
This graph gives excellent correlation with higher rated runners hitting higher win and placed strike rates; lower rated runners doing the reverse.
It is time now to look at the performance of the individual rating positions in terms of profit/loss to BSP. At this point it should be noted that in the whole data set for this article, there have been some unusually big-priced winners. The ten highest BSP prices of winners during this study were:
Within the type of sample size used for this piece, huge priced winners are a common problem when trying to use BSP as a value metric. That is why in other Geegeez articles, where appropriate, I have quoted BSP on shorter priced runners only. Unfortunately using this shorter priced idea will not work effectively on the rating positions data due to very small sample sizes (for the lower rated positions in particular). Therefore, in this case, the BSP profit/loss figures for individual ratings positions shown below may confuse matters for some readers, but hopefully you'll still get the gist at least:
We do have to take the profit/loss figures here with a pinch of salt however; especially the lower rated ones. Ordinarily strike rates of 4.1% and 5.2% are not going to produce stunning returns of 78.6 pence in the £ or 57.2 pence in the £.
Thus, instead of dwelling on these skewed figures, it makes more sense to dig deeper into the top-rated runner results as these prices as a whole are much shorter. In fact 88% of all top rated runners were priced 12.0 BSP shorter.
I want to look at two main areas when it comes to top-rated runners. Firstly I want to delve into profit / returns, so here are the profit/loss stats for top-rated runners in terms of their market position.
It is interesting to see that the top-rated runner has made a decent looking profit when actually starting as the favourite. From a ratings value perspective though, I would have liked to have seen slightly better figures from the poorer market positions. Having said that, 362 races is too small a sample to see a potential pattern emerge such as that, especially when just 63 top rated runners started 5th or higher in the actual betting market. It is heartening to know that top rated runners that were 5th or bigger in the betting did make a profit of £11.15 from these 63 runners, but we need much more data. Not just on top rated runners but on other rating positions too.
Secondly, I wanted a breakdown of how far clear of the 2nd rated the top rated was. This is something that I omitted to think about when penning the last article. Hence for this next table I have combined the relevant stats from both articles to include all 8 to 12 runner handicap results. This gives us a bigger data set for analysing the gap between the top two rated runners. Here are the findings:
Before commenting on this, it must be stressed that despite expanding the sample size, it is still a relatively modest one. However, one could not have dreamed of a much better set of results (well, I suppose I could have, but you have to have some sense of realism!) The bigger the gap, the better the results – both from a strike rate and returns perspective.
My next port of call was looking at ratings position versus betting market position. I wanted to compare the performance of horses that are rated better than their odds position, compared to those who are not. Here are the results:
As with the 8 to 9 runner data, horses ranked better than their price ranking have done best from a profit/loss perspective. There is a big differential here, but as I have already stated, the BSP data for all these races is not too reliable, and hence I would not read too much into this.
Before finishing, I have one more set of figures I want to share. As I did in the previous piece, I want to look at the actual rating scores and group the lower rated runners as a whole, and compare them with the higher rated runners. Tthe lowest possible rating using my PRB scores is 1.64; the highest possible is 2.39. The groupings I have used are horses that were rated 1.64 to 1.84, and horses that were rated 2.18 to 2.39. These are exactly the same groupings I used in the 8 to 9 runner article. Here is the comparison of wins, runs and strike rate for our two groups in 10 to 12 runner handicaps:
These stats are what one would have expected based on all the previous data shared in this article. However, it is always nice to have expectation validated in black and white.
I have not added the BSP profit figures as the 1.64 to 1.84 group had two of those huge priced winners I alluded to before (353.78 and 137.3). Such winners totally skew the profit/loss column making a comparison a mockery (as we have seen twice before in this piece, with the profit/loss figures for individual ratings positions, and with the rating rank v market rank BSP comparison). For the record, the 2.18 to 2.39 group, which did not have big priced winners skewing the results, lost a modest 20 points to BSP equating to an ROI of -4.7%.
So that’s currently where I’m at. There has definitely been further promise in this latest piece of research. I will decide where I go next with this over the next few days and any new ratings research will be written up and shared with you in the very near future.
- DR
https://www.geegeez.co.uk/wp-content/uploads/2023/08/ratings1.png320830Dave Renhamhttps://www.geegeez.co.uk/wp-content/uploads/2022/10/geegeez_banner_new_170x78.pngDave Renham2023-08-29 08:54:352023-08-29 08:58:20My Simple Ratings Method, Part 3
In the last piece I wrote on Geegeez I attempted to demonstrate to readers how you could go about creating a rating method in order to help you to analyse a race, writes Dave Renham. This article continues on from that simple ratings method, as I have decided to dig into the past and do some testing using historical results.
To recap, I was looking for a method for rating handicap races and, as far as weighting was concerned, I decided to give each factor a similar emphasis. In order to do this I used the PRB (Percentage of Rivals Beaten) metric.
The rating method I came up with involved five factors, as follows:
Draw – Using the Draw Analyser tool rather the draw tab in the racecard in order to define a more precise date range, I set what I felt were relevant parameters. These parameters had to ideally a) match the race in question; and b) give me a big enough sample size.
My ideal date range is a recent one such as 2016 to 2023, for handicaps only and, in terms of runners, covering a spread of plus or minus two runners compared with the field size of the race I was rating. So, for example, if it was a 10-runner race, I would set 8 to 12 runners on the Draw Analyser. The Draw Analyser gives PRB figures for individual stalls as well as grouping them into thirds. I used the thirds method for the ratings, grouping low draws together, middle draws together and high draws together.
Most Recent form – for this factor I used last time out (LTO) finishing position. For the relevant PRB figures I looked at two full years of handicap race data (2021 and 2022) to give what should be incredibly accurate readings. This amounted to several thousand races. The PRB figures had a range from 0.60 for winners last time out to 0.41 for horses that finished 9th or worse LTO.
Recent Market data – for this factor I used LTO price – so the Industry Starting Price the horse returned in its most recent race. Once again I used 2021 and 2022 handicap races to create these PRB figures. The PRB figures had a range from 0.60 for horses priced 6/4 or shorter LTO to 0.36 for horses that priced 40/1 or bigger LTO.
Long term form – for long term form I used career placed percentages in handicap races. Again the data for the PRB figures was taken from the two years of 2021-22 handicap results. The PRB had a range from 0.58 for career placed percentages of 51% or more, down to 0.44 for those who hit 20% or less.
Fitness – for this final factor I used days since last run to create the PRB figures. To give the most accurate scores I used the same data set (’21-’22 handicap races) as I had done for the previous factors. The PRB figures ranged from 0.61 for horses that returned to the track within three days to 0.43 for horses off the track for 71 or more days.
So, essentially when rating each race, I took the relevant five PRB figures for each horse and added them up. The horse with the highest total became top rated, the second highest total became second rated and so on.
After writing the original article my plan was to rate a few races and see how the figures worked out. This is something I am still in the process of doing and will feed back my findings in a future article.
[I initially had no intention of back testing results because I thought it would take far too long. However, using a bit of excel, a fair amount of copying / pasting, and a few shortcuts I thought of as I was going along, I managed to get a year’s worth of ratings data in a few hours. The only ‘problem’ is that to do this I had to ignore the draw factor. The main reason for this was that it would take me far too long to gather the draw data (probably several months). But there were other reasons as well, one being that a good proportion of course and distances do not have a significant draw bias so trawling through masses of these types of race would not really improve the ratings or make them more accurate.]
So my ratings would be created using the four other PRB figures based on LTO position, LTO price, career handicap placed percentages and days since last run.
When you create ratings or systems and then back test them on past results, it is important to ensure that you use a different data set. This is a common mistake people make – one I made the first time I tried to create systems back in the early 90s. Hence, having used a data set of 2021 to 2022 to create all the PRB figures, I needed to choose a different year for the testing phase. I chose 2018.
A year of handicaps gives me plenty of data to work with. I did however want to narrow that down by looking only at 3yo+ and 4yo+ handicaps, as this would avoid handicaps with younger, less exposed runners. My assumption is that these ratings will work far better in races that involve older horses. I also chose to try and eliminate any draw factors by choosing handicaps races of 1 mile 1 furlong or more. Without the draw in the ratings, it made no sense to test shorter distance handicaps where draw bias can be extremely relevant and potential skew some ratings results (without the draw PRB being considered). Finally, I looked at 8- or 9-runner 3yo+/4yo+ handicap races for the basis of this article.
Just to reiterate I am back testing my ratings on:
Year - 2018 (UK racing)
Age group - 3yo+ / 4yo+ handicaps
Distance - 1m1f or longer
Runners – 8 or 9
Before sharing the results of my testing, let me discuss briefly what I am hoping to find. For the ratings to have the potential to be useful/effective, more often than not, they need to show strong correlation with the actual betting market. If your top two rated horses are consistently near the head of the actual betting market this is a far more positive sign than if they are consistently near the foot of the betting market. Of course in terms of making a profit from your ratings, you are looking for them to be more accurate than the actual betting market and throw up value selections. Not easy!
OK, let’s dig into my findings:
Firstly let's see how the top rated horses matched up against the actual betting market.
This graph is extremely positive with over 41% of the top rated runners starting as the favourite. Indeed 79% of the top rated runners started in the top three in the betting. There is a definite sliding scale, too, showing the type of correlation you would be hoping for. Let me look at the second top rated runners next in the same way:
Again, this graph gives positive results. You would expect the higher bars in the graph to be on the left hand side once more, and they are. Just over 72% of the 2nd rated runners ended up in the top four of the betting.
So the top section of the ratings are looking good. How about the lowest rated? Firstly let us look at the market rank percentages for the lowest rated runners. To begin with let me look at the 8th rated runner in 8-runner handicap races:
The graph is reversed compared with the first two we saw, which is exactly what we are looking for. 44% of the lowest rated runners were at the bottom of the betting market in 8th place. Less than 17% of them ended up in the top four of the betting. Further positive news as far as the ratings are concerned.
Now a look at the 9-runner races (bottom rated):
A similar lay out to the 8-runner races with nearly 72% of 9th rated runners ending up 7th, 8th or 9th in the actual betting market.
I have to say that I am extremely pleased with the correlation to date between my ratings and the betting market. For something that is relatively simple (just four parameters), it is mirroring the betting market well.
So these ratings, on the evidence we have so far (based on 324 races), definitely show some potential. Time to see how the ratings fared in terms of winning – their win strike rate. Clearly I was hoping that the top rated runners would comfortably out-perform lower rated ones. Here are the findings:
More positive news with the top two rated runners both securing strike rates in excess of 20%. Also a clear break between the top four rated and those rated fifth to ninth. The 7th rated is very slightly out of kilter, but this can happen – the important fact is the trend from top rated to bottom rated is downwards.
What I now want to look at is how the ratings would have performed if betting on them. I am assuming that we are backing at £1 level stakes to Betfair Starting Price (BSP). Here are the findings:
The results for the top rated runner are a little disappointing, losses of around 17p in the £. Horses ranked 7th have made surprisingly high profits, but most of the big priced BSP winners happened to pop up in this specific ranking position. I doubt these figures would be replicated again – this is just the type of outlier you can get when analysing BSP profit/loss.
When taking the top four rated as a whole, they have outperformed horses rated fifth to ninth as the table below shows:
Considering how big priced runners on Betfair can skew the figures, these grouped results are very heartening.
