My series on all weather tracks continues with the first of a two-parter looking at the Irish racecourse at Dundalk, writes Dave Renham. This is my first detailed piece of research on Dundalk and I am hoping to find some credible angles which will help us all when having a bet there. The track is 1m2f in circumference and left handed and as you can see in the picture below the 5f distance starts from a chute which joins the round course at the penultimate bend.

 

 

Dundalk was resurfaced in April 2020 and reopened in July of that year. So, for this article, the focus will be on races run from 12^th July 2020 (the first fixture after the renovations) to 31^st August 2022. I will at times be comparing the new data with past data to try and gauge whether there are any significant differences compared to the results from the old surface.

I have used the Geegeez toolkit - Query Tool, Pace and Draw Analysers - for all the data collection: these can be accessed from the home page from the Tools dropdown menu.

Running Style at Dundalk

For regular readers of my Geegeez articles, they will know that I believe the run style of horses at certain distances is important and is still an underused approach when taking into account all punters as a whole. For the new readers, I will briefly discuss what is meant by run style before doing some digging. In essence, run style is the position a horse takes up very early on in the race. These are split into the following four categories:

Led (4) – front runners; horse or horses that take an early lead; Prominent (3) – horses that track the pace close behind the leader(s); Mid Division (2) – horses that race mid pack; Held Up (1) – horses that race at, or near the back of the field early.

The number in brackets is the run style score that is assigned to each section. These numbers can be a useful tool for crunchers like myself as they can be used to create different numerical representations.

When analysing Dundalk run style, and indeed later when analysing the draw, I will be looking at individual distances – mainly the shorter ones with the main focus being 8+ runner handicaps. The shorter distances are more prone to ‘bias’ both in terms of run style and draw.

Dundalk Racecourse 5f Run Style Bias

Let's look at the shortest trip first. Below are the run style (pace) figures taken from the Query Tool  for the period since the resurfacing (12/7/20 to 31/8/22):

 

 

Early leaders / front runners have had a strong advantage it seems, albeit from a relatively modest sample. This is a pattern that we have seen over the minimum trip of 5f at several courses, both on the all weather and on turf, especially on turning tracks; so, despite the sample size, the chances are that this front running bias is likely to continue. The A/E (Actual vs Expected) indices and Impact Values (IV) also correlate strongly which gives more confidence to the findings*. Combining win and placed percentages (each way) also seems to confirm the bias:

*For more on A/E and IV, as well as PRB, see this post

 

 

Hold up horses have a dreadful record in handicaps over five furlongs at Dundalk: a front runner is roughly four times more likely to make the frame than any individual hold up horse.

I looked back at the 5f data before the resurfacing of the track, going back as far as I could to 2009, and the front running bias existed then, too. The win percentage of 16.1% and the win / placed percentage of 40.3% are slightly below the more recent figures posted by front runners, but they were still much stronger than any other run style category during that time frame.

My conclusion is that the surface change has not made it more difficult to win from the front in 5f handicaps, indeed it has perhaps made it easier. The next year or two will help to confirm or deny this early view.

Before moving onto to 6f handicaps, I decided to dig a bit deeper into the more recent 5f stats and I looked at the performance of front running handicappers in terms of percentage of rivals beaten (PRB). This is a measure to help with determining the strength of a bias, and it helps by awarding every runner in every qualifying race with a 'score' based on the number of rivals they beat. The winner will have a score of 1.00 (100% of rivals beaten) and the last placed horse will have a score of 0.00 (0% of rivals beaten). The fifth horse home in a nine-horse race, for instance, will have beaten four rivals (6th, 7th, 8th and 9th) and lost to four rivals (the first four home) for a PRB of 0.50 (50% of rivals beaten).

A score of 0.55 or higher is considered a good advantage, while 0.45 or lower can be viewed as a negative.

Getting back to Dundalk five furlong handicaps, I found that front runners over this trip had a PRB of 0.64. This is further proof of the advantage front runners have. I did the same calculations for hold up horses, too, and their PRB stood at a lowly 0.40.

-------------------

SIDEBAR: How I calculated PRB for this article

PRB figures can be found on Geegeez in different areas. For example if you click on the record of a horse you will see something like the following:

 

 

Here we have not just the win, placed, profit, prize money data, but also the PRB figures. In this case the turf PRB (0.62), all-weather PRB (0.52) and the overall PRB (0.57). You can get this PRB info also when clicking on the records of trainers, jockeys and sires.

PRB figures are also shared in the Draw Analyser:

 

 

In this example (taken from Chester) we see the PRB figures back up the win% data showing a strong low draw bias and, particularly, a negative bias for high-drawn horses.

The PRB figures shared earlier for front runners and hold up horses in 5f handicaps at Dundalk were not taken directly from a page on the site; rather, I manually calculated them. It was not difficult to do and I’ll now offer a quick explanation of how I calculated that front running figure of 0.64 (in case you want to examine this type of idea yourself).

Firstly, in the Query Tool I looked for all horses that had gained a 4 pace rating