I must admit I am pleasantly surprised with these initial findings. I am intrigued to see how the ratings work with shorter distance races where I can include the fifth parameter – draw bias.
My next port of call was looking at ratings position versus betting market position. I wanted to compare the performance of horses that are rated better than their odds position, compared to those who are not.
Just to clarify, some examples of horses that are rated better than their odds position would be as follows (I appreciate for many I am just stating the obvious, but just in case there is any confusion in my English/grammar):
And here are examples where they are not (these include identical positions in the rank of the ratings compared to the actual market rank):
My hope is that I see better returns for horses that are rated better than their odds position, compared to those who are not. This would suggest that the ratings can potentially pinpoint some value selections.
Here are the returns for each:
These figures suggest the rankings are doing a pretty good job – it seems there has been more value when the rating rank has been better than the market position.
Before winding this piece up, I have one more set of data to share with you. I am looking at the actual rating scores and grouping the lower rated runners as a whole, and comparing them with the higher rated runners. Now the lowest possible rating using my PRB scores is 1.64; the highest possible is 2.39. The groupings I have used are horses that were rated 1.64 to 1.84, and horses that were rated 2.18 to 2.39. These groupings from 2018 3yo+/4yo+ handicaps would have produced the following results:
The strike rates should come as no surprise based on the evidence of the ‘Ratings Win SR%’ graph shown earlier, but the differences in returns are even wider than I had expected. It is another indication that these simple ratings have some real potential.
I'm to park things here for now and start further number crunching for the follow up article. The data set of 324 races is a decent one, but before making too many bold claims, I think we need to look to how these ratings fare in other races. Research wise, I plan to analyse the 2018 data from 10- to 12-runner races next. Once that’s done, I will write it up and share my findings.
In my previous article I went through some basic ideas in terms of trying to create your own odds line or betting tissue, writes Dave Renham. In this piece I am going to show you how to go about trying to create a rating method in order to help you when analysing a race. There is no perfect way to rate a race; there are no perfect ratings, so this idea / method I am sharing is just one of thousands of potential ways to rate a horse race.
Two problems in the past I have found with rating a race have been firstly which factors to use, and secondly what ‘weighting’ or importance do I give to each one. Let’s look at factors first:
Factors / variables to use – to start with, one important thing to be aware of is to make sure the factors you ‘rate’ do not overlap in any way. A simple example of this would be using ‘last time out (LTO) finishing position’ but using ‘LTO beaten distance’ as well. These two factors are very similar as they are both measuring last time out performance and they should not be used in combination in terms of rating races; rather, choose one or the other.
For me I do not want to over complicate things so I would be looking for a handful of factors/variables to use in my ratings. Here are the factors I tend to concentrate on when trying to develop a rating system, and what ‘measure’ I would use:
Most Recent form – either LTO finishing position or LTO distance beaten
Recent market data – LTO price or prices from last 2-3 runs
Long term form – some stat connected with the horses’ overall career
Fitness – days since their last race
Draw – past C&D draw stats split into thirds
Weighting of factors/variables – this is tricky in my opinion, and I have no magic bullet to share with you I’m afraid. What I have struggled with in the past is which stats to use for each factor – win strike rate, placed strike rate, A/E indices, etc. Not only that, but how on earth do you ‘weight’, for example, LTO position versus days since last run? How do the individual LTO finishing positions compare with a grouping of days since the horse ran last? What grouping for days since last run do I use? I cannot use individual days, so do I group it in weeks, blocks of 10 days, etc?
There are lots of questions, but no clear cut answers. Suffice to say, you just have to go with your gut instinct in terms of weighting factors. Once you have rated a few races, you will get a feel for what you may have to adjust to improve them.
For this article I will be using the five variables mentioned above in an attempt to create simple ratings for horse races. As far as weighting is concerned, I am basically going to weight each factor in a similar way. In order to do this, my stat of choice is going to be the PRB stat (Percentage of Rivals Beaten).
Percentage of rivals beaten (PRB) – Before becoming a member of the Geegeez family I had not really delved into this metric much. However, now, I think it is arguably the most important racing stat I consider. For more information on PRB (and the other metrics used on geegeez.co.uk) check out this article.
On Geegeez you can find the PRB stat in a variety of areas which can be accessed from the racecard – individual horse records such as:
On the Profiler tab:
In the pace / run style tab:
And in the draw tab:
Hence, we can find the draw PRB stats needed for my simple rating method on Geegeez. For the remaining stats we need to make use of several hours of number crunching I did prior to writing this article.
How is this simple rating system going to work?
Essentially, I am going to use the five factors mentioned earlier and find the relevant PRB figures for each horse within each factor. Then I am going to add up the five PRB scores to give me their final scores or rating. I said it was simple! I would suggest trying this idea in handicap races; I would use a different idea for say 2yo races or 3yo maiden races.
Right, let’s go through each factor one by one:
Draw – I would like to start with one of the tabs you can use on Geegeez. Let us imagine we have a 1m handicap at Pontefract with 9 runners. I would actually go to the Draw Analyser tool rather the draw tab in the racecard in order to use a more precise date range. Hence this is what I would enter in terms of parameters:
As you can see, I have chosen a recent date range (2016 to 2023); handicaps only due to it being a handicap, 7 to 11 runners (+ or -2 from 9 runners), full ‘going’ range from hard to heavy, and ‘Actual’ rather than ‘Card’ as this takes non-runners into account.
In this imaginary 9-runner handicap example, any horse drawn 1 to 3 would get a ‘0.60’ PRB ‘rating’ figure, those drawn 4 to 6 would get ‘0.48’ and draws 7 to 9 the figure would be ‘0.42’.
Most Recent form – for this factor I am going to use LTO finishing position. For the PRB figures I have looked at thousands of handicap races to give the most accurate readings. Here are the PRB figures for LTO finishing position:
As you can see LTO winners have a PRB figure of 0.60 when running in their next race; compare this with horses that finished 9th or worse LTO whose figure is much lower, not surprisingly, at just 0.41.
Therefore, when rating each horse you simply need to look for their LTO finishing position and assign the relevant figure from this graph.
Recent Market data – for this factor I am going to use LTO price – so the price the horse was in its most recent race. Here are the PRB figures, again taken from thousands of races:
Another sliding scale here as you would expect with shorter priced runners LTO producing higher PRB figures. Hence a horse that was priced 11/2 LTO would be assigned a rating figure of 0.55, if they were priced 25/1 LTO their figure would be 0.44, etc.
Long term form – for long term form I am going to use career placed percentages in handicap races. The data shared again is taken from thousands of races to give us the most accurate figures possible. I have split the percentages into four groups as the graph below shows:
The beauty of this stat from a Geegeez perspective is that you can find these percentages by using the ‘Instant Expert’ tab from the racecard. All you need to do is to adjust the distance tab (top left of screenshot) to include all races (I use the full range from 5f to 4m4f), click the handicap tab (top right of screenshot), and for ALL flat races click the ‘Flat & AW’.
In the above example, the horse at the top has a career placed percentage of 43% and as this lies between the ’36 to 50%’ grouping, this horse would be worth a PRB figure of 0.54.
Fitness – for this metric I am using days since last run (DSLR) with once more thousands of races analysed to create accurate PRB figures. Here are the splits:
As you can see this metric is going to be quite even for most horses, as the vast majority of runners will have been off the track for between 8 and 70 days. The 8 to 14, 15 to 28, 29 to 42 and 43 to 70 groupings are very close together in terms of PRB figures.
And that’s it for configuring my factors. Hopefully it has been fairly self-explanatory.
However, let me give you a fictitious example which hopefully will help. I will stick to the 9-runner mile handicap race at Pontefract that I used for the draw data earlier.
Firstly here are our imaginary horses and their relevant stats:
From here we can convert these into the relevant PRB figures for each individual stat:
I have totalled up the five PRB figures for each horse to give them a final total (furthest column on the right). These totals are their final rating figures and I have ordered our imaginary horses highest to lowest below:
From these ratings, you would hope there is good chance that the winner will come from one of the top three (Fireball, Frazzle and Dobbin); likewise you would hope the bottom three rated (Monty, The Closer and The Gooner) are unlikely to produce the winner.
How you deploy your ratings is going to be different for each individual. You may want to use them as a basis for an odds line – in a 9-runner race, each horse theoretically has an 8/1 chance of winning so you could initially price up the middle rated horse at 8/1. This horse is Plodder – from here you could assign the rest of the prices using Plodder as your starting point, and then adjust them to get a book percentage of close to 100%. Once done you could compare them to the actual prices on offer to see if there are any horses that potentially offer you value.
An alternative approach is to simply compare the actual market position with your rating positions. Let’s say for argument’s sake Dobbin, your third rated horse, is priced up at 10/1 and is 6th in the betting, you might perceive this to offer value. Likewise if the top rated horse Fireball is third in the betting priced 5/1, again you might perceive this horse to offer you some value.
Essentially how you interpret the ratings is going to be personal to you – again there is no right or wrong way. What ultimately matters is how predictive your ratings are. I have not back tested this rating idea as yet, but it is on my ‘to do list’, as I have enjoyed researching and writing this piece. What is likely with a first ratings attempt is that I will need to make some adjustments – for this idea it may be that I am underrating one particular factor and overrating another. If that proves to be the case, I could apply some multipliers to the relevant PRB figures to help with that adjustment. For example, I may need to multiply the underrated PRB figures by 1.10 and the overrated ones by 0.90 to give more accurate overall ratings. However, that is for another time.
Until next time, I’d encourage you to experiment for yourself and if you find anything interesting, feel free to share in the comments below.
DR
https://www.geegeez.co.uk/wp-content/uploads/2023/08/LTO_PRB.png320830Dave Renhamhttps://www.geegeez.co.uk/wp-content/uploads/2022/10/geegeez_banner_new_170x78.pngDave Renham2023-08-07 12:04:222023-08-07 12:04:22How To Create Simple Horse Racing Ratings: Example
In this article I am going to discuss some ideas when it comes to creating your own odds line (betting odds / tissue), writes Dave Renham. As punters we need to take the odds of horses into account when contemplating a bet so 'finding value' is clearly an area of real importance. I wonder, though, what percentage of punters actually produce their own odds lines? I would guess less than 1%.
The modern world seems all about trying to do things quickly, and most punters haven’t got the time or inclination to spend hours and hours creating an odds line or betting tissue for races. Although there is no magic bullet for creating a line, in this article I will look at three basic DIY methods for your consideration, none of which take too much time to produce.
Why create an odds line?
Before delving into some ideas, let me first talk a little about why you may want to create an odds line and why an accurate one is extremely useful.
Making money from horse racing is not about the ratio of winners to losers you get, it is about whether you are able to achieve value prices. If you can achieve value prices on most of your selections you WILL come out in front. I’ve lost count of the number of times friends who have gone to a race meeting, usually on a ‘jolly’, have asked me to give them as many winners as possible. My reply is always the same: ‘if you want the most number of winners, just bet the favourite’. I tell them that favourites are the most likely horse to win so bet them. Of course this is true – they have the best chance of winning in percentage terms. However, I do also tell them that this is not the right way to approach betting in reality and that I am simply answering their question with a percentage-based answer.
So let's explore what ‘value’ is in more detail. Sticking with favourites for the moment, let us imagine over a long period of time we back a hundred identical favourites who all theoretically have the same chance of winning, and are all priced at Even money. According to their betting odds, each horse has a 50% chance of winning their respective race. For you to achieve ‘value’ the true chance of each horse winning needs to in excess of 50%. In reality is more likely to be around 45%.
Let us now imagine that each horse has a true percentage chance of winning of 55%. Over the 100 races, you will win 55 and lose 45, and make a profit of 10p for every £1 bet.
It would be interesting to know how many punters have a good understanding of probability in terms of betting odds. I am lucky that my main profession was as a Maths teacher and hence probability is something I understand well.
I must admit, I am often amused when I hear about certain ‘significant’ market movers. For example you may hear a pundit say, ‘Horse A has halved in priced from 33/1 to 16/1’, which I think gives people the wrong impression. OK, if these two prices reflect the true chance of Horse A winning, then the probability of success has improved from 2.9% to 5.9%. Whilst it is technically correct that the horse has halved in price, the percentages have changed by just 3%, which is not as big a swing as the numbers 33 and 16 perhaps suggest.
Let us imagine another horse, Horse B, whose price has come in from 9/4 to 13/8. On first glance, for some, I would guess that this shortening in price does not appear anywhere near as big as the 33/1 to 16/1 one. However, this second example is actually a much bigger percentage swing, moving from a 30.8% chance of winning (9/4) to 38.1% (13/8) - a difference of 7.3%. So understanding the percentages when it comes to betting prices is important.
When to use a 'tissue'
Before I look at some different ideas regarding creating tissues, or odds lines, I want to suggest what type of races you might start with. I think it is best to look at races where most of the runners are 'exposed' – in other words, they have plenty of form in the book. I would probably avoid Group races and low grade handicaps, however, and look for something in between – maybe class 2, 3 or 4 handicaps. Also I would personally focus on races with smallish fields for the first few races – around 7 to 9 runners seems a sensible option - in order to manage the time aspect of creating a tissue. Ok, let’s look at method 1.
Method 1. The ABCD method
This is a very simplistic idea but is a potential starting point if trying this for the first time. I would grade each runner depending on what I felt in terms of their general chance of winning, where:
A – very strong chance
B – good chance
C – worth considering
D – unlikely, but cannot dismiss
E – poor chance
F – virtually no chance
How you assess each horse to arrive at your ranking is totally up to you. I personally would look at a combination of factors using some, or all, of the following:
i) form, both recent and long term
ii) fitness
iii) trainer/stable form – recent and course form
iv) the draw in flat races
v) potential run style if known
vi) starting prices of most recent starts
vii) speed ratings
Of course this is not a definitive list, there are other factors you may wish to consider; essentially it is up to you.
Having graded the horses I would then try and assign a rough percentage chance of them winning. Let us imagine a 7 runner race where we have the following grades:
From here I would add up all the percentages and here they total 103%. Ideally I would like them to add up to 100% to give me a ‘perfect’ book, but 103% is perfectly acceptable. If your percentages are further away from 100%, massage the 'percentage chance' figures until they fit.
From here I would assign the closest betting price to each percentage. Some percentages such as 25% have an exact betting price of 3/1; others, like 15%, have not, but the closest price to that is 11/2 (which equates to 15.4% chance). Here is my grid with the prices added and the percentage chance rounded:
The overall percentage book has edged up slightly to 103.9%, but this still pretty good.
From here I would simply look at the best prices available across all bookmakers, and/or check the Betfair market. The Betfair market and most odds comparison sites (including the geegeez cards 'odds' tab - see image) give you the book percentage for the race so you can see how close that is to your book percentage.
The book percentages should be similar to yours, within a few percentage points at least. Betfair Exchange will always have a figure closer to 100% than the best bookie prices.
If you prefer to work with odds and then convert those to percentages, the 'My Ratings' icon will enable you to do this. Click the icon in the menu bar to open the ratings boxes for all runners, and then simply add odds to each runner. You'll see the percentages calculated for you, both for the individual runner and for the race overall:
At this point you need to compare the actual prices on offer with your own prices. We are looking for the positive outliers - horses that are priced higher in the real betting market when comparing them with your prices. Let us imagine in our example above that Horse C is best priced with the bookies at 9/1 (10.0) which equates to a percentage chance of winning of 10%. This gives us an ‘edge’ of around 5% assuming that our price is accurate! If our price is an accurate representation of the true percentage chance of the horse in question, then we have a value bet.
Method 2. Simple Rating Method
The second idea I want to look at is a basic rating idea. Once more I’ll assume we have a 7 runner race with exposed horses. Look at each of the horses in the race and their finishing position last time out. I would then look at the long term data in relation to that finishing position when racing in similar 7 runner races. For example, last time out winners win roughly 22% of the time in these races, horses that finished 5th last time win around 12%, etc. From here I halve each percentage and assign that as a rating figure to each horse.
Then I would look at four or five different factors (such as those I shared earlier in the ABCD method above), giving each horse either a positive, neutral or negative mark. For positives I would add two points to their total, neutral marks I would leave the rating as it is, for negatives I would subtract two. One of these factors could be, for example, last time out starting price. If a horse started at 3/1 last time, say, this would count as a positive and gain a ‘+2’. If another horse started at 25/1 on its prior run then I would give that a ‘-2’.
Once you’ve done this across all factors / variables you will have ‘rating’ totals for each horse such as below:
From here you can add up the final rating totals which in this case equals 62. Then divide the rating for each horse by 62 and multiply by 100 to give them a percentage chance. In this example we get the following:
With the percentages of the seven horses adding up to 100% we have a perfect betting book. From here we would change these percentages into betting odds in the same way as before. If you are unsure how to convert percentages into odds, then there are odds calculators that you can google and use; or enter the odds into the geegeez ratings boxes like in the image above.
I personally use an excel sheet with a simple formula to calculate the individual horse percentage chances, and the sheet also calculates the percentage book for the race. It is not too difficult to set up and anybody interested please request in the comments below.
Once we have our odds, as previously, we would cross check our prices against the actual odds on offer looking for any horse that offers ‘value’.
Method 3. Using a ‘benchmark’ horse or horses
This is the method I used for these types of races when I was writing Spotlight for the Racing Post many moons ago. Interestingly, Andrew Mount, who co-wrote a draw bias book with me 20 years ago and also worked as a Spotlight writer for many years, used the same idea in certain races for more exposed runners.
The ‘benchmark’ horse is basically the horse which you think is the easiest to accurately price up / give a percentage chance of winning. Each race will of course be slightly different and hence each ‘benchmark’ horse is likely to fit a slightly different profile. The types of horses I’d be looking for are either:
a horse that stands out on form – in other words the horse you believe should start favourite. The potential price of your favourite will depend on your perceived chance of the horse winning. If you thought the horse was likely to win this type of race one in every three starts you would price the horse at 2/1 (3.0); if you thought it was likely to win the race one in two you’d price it up at Evens (2.0), etc.
a horse with a consistent recent record; hopefully one with solid form figures over the last three or four races. Ideally this horse would have gone off at similar prices too in these races. So if we say had a horse that had finished 4th, 2nd and 3rd in their last three runs, starting at prices of between 3-1 and 5-1, then you would probably could price the horse up at around the 4-1 mark.
Once you have a price for your ‘benchmark’ horse you can build the prices of other horses around it. Horses you think that have a better chance than your ‘benchmark’ will be put in at a shorter price, those with a worse chance in your opinion will be priced higher. You may find that two or more horses in a race have either profile 1 or 2 mentioned above, which will speed up the process slightly. Naturally, you can use the odds boxes behind the My Ratings icon on geegeez racecards to help you price up.
Once you have priced up the other horses around your ‘benchmark’ horse(s), then you will need to see what the percentage book for the race equals and it may need some slight adjustment of prices up or down. Once you have a book close to 100% then you can compare prices with the best prices on offer as we did for the previous two methods.
**
For those of you who have not attempted to create odds lines before, I hope these ideas make the task feel less daunting. Of course, using any of the three methods suggested does not mean you will be getting that much needed value race in, race out. It is going to essentially depend on how accurate your odds line is. The only way to find out how accurate it is, is by choosing a method, either one of the above, or a method of your own, and testing it out. One indication that you have a sound method is if your prices generally ‘mirror’ the actual market. If they don’t and, say, you price up horses at 2/1 that are regularly available at 5/1 in the real world, then you need to go back to square one and figure out why the disparity.
If the prices you create generally look reasonable compared to the actual prices on offer then I would suggest pinpointing those horses that look good value when comparing your odds with those of the bookmakers. Paper test these value selections with imaginary £1 stakes and see what happens over a series of ‘bets’. If you are in profit by the end of your testing phase then there are hopeful signs that your odds line is pretty accurate.
How many of these imaginary £1 bets you should work through during the testing phase is not easy to say, but I personally think you need to have at least 200 to give the test some validity. Of course, you could also look for horses that seem poor value and see how they performed as well. All of this will give you a decent overall picture of the effectiveness of your tissue making skills. Once you are happy with your method, then I would suggest looking at races with bigger fields but still races where the horses are exposed.
One final point to note is that I would look at a race as early as possible. With 48 hour declarations you can price up the race before the bookmakers and give yourself plenty of time. Also pricing it up early avoids you being influenced by early odds or the Racing Post forecast prices.
I would not advocate the same method for every single race – for example you cannot use any of the above ideas with a 2yo race with little or no form. Perhaps in the future I will revisit this topic and look at how you may try to create odds lines in other races such as these. For now, though, I hope this article has encouraged you to look into creating your own odds lines – even if you do not find the holy grail, the process you go through will, I am sure, be useful and enlightening for your overall betting / value understanding.
You type in the odds (traditional) and the spreadsheet will calculate the percentage for each horse, the book percentage and the overround. As long as users take care to retain the formula in cell C2 then it is merely a case of deleting the column B odds and the column C percentages (cells C3 and below). Then, once you've typed in the new odds for your race you can simply copy and paste the formula in cell C2 down column C as far as you need to.
If you lose the formula in cell C2 just re-type =100/(B2+1) in to that cell. The total book percentage and overround will provide the relevant percentages for each race.
https://www.geegeez.co.uk/wp-content/uploads/2016/03/CheltenhamFestivalBookmakerOffers5.jpg320830Dave Renhamhttps://www.geegeez.co.uk/wp-content/uploads/2022/10/geegeez_banner_new_170x78.pngDave Renham2023-07-18 16:59:282023-07-23 21:19:01How to Create an Odds Line / ‘Tissue’
Regular readers of my Geegeez articles will know that probably my favourite area of horse racing research is connected with the run style of horses, writes Dave Renham. This is mainly due to the fact that at shorter distances early leaders/front runners tend to have an edge over horses which initially take up a prominent, mid-division, or held up position. Indeed, as I've observed before, this front running edge is extremely potent at a good number of course and distances. However, there are plenty of races where front runners do not have an edge, and hold up horses start to become much more competitive. In this article I am going to explore this area, and I will begin by digging into some general stats.
For this piece I will be looking at UK racing from 2015 to 2022 with the focus on 8+ runner handicaps.
General Hold Up Run Style Data
Let's start by looking at a graph comparing front runners with hold up horses across all the flat distances, looking firstly at win strike rate:
This graph illustrates neatly how the general advantage to front runners drops as the race distance increases. We do not really have to worry about different field size averages for different distances, because we are basically comparing the strike rates for one run style group against the other at each specific distance range. However, it should be noted that in any race there are almost always going to be more hold up horses than front runners. In a 12-runner race for example, we might expect to see one front runner, maybe two; but in terms of hold up horses we are likely to have three or perhaps four. This is something to keep in mind when comparing run style win percentages.
If we look at the A/E indices*, a measure of value, we see excellent correlation with the win SR% graph:
*You can read more about A/E here
In 8+ runner handicaps of 1m5f or more there is virtual parity in terms of betting value between front runners and hold up horses. Although just about equal, however, following either run style as a betting approach is a sure route to potlessness! Sticking with these longer races, there are some interesting findings when we break down results by going. Here are the win strike rates for front runners / hold up horses when comparing results on good or firmer ground with good to soft or softer:
As we can see the going on turf does seem to make a difference in 1m5f+ handicaps. On firmer ground there is a smaller difference between the records of both run style groups, when compared with data on softer ground. It seems harder to make up ground from the back on a softer surface.
If we look at the all-weather data for these longer races, we can see a different outcome from the turf:
Hold up horses actually have a better strike rate in longer handicap races on the synthetics with front runners struggling, winning less than 1 race in every 14. There is also a big difference between the all-weather A/E indices, with front runners standing on a lowly 0.61 figure (akin to punting suicide) and hold up horses at 0.86.
Let's now dig into some individual course data in terms of hold up horse performance. In the graph below we can see a comparison between courses that have the highest A/E indices for hold up horses versus those with the lowest. The top 10 course figures and the lowest 10 figures are shown – all distances have been combined:
There is a huge difference between the top figures and the bottom ones: Yarmouth heads the list with a highly credible A/E of 0.95 which is a world away from Ripon’s hideous 0.53 figure. The ‘returns’ for hold up horses at each of these courses mirror the A/E index chasm with Yarmouth hold up horses losing 18.7 pence in the £ at Starting Price, while Ripon hold ups lost a massive 53.7 pence in the £.
Course Specific Hold Up Run Style Data
We will look now at some specific track statistics concerning hold up horses, beginning at the Norfolk seaside.
Yarmouth 5f
Here are the win strike rates for each run style category over Yarmouth's 5f distance:
This is a highly unusual set of run style figures: the minimum trip at most flat tracks shows the highest positive edge to front-runners. Looking at the ALL courses data combined over five furlongs, front runners win 18.2% of races while hold up horses are down at 6.5%. But at Yarmouth we have the reverse with the strike rate for hold up horses being three times that of front runners.
In addition to the win stats, the A/E index for hold up horses over 5f at Yarmouth stands at an impressive 1.04. Sticking with A/E indices, at five of the eight distances run at Yarmouth hold up horses have secured a figure of above 1.00, suggesting the advantage to patiently ridden horses is underbet.
Newcastle 1m
I have always felt that the all-weather track at Newcastle is one where hold up horses are competitive due to the stamina-sapping nature of the configuration coupled with the uphill half mile finish in the straight. I am pleased to see the stats back this up. It should be noted that for Newcastle I am looking at data from 2016 onwards, when the course changed from a turf course to an all-weather one.
The distance where hold up horses do best at Newcastle is over 1 mile. This trip is the longest of the four distances on the straight course, and it clearly plays more to the strengths of hold up horses.
These strong figures for late runners are replicated when we look at the Percentage of Rivals Beaten (PRB) data:
The held up PRB figure of 0.55 is a strong one. Closers have actually made a blind profit to Industry SP despite there being nearly 1400 of them. Such runners have secured returns of just over 3p in the £. Compare this to the combined losses of the other three run style groups which stands a depressing 29p in the £.
Digging deeper into hold up horses over 1 mile at Newcastle, when they have started in the top three of the betting they have won 80 races from 360 (SR 22.2%) for an SP profit of £49.98 (ROI +13.9%). Hence, a well fancied hold up horse over this course and distance is definitely worth a second glance.
The longer distances of 1m4f and 2m at Newcastle see front runners having a really hard time of it winning under 6% of all races and backing all front runners would have yielded heavy losses of 54p in the £.
Doncaster 1m4f+
In races of 1m4f or more on Town Moor, hold up horses perform strongly as any group, while front runners have really found it hard going. Taking the data straight from the Geegeez Pace Query Tool we see the following:
There were just two wins from 92 runners for horses that took the early lead in such races between 2015 and 2022; and front runners as a group also had notably the poorest placed record. One plausible reason for these humbling figures, along with the fact that we are dealing with longer distances, is that the Doncaster straight is 4½ furlongs in length. I am a believer that longer straights as a rule tend to be harder for front runners to maintain their advantage when compared to tracks with shorter straights.
Over 5f at Doncaster front runners also had a hard time of it, winning just twice from 54 runs. Hold up horses at that trip edged it over the other three run styles winning just shy of 10% of the time (A/E 0.98).
Ascot 1m
1 mile handicaps at Ascot are often extremely competitive and, from a run style perspective, hold up horses do best. Here are the splits:
These are highly noteworthy figures for hold up horses considering the stats for ALL courses combined over 1 mile (8+ runner handicaps) sees front runners winning 12.8% of the time and hold up horses just 7.6% of the time.
The PRB figures are also very strong for hold up horses as the graph below shows:
Backing these 2015-2022 data up, both the Royal Hunt Cup (22/1 Jimi Hendrix) and the Britannia Handicap (6/1 Docklands) were won by hold up horses at the recent Royal Ascot meeting.
*
Earlier I looked at some data for all courses across all distances. Having looked at a few specific course and distances, I want to now share data for more courses at two different distance ranges.
Run Style Negatives: Front runners in handicaps of 1m4f+
At the beginning of the article when looking at long distance races I used 1m5f or more as my cut off point. However, in order to get better and bigger data sets when looking at individual courses (rather than ALL courses), we need to include races of 1m4f or more.
Below are the courses where front runners have the lowest win strike rates at distances of 1m4f+ – the ten lowest in fact (at least 45 races during the study period):
Doncaster and Newcastle, which we have previously discussed, top the list. It is also no surprise to see four of the six all-weather courses in this table considering what we found earlier with the overall 1m5f+ AW data.
Here now are those courses with the lowest A/E indices:
This table correlates well with the first one, eight of the ten tracks appearing on both lists – Doncaster, Goodwood, Newcastle, Brighton, Chelmsford, Ayr, Wolverhampton and Kempton.
It is clear that if we are ‘sweet’ on a front runner at any of these courses in handicaps races of 1m4f or more, we need to be really sweet! The stats are truly against us.
Run Style Negatives: Front runners in handicaps of 1m 1f to 1m3f
My final port of call in this piece is 8+ runner handicaps incorporating races from 1m1f to 1m3f. Below is table showing performance data for all courses with at least 45 qualifying races, ordered by win strike rate. As can be seen, there is quite a difference between York, with the poorest front running record, and Musselburgh (the best):
This table illustrates how important it is to appreciate that UK courses vary so much when analysing certain stats sun as run style ones. That should come as no surprise because the turf courses especially are so different: some sharp, some undulating, some stiff, and so on. That is why the pace maps on the Geegeez racecard are like gold dust. Having past run style insights for a specific course and distance (and going and field size) gives us a huge edge when gauging how important a factor run style is likely to be.
As you know, I am a huge believer that run style can be the key to unlocking the winner of many races. It is something I strongly feel that all punters should include in their form analysis. I hope that, if you're not already, the findings in this article might encourage you to start!
Good luck.
- DR
https://www.geegeez.co.uk/wp-content/uploads/2020/04/HighlySprung_SdS.jpg319830Dave Renhamhttps://www.geegeez.co.uk/wp-content/uploads/2022/10/geegeez_banner_new_170x78.pngDave Renham2023-07-11 08:01:182023-07-11 08:01:18Run Style: When Early Leaders *Don’t* Have The Edge
This is the fifth and final article in my series of articles on jockeys, writes Dave Renham. In this piece I will be looking at three more top jockeys trying to pinpoint their strongest stats, be it positive or negative. As with the previous four articles I have analysed the last eight full years of flat racing in the UK and Ireland (2015-2022). I have used the Geegeez Query Tool as well as the Profiler Tool, amongst other things. In all the tables the profits/losses quoted are to Industry SP, but I have shared Betfair Starting Price where appropriate. Let's start with last season's champion jockey...
William Buick Jockey Profile
William Buick became Godolphin’s first choice jockey in 2016 and hence it should come as no surprise that within a year his win strike rate soon began to edge up:
As we can see from 2017 onwards he has achieved yearly strike rates in excess of 20%, with 2022 being a particularly good year. His overall record reads as follows:
Buick backers incurred relatively modest losses to Industry SP when we look at all races as a whole. Considering he has had over 4000 rides this is quite impressive. To BSP, backing Buick ‘blind’, you would have made a profit of £317.71 (ROI +7.5%).
Let us now look at his performance for different trainers over this eight year period (minimum 100 rides):
Buick when teaming up with his boss, Godolphin trainer Charlie Appleby, has secured a strike rate edging close to three wins in every ten rides. Not only that, they have combined to virtually break even to SP, with profits to BSP hitting £139.33 (ROI 9.6%). Indeed, to BSP they have secured profits in six of the last seven seasons. His record is less impressive when riding for the Gosden stable – a stable for whom he has been stable jockey in the past - with a modest strike rate of under 15% and poor returns.
One trainer not in the table due to the minimum ride stipulation is Sir Michael Stoute. Buick and Stoute do not team up that regularly, but when they do their record is excellent – 19 wins from 73 (SR 26%) for a profit of £42.07 (ROI +57.6%). To BSP profits that increases to £57.59 (ROI +78.9%). Their PRB figure is excellent also standing at 0.68.
One thing I like about Buick is that he is an excellent rider from the front. He wins on board virtually one ride in three when taking the early lead. Here are his win percentage splits for the four main run styles:
Buick follows the usual trend in that his front running rides win more often than his prominent ones which in turn out-perform mid div / hold up rides. For the record, at distances of 1m2f or less his front running strike rate stands at 35.1%; at 1m3f or longer it drops to 19.1%.
As regular readers will know, I like to look at favourite run style data, too, as this eliminates any potential selection bias regarding ‘good horses at the front, bad ones at the back’. Here are the relative win strike rates for Buick-ridden favourites in terms of the four main run styles:
Again, front running market leaders did best by some margin, while favourites that raced mid-division early had a very poor record: these runners would have lost you 26p for every £1 bet. Buick's record on held up favourites are a lot stronger than most jockeys, presumably because of the number of Godolphin horses able to outclass their opposition.
Before moving on, let us look at some additional statistics for the reigning champ:
Buick has a great record at Newmarket from a significant number of rides. Specifically, he scored on 212 winners from 843 (SR 25.2%) for a BSP profit of £94.09 (ROI +11.2%). When riding for Charlie Appleby at HQ the record is even more impressive – 132 winners from 412 rides (SR 32%) for a BSP profit of £93.34 (ROI +22.7%).
In contrast, at York his record reads 24 wins from 197 (SR 12.2%) for a BSP loss of £41.53 (ROI -21.1%).
On 2yos Buick has won 25% of races returning a BSP return of 6p in the £.
On 2yos having their second career start Buick has a strike rate of 1 in 3 and has returned a profit to BSP of just over 15 pence in the £.
Buick is a very good all round jockey who I am always happy to see riding a horse I fancy.
Jim Crowley Jockey Profile
Jim Crowley is a seasoned campaigner, and retained rider for the Shadwell operation, who is right up there when it comes to win rate. Here is his overall record going back to 2015:
These are excellent stats and backing all Crowley runners to BSP would have yielded a profit of £424.79 to £1 level stakes, equating to returns of nearly 8p in the £.
Crowley rides for numerous different trainers and there are 16 trainers for whom he has ridden more than 100 times. Here are their stats:
We see some very good stats here with seven of the 16 trainers showing a blind profit to Industry SP; and 11 trainers showing a profit to BSP.
Crowley has produced excellent results with horses from the top two in the betting when riding for Owen Burrows, William Haggas and the Gosden stable. All three have yielded good BSP returns on investment (Burrows +19.8%, Haggas +16% and the Gosden stable +8.6%).
In terms of courses, Crowley has ridden more than 100 times at 18 different venues. Here are the A/E indices at these tracks:
It is very impressive to note that eight courses have A/E indices in excess of 1.00 with Nottingham hitting a remarkable 1.57. His overall Nottingham stats are unsurprisingly outstanding – 43 wins from 131 rides (SR 32.8%) for an SP profit of £159.14 (ROI +121.5%). To BSP this improves to a profit of £186.86 (ROI +142.6%). His PRB course figure is also very strong standing at 0.65.
Here are a couple of stats for Crowley that are also worth sharing:
He has an excellent record in very small fields. In races of five runners or fewer he has won 144 races from 410 rides (SR 35.1%) for a BSP profit of £130.46 (ROI +31.8%). He has made a profit to industry SP also of £84.64 (ROI +20.6%).
On front runners he has performed especially well for trainers Charlie Hills and Owen Burrows. This is particularly true in races of 1 mile or less where Crowley hits the 34% win percentage mark for both trainers.
Crowley is hugely experienced and this shows in his stats.
Oisin Murphy Jockey Profile
Oisin Murphy was British Champion Jockey in 2019, 2020 and 2021. He did not race in 2022 as he was banned for two failed breath tests and breaking coronavirus rules but has resumed riding with a win percentage of 17.5% in 2023, slightly above his overall record as can be seen in the table below:
These are sound stats given Murphy has taken over 2000 more rides than Buick and 1000 more than Crowley, despite missing the whole of 2022! He clearly is a rider who does not have an issue with being busy. Like Crowley he has ridden 100 times or more for several trainers and here are the stats (ordered by strike rate):
Although he has not made a profit to SP when riding for Saeed bin Suroor, they are a combination to keep an eye on. The PRB of 0.70 is particularly high and, when betting to BSP, they have snuck into profit. Indeed to BSP, all bar Simcock and Williams have produced a profit with Oisin in the plate. Keeping with the BSP theme, if we combine all nine trainers, then Murphy has made a profit with them as a group in every year from 2015 to 2021. The combined yearly returns to BSP are shown in the graph below:
It is rare to get seven profitable years in a row when combining as many as nine different trainers.
There are four other trainers to keep an eye out for where Murphy has had less than 100 rides in each case. They are the Harry & Roger Charlton barn (10 wins from 32), Mick Appleby (16 wins from 66), John & Thady Gosden (31 wins from 84) and John Butler (8 wins from 21).
Murphy has a notably good record on 2yos with an overall strike rate in the review period of 17.4% thanks to 256 winners from 1473 runners. To Industry SP these runners yielded small losses of just under 4p in the £; to BSP, however, this turns into a profit of over 13 pence in the £. Here are three additional 2yo stats worth sharing:
2yos that have started in the top four of the betting have provided 226 wins from 971 runners (SR 23.3%) for a BSP profit of £92.32 (ROI +9.5%)
For the Gosden stable he has had 14 2yo winners from just 39 runners (SR 35.9%) for a BSP profit of £12.08 (ROI +31.0%)
2yos that Murphy has taken into the lead early have won over 30% of their races. But...
2yos that were held up by Murphy have won just 8.4% of the time
Continuing with the run style theme, I have always liked Murphy from the front as an angle. Indeed, if your crystal ball was in mint condition and you had predicted pre-race all of Oisin's front runners in all races (not just 2yo ones), you would have been rewarded with an SP profit of £312.85 (ROI +30.9%). To BSP returns were nearer 45p in the £.
Looking at his run style record on favourites we see the same pattern we have seen numerous time before:
Front running favourites do best as is the norm and they would have been profitable to the tune of 12p in the £. Prominent racers would have seen you lose 2p in £, mid div 'jollies' lost 24p for every £1 bet, while hold ups lost 19p.
Here are some additional stats for Murphy, starting with two negative ones:
Murphy has a poor record with very short priced runners. On horses priced 8/13 or shorter he has had 61 wins from 112 (SR 54.5%) for losses to Industry SP of £28.10 (ROI –25.1%)
With big-priced runners his record is poor also. Horses priced 28/1 or bigger accounted for just four winners and nine placed runners from 337. Losses to Industry SP stood at £206.00 (ROI –61.1%). To BSP it improves a little but he still lost over 42p in the £
Murphy has achieved a strike rate of 20% or more at five courses (with a minimum of 100 rides) – these are Chelmsford 20.1%, Newcastle 21.5%, Nottingham 20%, Salisbury 21.1% and Wolverhampton 20.5%. Four of the five have yielded blind profits to BSP (Wolverhampton being the only track that has not)
When teaming up with Hughie Morrison at Nottingham they are 6 wins from just 13 runners. They have also had two seconds at 14/1 and 12/1. When riding at Lingfield for Archie Watson, Murphy is 12 wins from 35 (SR 34.3%) for a BSP profit of £11.80 (ROI +33.7%)
I really like Murphy as a jockey and I especially look for horses he is riding that may take the lead early.
MAIN TAKEWAYS
Below is a summary of my key findings, firstly for William Buick:
Buick has a good record riding for his boss Charlie Appleby, making a blind profit to BSP with a decent strike rate. He also has a good record when booked to ride for the Stoute stable
He is outstanding from the front especially in races of ten furlongs or less.
He has a very good record at Newmarket for all trainers, but especially with Appleby. At York his record is relatively poor.
His record with 2yos is decent, with second starters doing particularly well.
Onto Jim Crowley now:
Crowley has a strike rate of better than one win in four with four trainers (100 rides or more) – John & Thady Gosden, William Haggas, Roger Varian and Owen Burrows. Three of the four have yielded a profit to Industry SP
He has an outstanding record when riding at Nottingham
In small fields of five runners or less Crowley has been exceptional
And finally Oisin Murphy:
Murphy has a good record with many trainers he rides regularly for.
Harry & Roger Charlton, Mick Appleby, John & Thady Gosden, and John Butler are trainers he rides less often for but his record with all four is excellent.
He goes well on 2yo runners.
He is excellent when riding from the front.
He has a relatively poor record with very short priced runners (8/13 or shorter); likewise with outsiders priced 28/1 or bigger.
Two trainer/jockey course combinations to note are Murphy and Morrison at Nottingham, and Murphy and Watson at Lingfield.
*
So I have come to the end of this series on jockeys. Of course, I have barely scratched the surface as there are hundreds of riders I have not analysed at all. Most punters have favourite jockeys or indeed ‘lucky’ ones, but digging into the stats is a worthwhile use of all of our time. Building up a picture of strengths and weaknesses is important, and with Geegeez’s tools - especially the Profiler and Query Tool - it is not difficult to do or time consuming. In fact, it's fun!
Other jockeys you may want to look at in your own time include James Doyle, Andrea Atzeni, Jack Mitchell, Kevin Stott and Adam Kirby; or indeed whoever interests you. If you find anything noteworthy, feel free to comment below as it will help the Geegeez community. Until next time, when I'll be looking at something different, stay lucky.
- DR
https://www.geegeez.co.uk/wp-content/uploads/2023/07/StarofMystery_WilliamBuick_CharlieAppleby.png319830Dave Renhamhttps://www.geegeez.co.uk/wp-content/uploads/2022/10/geegeez_banner_new_170x78.pngDave Renham2023-07-04 18:40:162023-07-05 16:42:29Jockey Profiles: Best of the Rest
This is the fourth in my series of articles on jockeys, and this time I am examining the performance of some riders in Ireland, writes Dave Renham.
As with the previous three articles I am analysing the last eight full years of flat racing (2015-2022) but focusing solely on Irish results. For the majority of the number crunching I will be using the Geegeez Query Tool, but I will also use the Profiler tool amongst other things. In all the tables the profits/losses quoted will be to Industry SP, but I will share Betfair SP where appropriate.
The first point to make is that you should not blindly compare Irish jockey strike rates with their UK counterparts. This is because the average field size in Ireland is bigger than it is in the UK. In the past eight seasons, the average number of runners in a UK flat race stands at 9.2; in Ireland this jumps markedly to 11.7. If we compare by year we see that the gap in the last couple of years has increased further:
Hence strike rates for jockeys racing in Ireland are going to be lower than for jockeys racing in the UK. If we want to compare jockeys across the Irish Sea against each other, then the PRB figure (Percentage of Rivals Beaten) is a better barometer to use.
Jockey Performance in All Races
Let us first look at all jockeys that have ridden at least 400 times in the past eight seasons in Ireland. I have included all of them, rather than hand pick those with the highest strike rates. The reason for this is that I do not know that much about some Irish jockeys so I am keen to absorb all the stats I can:
As we can see there are no jockeys in profit to SP with many heavily in the minus.
Anyone who read my Ryan Moore article will be familiar with his overall stats. Moore is comfortably ahead of the rest with a crazy strike rate, thanks as we know in the main to his partnership with trainer Aidan O’Brien. The next best strike rate is owned by Colin Keane, on 14.83%, which is less than half the figure of Moore! Speaking of Keane, let us dig a little further into his stats:
Colin Keane
Keane is stable jockey to Ger Lyons, a relationship that began in 2014. Keane has been Champion Jockey in Ireland in four of the last six seasons (2017, 2020, 2021, 2022), and in 2021 he had his highest number of wins in a season with 156. Let us look at his record with different trainers (minimum 50 rides), ordered by number of runs:
There are some strong PRB figures here, with Keane securing a PRB of 0.60 or better with eight different trainers. Naturally, the most rides have been for Lyons, but the O’Callaghan combination completely stands out. In 2022, they teamed up 18 times and nine of those horses ended up in the winner’s enclosure. They have partnered up at 14 different Irish courses and had winners at ten of them! Three of the courses where they have not had a winner have been at courses where Keane has ridden for O’Callaghan just once.
Two trainers perhaps to avoid are Martin and Mc Court – both have relatively poor figures in comparison to the average, though are still at least 50% of rivals beaten together.
For Ger Lyons, Keane is close to one win in five which is excellent. Here is a graph of their trainer/jockey combo in terms of yearly strike rate – looking at both win strike rate and each way strike rate:
There were slight dips in 2018 and 2022, but generally quite consistent figures. It should be noted that 2023 started very poorly, so this is something we need to keep our eye on. Things have improved in June and hopefully the pairing are back to normal service resumed now.
Here are three of the most potent Keane/Lyons stats:
2yo horses have done extremely well with 177 wins from 759 runners (SR 23.3%). To SP, returns have shown a small 3p in the £ loss. To BSP though a profit of £65.91 would have been made equating to returns of over 8p in the £.
Clear favourites have won 233 races from 572 races (SR 40.7%) for a BSP profit of £49.66 (ROI +8.7%).
Horses making their debut have an outstanding record. Of the 333 debutants, 71 have won (SR 21.3%) for a BSP profit of £158.33 (ROI +47.6%). Profits to SP were smaller but still returned just under 20p in the £.
Moving back to looking at Keane’s overall record again, it is time to consider some of his run style data. Geegeez members will know I am a big fan of looking at favourite run style data as this eliminates any potential selection bias regarding ‘good horses at the front, bad ones at the back’. Here are the relative win strike rates for Keane horses that have started as the market leader in terms of the four main run styles:
No surprises here with front running favourites doing clearly best. This is an even stronger bias than we normally see with hold up favourites scoring less than one win in every four. Front runners, meanwhile, would have secured a profit of around 24p in the £ to SP assuming our crystal ball could have accurately predicted that they would all go forward as well as being favourite. This profit would jump to 30p in the £ if backing them all to BSP.
Seamie Heffernan
Heffernan has some interesting run style stats when we focus on shorter distance races of 5f to 7f. Below are his strike rates both from a win and each way perspective:
As can be seen, Heffernan’s record on front runners from both a win and placed viewpoint is top notch. The figures for hold up horses in these shorter distance races are very weak – fewer than one in twenty winning, fewer than one in eight placing.
Heffernan has ridden 103 front runners in these 5-7f races for trainer Aidan O’Brien and has won on 38 of them (SR 36.9%). For the same trainer over the same distance spread, we see hold up horses claiming just 12 wins from 153 (SR 7.8%). Of the 153, 77 came from the top three in the betting! Now I appreciate I probably have the largest and noisiest ‘drum’ when it comes to run style stats in the whole of the racing world but when the numbers look like this, I just have to make you aware.
Horses from top three in the betting, by jockey
As the main table indicated, most jockeys have modest profit records at best when looking at their rides as a whole. Let’s look at how they have performed in terms of when they are riding a fancied runner – specifically, a horse in the top three in the betting. Here are the jockeys who have secured the best strike rates (minimum 100 runners):
Moore tops the list once more; Keane is in 5th, while three jockeys have managed to secure a profit to SP, namely Shane B Kelly, Ben Coen and Connor King. The average A/E figure for all Irish riders on horses from the top three in the betting is 0.88, so a few of them are nicely above this figure.
A look now at the jockeys with the lowest strike rates (below 16%) with the same group of fancied horses:
These jockeys are probably ones to be wary of even if riding a horse that heads, or is near the head of, the market. They have produced some hefty losses as a group.
Jockeys on front running favourites
Earlier we saw that Colin Keane had an excellent record on favourites that took the lead early. Here are the jockeys with the highest strike rates with such runners, of which Keane is one of them:
Absolutely exceptional figures for Moore; in the previous article on Ryan I noted his excellent record on front running favourites when combining UK and Irish stats. To that we can now add that his Irish front running win stats are 15% higher than his UK ones. I also mentioned in that piece that Moore does not go to the front early as often as he should – this cements my feeling for time immemorial. Of course, many of Aidan O’Brien’s horses are steadily away which makes getting to the front more difficult.
Jockey Performance, by Racecourse
For this section I decided I would look for any big positives or negatives at the Irish courses as regards to jockeys. Here are my findings:
Ballinrobe – Shane Foley has the highest strike rate at the course (19.3%) thanks to 11 wins from 57; Rory Cleary is 0 from 41. To be fair Cleary has not had many good chances at the course;
Bellewstown – Declan McDonogh has a 20% win rate at the course (11 from 55) and provided a BSP profit of £32.79 (ROI +59.6%); he has a decent placed record too. Dylan Browne McMonagle has managed just two successes from 52 including just one from 22 with horses 7/1 or shorter;
Cork – Billy Lee has ridden 45 winners at the track in the past eight seasons (Colin Keane also has won 45) from 245 rides. He has secured strike rates above 20% at Cork in three of the past five seasons, and in six of the eight seasons you would have made a profit backing his runners to BSP. When teaming up with trainer Paddy Twomey, Lee has ridden 13 winners from just 33 runners which equates to a strike rate of 39.4%.
Curragh – Ryan Moore has a good record here with 109 winners from 393 rides (SR 27.7%). To BSP his mounts have virtually broken even. His record in Group 3 contests is eye catching – 29 wins from 59 (SR 49.2%) for a BSP profit of £48.07 (ROI +81.5%). Contrast Moore with the Curragh stats for Connor Hoban who has managed just two wins from 197 runners, though again opportunity is obviously not created equally for the two riders.
Dundalk – the course that stages by far the most Irish racing due to it being an all-weather track. Colin Keane seems to ride the track as well as any – he has had 1210 rides with 198 successes (SR 16.4%). A BSP profit of £81.56 (ROI +6.7%) would have been achieved backing all his rides blind. His record is quite consistent when analysing the data by year. Keane’s strike rate exceeds 20% when riding for his boss Ger Lyons and when riding for Noel Meade.
Fairyhouse – Rory Cleary is 0 from 101 at the track in the study period.
Killarney – Declan McDonogh is head and shoulders above the rest here with 26 wins from 104 rides (SR 25%) for a BSP profit of £170.64 (ROI +164.1%).
Leopardstown – Ryan Moore has 42 victories here and is just half a percentage off hitting a 30% win strike rate. You would have lost 11p in the £ however, even to BSP.
Naas – it is Ryan Moore again who has by far the best strike rate at 35.8% (29 wins from 81) for a 4p in the £ BSP return. Colin Keane and Seamie Heffernan are the only other two jockeys above the 15% mark.
Navan – Moore is a rare visitor here but has an impressive 13 from 31.
Tipperary – Billy Lee has the best record here – 39 wins from 214 (SR 18.2%) for a BSP profit of £72.48 (ROI +33.9%).
MAIN TAKEWAYS
Let me summarise the key findings:
Irish races have bigger field sizes so we need to appreciate that when we compare Irish strike rates with UK ones;
Ryan Moore has a 3 in 10 strike rate in all races. He has a fantastic record on front running favourites. He has a decent record at several tracks but take note whenever he makes a trip to Navan;
Colin Keane has a very good record on debutants when riding for Ger Lyons. His ‘all runners’ record is outstanding for Michael O'Callaghan (though steadier so far in 2023). He also rides Dundalk as well as anyone and has an excellent strike rate on front running favourites;
Shane B Kelly, Ben Coen and Connor King has proved profitable to follow when riding a fancied horse (first three in the betting);
Seamie Heffernan rides from the front exceptionally well in sprint races (5f to 7f). His record over the same distances on hold up horses is very poor;
Billy Lee has good records at both Cork and Tipperary – each was highly profitable during the review period;
Declan McDonogh is a jockey to follow at Killarney. His record is far superior to the rest.
----
This article has been very interesting to me to research because I personally rarely bet in Irish races; but during the research I’ve found a number of avenues to attack the flat racing puzzle there. I hope the findings have been interesting for you, too.
This is the third in my series of articles on jockeys, and in this one I am examining the two jockeys who have ridden the most winners at northern or Scottish tracks in the past eight seasons, namely Danny Tudhope and Ben Curtis, writes Dave Renham. Between them they have ridden over 7500 horses in this part of the UK, winning 1242 races (Tudhope 687 wins, Curtis 555), and these runners have accounted for about 75% of Tudhope’s total rides in the UK/Ireland and about 64% of Curtis’s. They have both been successful ‘down south’ as well; Tudhope, for example, has ridden four winners on two separate occasions in Royal Ascot festivals – once in 2019 and then again in 2022.
As with the previous two articles I am analysing the last eight full years of flat racing in the UK and Ireland (2015-2022). I am using the Profiler Tool along with the Query Tool as the main vehicles for my data gathering. In all the tables profits/losses quoted are to Industry SP, but I will quote Betfair SP where appropriate.
Let’s start with Tudhope.
Danny Tudhope Jockey Profile
Danny Tudhope: Overall Record
Let me first review Tudhope’s overall stats by looking at his performance on every single runner during this eight-year period:
This is a very presentable set of figures – a win rate of roughly one win in every six and very modest losses of just over 7½ pence in the £ to SP. Indeed, to BSP this would have been converted into a profit of £317.49 (ROI +6.1%), with five of the individual years showing 'in the black' against the machine. Tudhope's A/E index, a ratio that essentially determines value, is above the average for all jockeys, as is his PRB figure.
Danny Tudhope: Record by Year
Yearly stats are the next port of call. Here is a breakdown by both win and win/placed (Each Way) percentage / Strike Rate (SR%):
As can be seen, 2019 was his best year hitting the winners' enclosure on nearly one in five of his rides. Overall, Tudhope's performance has been very consistent both from a win and placed perspective, which is something one always likes to see. This consistency can be viewed even more clearly when we look at his yearly PRB (Percentage of Rivals Beaten) figures:
Danny Tudhope: Record by Betting Odds / Price (SP)
The Profiler on Geegeez gives a breakdown of performance by Starting Price splitting the market into seven price brackets. Tudhope’s figures are as follows:
At the shorter prices (9/4 or less) his figures are slightly below what would be expected, certainly in terms of returns. The remaining figures are slightly above what might be expected in terms of returns. His strike rate of 5.16% on horses priced 16/1 to 25/1 is well above the figure for ALL jockeys, which stands at only 3.47%. The same is true when looking at the 9/1 to 14/1 price bracket – Tudhope’s SR% stands at 7.96%, the ALL jockey figure stands at 6.56%. These mid- to bigger-priced horses have definitely offered some value for punters over the past eight seasons, though with single digit hit rates, it can be a long time between drinks!
Danny Tudhope: Record by Distance
A look at Tudhope’s record at different distances now. I have grouped them into five distance bands as with the previous two jockey pieces, and once again it is win strike rates that are being compared:
Similar strike rates, although the longest distance win percentage is slightly below the others. This might be due to the smaller sample size of 139 races. Tudhope primarily rides in races of 1 mile or less – roughly 75% of all his rides have been over these shorter distances with an even split between 5-6f and 7f-1m. Personally, I am a fan of Tudhope in sprints – he is excellent when on a front runner in these 5-6f races, winning over 30% of the time (92 wins from 300). Backing all these runners would have yielded a profit of £161.01 to SP (ROI +53.7%). Clearly we are never totally sure which horse is going to front run, but if a Tudhope sprinter does go to the front early it is cause for optimism. For the record, his returns on front runners have been virtually identical in handicap sprints and non-handicap sprints.
He also has a winning strike rate of around 27% in 7f-1m races on front runners which is equally as eye-catching, if not more so (N.B. average SR% for ALL jockeys on 7f-1m front runners is 18%).
Danny Tudhope: Record By Racecourse
I am now going to look at all courses where Tudhope has had at least 150 rides. The courses are listed alphabetically:
As expected, the majority of the courses in the table are northern English or Scottish tracks. Overall there is a fair smattering of profitable courses. When looking at market factors and taking out some of the bigger priced winners, four courses stand out, namely Beverley, Musselburgh, Redcar and Ripon. Tudhope riding at any of these should generally be considered to be a positive. At Beverley it is worth noting that restricting Tudhope to horses that were either favourite or second favourite would have yielded 46 winners from 142 rides (SR 32.4%) for an SP profit of £31.63 (ROI +22.3%). To BSP this edges up to +£47.88 (ROI +33.7%).
Danny Tudhope: Record by Trainer
Time to examine the trainer stats and below are all the trainers (still in business) for whom Tudhope has had at least 100 rides. They are ordered by strike rate.
Tudhope is David O’Meara’s stable jockey which explains the huge number of rides for the Yorkshire handler. He has a very good strike rate when riding for the William Haggas stable, but a good proportion of these rides have been at short prices. A couple of courses stand out with the Haggas / Tudhope combination: firstly they are 8 from 17 at Redcar, while at Newcastle they have secured 11 wins from just 23 runs.
He has profitable records to SP when riding for Archie Watson and Karl Burke. In fact, when riding for Watson, which he has done between 2017 and 2022, five of those six years produced a profit to BSP.
There are three trainers that should be mentioned, although none made the above table due to not having enough rides to qualify. They are Kevin Ryan, James Bethell and Sir Michael Stoute. Tudhope has a good record with all three both from a strike rate and a returns perspective.
Danny Tudhope: Record by Run Style
Onto run style now. I have already shared the fact that Tudhope has an excellent front running record on horses that race between 5f and 1m. Here is a breakdown of his run style performance in terms of percentage of runners that match each one:
These figures are very similar to those you would find if you averaged out all jockeys in the weighing room. It is a shame he has not led early on more than 15.4% of all horses considering how effective he is from the front. Of course, that style doesn't suit all horses on all occasions.
Tudhope wins more often with front running horses than with prominent racers, which in turn out-perform midfield racers and those held up early. This is the normal pattern we see for virtually all jockeys on the flat.
Danny Tudhope: Record by Market Factors
As regular readers will know I am a big fan of looking at favourite run style data, too, as this eliminates any potential selection bias regarding ‘good horses at the front, bad ones at the back’. Here are the relative win strike rates for Tudhope horses that have started as the market leader in terms of the four main run styles:
Front running favourites perform much the best. They secured a profit of around 15p in the £ assuming your crystal ball could have accurately predicted that they would all front run as well as being favourite. Tudhope has won from the front on favourites at all distances so is clearly an excellent judge of pace when leading, regardless of distance.
He is also one of the better jockeys from off the pace, especially in races at beyond a mile. In longer distance races I would not be put off by a Tudhope runner that habitually is held up.
I will summarize Danny Tudhope main takeaways at the end of the article, but now it is time to look at Ben Curtis.
Ben Curtis Jockey Profile
Ben Curtis Overall Record
Here are the overall stats for Ben Curtis:
Curtis has a slightly lower win strike rate than Tudhope, but still highly respectable, around the one win in seven mark. The A/E index of 0.95 is close to ‘value’ and to BSP Curtis would have secured punters a 4p in the £ profit across all his 5796 rides, which is mightily impressive.
Ben Curtis: Record by Year
Here is a breakdown by both win, and win/placed (Each Way) percentage / Strike Rate (SR%):
There has been a clear uptick when comparing 2018-2022 data with that from 2015-2017. This has occurred both from a win and each way perspective. There is a reason for this, as the improvement coincided with getting better rides as a whole from 2018: we can see this when we look at the prices of his runners year by year, especially the shorter end of the market. Here is a graph looking at the percentage of Curtis's rides by year that have been on horses priced 9/2 or shorter.
As the graph indicates, in 2022 compared with 2015 he rode more than double the number of horses sent off at 9/2 or shorter (in terms of percentage of his rides). Riding shorter priced runners improves the strike rate and that has been the driving force in the more recent past. I have said before that, where possible, we cannot be dependent on just one type of statistic. The more data and info we have at our fingertips the better, especially when it helps us understand why certain stats look the way they do.
Ben Curtis: Record by Betting Odds / Price (SP)
A look now at the Profiler splits in terms of Industry Starting Price:
The shortest priced runners (odds on) have, amazingly, nudged into SP profit. That is unusual. The 9/1 to 14/1 bracket has also seen him out-perform the average, certainly in terms of strike rate, as we saw with Tudhope earlier. It looks like the very big-priced runners (28/1 or more) are worth avoiding though – just 5 winners from 553 with losses of just over 62 pence for every £1 staked.
Ben Curtis: Record by Distance
Time to see if there are any clear differences when we look at Ben's record at different distances. Normally these figures are very similar, but it is always worth checking just in case:
As with Tudhope the very longest distances have the lowest strike rate, but again the sample size is smaller than the other categories – 197 races in the 1m7f+ sample. I would say Curtis has no major strengths or weaknesses when it comes to riding at different distances.
Ben Curtis: Record by Racecourse
Let's take a look at ‘Curtis by course’ – as before the courses are listed alphabetically and the minimum number of rides to qualify is 150:
Seven of the 19 courses have produced a profit to SP, with the Carlisle stats leaping off the page. At Carlisle Curtis has secured comfortably the highest strike rate compared to other courses, likewise the A/E index is the highest of all courses as is the PRB figure. Profits are extremely high, but we need to dig a bit deeper to see how many big-priced winners have affected the bottom line.
The biggest priced winner for Curtis at Carlisle has actually only been 25/1 so that makes these figures even more impressive. Below is the Carlisle breakdown by year, which is always useful to review for consistency:
Probably two things stand out initially. Firstly the eye is drawn to the poor performance in 2019 and, secondly, the 2015 profit figure accounts for over half of the eight year bottom line. Dealing with the poor 2019 – this is bound to happen when examining course/jockey stats. If you look at the PRB figure for that year it was similar to four of the six other years, so things are not as bad as they look at first glance. Also when delving in more detail into 2019, it emerges that Curtis rode eight horses at the course which finished second including at some reasonable prices – 6/1, 7/1, 8/1. With smaller data samples these ‘poor’ years will happen. Statistical variance, luck, quality of rides will all play a part too.
In terms of 2015 providing more than half the profit, it should be noted that four of the other six seasons made a profit, and decent profits at that. 2022 was a losing year, but his placed strike rate was actually the highest of any of the seasons (59%) so again perhaps not as ‘bad’ a year as the raw stats suggest.
All in all it is clear that Curtis rides Carlisle very well and, for the record, he has won for 24 different trainers at the course, so he is not reliant on a single handler, like so many jockeys are. He has also won for 30 different trainers at Beverley (57 winners).
Ben Curtis: Record by Trainer
That leads us nicely onto Curtis's performance for different trainers now. Below are all the trainers (still operating) for whom Curtis has had at least 100 rides. They are ordered by strike rate.
* including singular trainer name entities at the same yard
The combination with William Haggas is extremely good, although Curtis has only had nine rides for the stable in the past two seasons. Curtis has produced a profit to SP with horses from the top three in the betting for five of the trainers in the table; namely Haggas, Boughey, Palmer, Ellison and Appleby. He has only started riding for Boughey in the past three seasons but it is worth noting that on horses priced 6/1 or shorter the combo has produced a highly impressive 38 winners from 81 runners (SR 38.8%) for an SP profit of £22.10 (ROI +27.6%). To BSP returns increase to +36%. He has ridden a lot for Karl Burke in the past, but last year saw him have just six rides so it is not a combination that is going to produce many more runners it seems.
Before moving away from trainers, Curtis also has an excellent record when riding for two other trainers – for Charlie Hills he is 13 wins from 44 (SR 29.6%) and for George Scott 16 wins from 59 (SR 27.1%).
Ben Curtis: Record by Run Style
Onto run style now, and here is a breakdown of Curtis’s run style in terms of percentage of runners that match each one:
It is good to see he is above the average figure for ALL jockeys when it comes to front runners and also he is below the average for hold up horses. This to me suggests that he has some appreciation about the value of early track position.
Onto his win record on favourites in terms of run style:
Curtis has a slightly higher SR% figure on prominent favourites to the norm, but the general pattern is seen once more – there is such a simple answer to the question, ‘Would you prefer to be on a front running favourite or a held up favourite?’
It should also be shared that Curtis's front-running favourites were profitable to SP, as were the prominent racing favourites.
Danny Tudhope and Ben Curtis: Main Takeaways
The table below summarises the key takeaways for these two giants of the northern circuit:
So there you have it – two for the price of one!
I hope this article has uncovered a few more angles that may prove profitable for you to deploy over the coming months.
- DR
https://www.geegeez.co.uk/wp-content/uploads/2023/06/DannyTudhope_BenCurtis.png319830Dave Renhamhttps://www.geegeez.co.uk/wp-content/uploads/2022/10/geegeez_banner_new_170x78.pngDave Renham2023-06-12 14:59:052023-06-12 14:59:05Jockey Profiles: Danny Tudhope and Ben Curtis
The second in my series of articles on jockeys and, this time, Ryan Moore comes under the microscope.
Ryan Moore Introduction
Ryan Moore was born in Brighton in 1983 and he rode his first winner in the year 2000. Three years later, he broke through the 50 winners in a year barrier and, in 2004, he notched up his first century (132). In his early career he rode primarily for Richard Hannon but, by the mid-2000s, Moore was getting an increasing number of rides for Sir Michael Stoute. It was for Sir Michael that he recorded his first Group 1 success with Notnowcato in the Juddmonte Stakes at York in August 2006. In 2011 he started being noticed by Aidan O’Brien and, by 2016, he had ridden over 100 times in a season for the Irish maestro in the UK and Ireland combined. The Coolmore Stud provided the vast majority of these rides from the Ballydoyle handler, giving Moore the opportunity to ride some of the very top horses in training. In 2017 he secured his 2000th British winner and Moore is a definitely a jockey who justifies a deep dive into his statistical performance.
As with the Hollie Doyle piece I have analysed the last eight full years of flat racing in the UK and Ireland (2015-2023). I have used the Profiler Tool along with the Query Tool as the main vehicles for my data gathering. In all the tables profit/loss quoted is to Industry SP, but I will quote Betfair SP where appropriate.
Ryan Moore: Overall Record
Let's first look at Moore’s overall stats by reviewing his performance on every single runner during this eight-year period:
An excellent strike rate for Moore, in excess of one win in every five, primarily due to the fact that a sizable percentage of his rides are on fancied runners at shorter prices. This market detail also partly accounts for the fact that the PRB figure is very high at 0.63. His A/E index, a ratio that essentially determines value, is around the average for all jockeys.
We can also see that backing all his rides blind would have secured losses of nearly 21p in the £ to SP; to BSP the returns improve, but we still would have lost around 12p for every £1 staked.
Ryan Moore: UK v Ireland
It is relevant to distinguish performance in the UK versus Ireland for Moore because there is a quite a difference:
As can be seen, Moore's record in Ireland is far superior in terms of win percentage. This is mainly due to the fact that, in Ireland, 93% of his rides have been for Aidan O’Brien, whereas in the UK this combo stands at just 17% of total rides. O’Brien runners are rarely big prices so as a result of this one would expect to see that high strike rate for Moore in Ireland. However, perhaps what is more significant is if we look at the data for horses from the top three in the betting, comparing Ryan's record in the UK with his record in Ireland.
We are now comparing like for like from a betting market perspective. And yet still we see a stronger performance in Ireland and a much higher strike rate, as well as significantly better returns and a stronger A/E index. It should be noted we get a similar set of results if using a price bracket of say 5/1 or less. Already I am thinking Moore riding in Ireland is something to keep an eye on.
Ryan Moore: Record by Year
Annual data are the next port of call. Here is a breakdown by win percentage / Strike Rate (SR%):
Six of the eight years have seen a strike rate of over 20%; 2019 and 2020 were the years to dip below that figure. One obvious reason that may help explain this lower level was that Aidan O’Brien slightly under-performed at the same time. Obviously that would have affected Moore’s record as he rides so regularly for the stable. Moreover, 2020 was Covid-affected with Moore largely unable to ride in Ireland: he had just 15 rides, across Irish Champions Weekend, with two wins and another five placed horses.
If we track the yearly strike rates of both trainer and jockey we can see there is a clear correlation:
As punters we need to appreciate that in most cases jockeys are only as good as the horses they are riding, and those primarily riding for top stables will win more often than jockeys who ride regularly for ‘lesser’ stables. This is why when researchers drill into data they often use price bands to compare in order to offer a fairer comparison (like I did earlier in the UK v Ireland – top three in the betting stats). Talking of price, let's look at this area next:
Ryan Moore: Record by Betting Odds / Price (SP)
The Profiler offers a breakdown of performance by Starting Price splitting the market into seven price brackets. I have taken Moore’s record straight from that table:
As can be seen, Moore does not ride many genuine outsiders – less than 50 rides on horses priced 28/1 or bigger in the last eight years. From the table, then, it looks sensible to concentrate on horses priced 17/2 or shorter. When using BSP with these shorter priced runners one would have lost only around 6p in the £ across 3549 qualifiers. That's not too bad given the huge sample. In fact we would have made a small profit to BSP last year (2022) on horses with an industry SP of 17/2 or shorter. Hindsight, eh?
One clear problem with jockeys as well renowned as Moore is securing value. How easy is it to obtain value on a Moore mount? Clearly it is not easy, so we need to keep digging!
Ryan Moore: Record by Distance
A look at Ryan's record at different distances now. I have grouped them into five distance bands. Again I am comparing strike rates:
The one distance bracket that stands out from a strike rate perspective is 1m1f to 1m3f. The data sample is considerable so one would guess there is something going on here. But what could be happening? The first point to clarify is there is not a field size-related bias, even if 7f-1m races have a slightly bigger average field size than other distances.
One factor could be that Moore rarely blasts his runners out of the gates and hence tends to front run in races less than the average jockey. With that in mind, this might be what is hindering his strike rate figures at shorter distances (less than a mile). Over longer distances the front running bias declines considerably and hence in 1m1f to 1m3f this is not such an issue. That is one plausible idea.
Another theory is linked to the fact he rides many of the best bred middle distance horses in the world, usually for O'Brien / Coolmore Stud. Indeed if you look at the distance stats for Moore when riding for O’Brien, the best distance range for the pair is also 1m1f to 1m3f – hitting close to a 31% success rate. Backing this combo over these distances would have yielded a BSP profit of over 15p in the £. This theory, which initially had plenty of logic to it, now has some evidence to give it 'real world' credibility.
My final word on this distance section is simply that Moore may just judge the pace of these 1m1f-1m3f races better than any other distance. That may also have some validity.
Ryan Moore: Record by Course
I am now going to look at all courses where Moore has had at least 75 rides in the eight year sample period. The courses are listed alphabetically:
As one might expect, achieving blind profits at individual courses is unlikely, but Moore has snuck into SP profit at Chelmsford and Sandown. Using BSP actually does not change things too much with only Naas additionally edging into profit and Lingfield hitting break even.
Moore's record at Goodwood offers up some interesting stats when we compare his data on favourites with other market ranks:
The ‘not favourite’ stats include plenty of runners that were near the head of the market – combining second and third favourites produced just 6 winners from 73! Goodwood obviously hosts highly competitive racing so we do have to factor that in when noting poor or modest looking results. But perhaps a crucial note is that Aidan O'Brien doesn't really target the Glorious Goodwood festival like he does other meetings. Indeed, of the 16 tracks where O'Brien has saddled 20+ runners in the months of July and August, Goodwood has the lowest each way strike rate of all. Moore rode 55 of APOB's 80 such runners during the study period.
Considering Grade 1 UK courses more broadly, punters need to be cautious when focusing strongly on one particular jockey. For example, I think the following table is quite an eye opener. It compares Moore riding favourites at Grade 1 UK tracks with favourites at non-Grade 1 UK tracks. The Grade 1 UK tracks are Ascot, Doncaster, Epsom, Goodwood, Newbury, Newmarket, Sandown and York:
It should be noted that the average price of the favourites at the UK Grade 1 tracks was higher, which will have a bearing on the strike rate, but even taking that into account the numbers are still poles apart. I did check horses priced 2/1 or shorter across both types of track and the non-Grade 1 UK courses secured an 11% better strike rate then as well and much better returns of an extra 19p in the £. I rarely back favourites myself, but if there are favourite backers out there, bear those stats in mind if looking to back a Moore 'jolly'.
Before moving away from courses, the stats from these five courses where Moore did not ride at least 75 runners are actually worth sharing:
The sample sizes are not that small and the two stand out stats are the PRB figures for Wolves (0.84) and Navan (0.80) – these are exceptionally high.
Ryan Moore: Record by Trainer
Here are the trainers that Moore has ridden for at least 50 times (ordered by strike rate) – there are 11 in total:
* includes prior trainer entities at the same establishment
Moore has a very good record when riding for the Charlton stable, especially with horses from the top three in the betting – with these runners his figures read 21 wins from 54 (SR 38.9%) for an SP profit of £34.03 (ROI +63.0%). William Haggas and Charlie Hills are also trainers that Moore has done well for and, as a general rule, when the jockey teams up with either of these trainers I would look at it as a positive.
As expected Aidan O'Brien and Sir Michael Stoute provide Moore with the vast majority of his rides, with O'Brien offering better stats in that particular battle.
We saw earlier that the overall Ireland versus UK stats differed markedly for Moore. It makes sense therefore to compare Moore’s record with O'Brien when riding in the UK compared with Ireland. The graph below plots the relative win and win/placed (each way) strike rates:
We can see a much stronger set of results for Irish races in terms of wins and places. This was to be expected, with there being a heavy selection bias when Moore catches a plane to ride, but it is still nice to see that confirmed. Losses to level stakes correlated with the strike rates meaning they were much steeper in the UK than in Ireland for this jockey trainer combination - 16.5% in the UK, 5.8% in Ireland. This equates to a difference of nearly 11 pence in the £.
Ryan Moore: Record by Run Style
Onto run style now. Here is a breakdown of Moore’s run style in terms of percentage of runners that match each of the four styles measured on geegeez.co.uk:
These figures are very similar to those you would find if you averaged out all the jockeys in the weighing room. Ryan has raced from the front on 14% of his rides which equates to roughly one in every seven. However, there is a big difference if we compare the percentage of Moore front runners in handicaps to non-handicaps. In handicaps he has taken the early lead in just 9.7% of races, in non-handicaps the figure is 16.7%.
In sprint handicaps (5-6f) Moore has led early just 20 times in 264 races, which equates to just 7.6% of the time. This stat does baffle me. As regular readers will know, front runners in sprint handicaps generally have a huge edge. Moore clearly does not think like this – if he did that figure would be much much higher.
Moore follows the usual trend of jockeys where his front runners win more often than his prominent racers who in turn out-perform mid div and those held up early. I always look at favourite run style data, too, as this eliminates any potential selection bias regarding 'good horses at the front, bad ones at the back'. Here are the relative win strike rates for Moore-ridden favourites in terms of the four main run styles:
Over half of his front-running favourites went onto win. It should come as no surprise therefore that one would have made a healthy profit on Moore-ridden front-running favourites, while significant losses were incurred on favourites that were held up or raced midfield early. Moore on Aidan O'Brien-trained front-running favourites have an astonishing record: 60 wins from 94 runners (SR 63.8%). If your crystal ball had predicted these runners pre-race, you would been able to secure a huge profit of £52.36 (ROI +55.7%).
Ryan "More": Extra stats and nuggets
With the main body of the article complete allow me to share a few extra statistics that may be of interest:
When riding a horse making its debut in the UK, Moore has won 44 times from 333 runs (SR 13.2%) for significant losses of £143.36 (ROI -43.1%). Even when these debutants have started favourite such runners made losses of around 29p in the £. Compare this to Irish debutants who have won over 25% of the time (23 wins from 90). This is another example of the O'Brien factor.
Keep an eye on horses that are having their second career start where Moore was also on board for their debut. This cohort has produced 39 winners from 111 (SR 35.1%) for a small SP profit of £3.27 (ROI +3.0%). To BSP this improves to +£18.73 (ROI +16.9%).
Moore has a better strike rate at Royal Ascot compared with all other Ascot meetings combined. At Royal Ascot his strike rate is 18.6%; all other Ascot meetings combined this figure is just 12.7%. At Royal Ascot (2015-2022) backing Moore blind would have yielded a BSP profit of £44.91 (ROI +18.2%).
Ryan Moore Main Takeaways
Moore has a much higher strike rate in Ireland than in the UK (the O'Brien factor).
Moore's form is heavily influenced by the form of the Aidan O'Brien stable, especially when racing over the Irish Sea.
Moore has excelled at middle distances of 1m1f to 1m3f for all trainers, but especially so for O'Brien.
At Grade 1 UK tracks it is difficult to find value when Moore is riding.
Away from Grade 1 UK tracks Moore has made a small profit on all rides sent off favourite.
He has an excellent record at both Navan and Wolverhampton (samples are modest but the PRB figures are insane).
He has a very good record when riding for the Charlton stable, especially if they are in the top three of the betting. Charles Hills and William Haggas are trainers for whom he has solid records also.
Moore has an outstanding record on front runners that start favourite. This is especially true if trained by O'Brien.
The three extra nuggets shared immediately above.
*
So that wraps up my Ryan Moore profile. There is clearly no doubting Moore's qualities as a jockey – from a personal point of view, I just wish he would race close to or up with pace more often, especially in races of a mile or less. Given his superstar profile it is difficult but, as I hope you've discovered, not impossible to squeeze some juice out of Ryan Moore's value lemon.
Until next time...
- DR
https://www.geegeez.co.uk/wp-content/uploads/2021/09/StMarksBasilica_IrishChampionsStakes2021.jpg320826Dave Renhamhttps://www.geegeez.co.uk/wp-content/uploads/2022/10/geegeez_banner_new_170x78.pngDave Renham2023-06-07 08:35:262023-06-12 15:00:17Jockey Profiles: Ryan Moore
geegeez.co.uk uses cookies to improve your experience. We assume that's OK, but you may opt-out from the settings. Cookie settingsACCEPT
Privacy & Cookies Policy
Privacy Overview
This website uses cookies to improve your experience while you navigate through the website. Out of these cookies, the cookies that are categorized as necessary are stored on your browser as they are essential for the working of basic functionalities of the website. We also use third-party cookies that help us analyze and understand how you use this website. These cookies will be stored in your browser only with your consent. You also have the option to opt-out of these cookies. But opting out of some of these cookies may have an effect on your browsing experience.
Necessary cookies are absolutely essential for the website to function properly. This category only includes cookies that ensures basic functionalities and security features of the website. These cookies do not store any personal information.
Any cookies that may not be particularly necessary for the website to function and is used specifically to collect user personal data via analytics, ads, other embedded contents are termed as non-necessary cookies. It is mandatory to procure user consent prior to running these cookies on your website.
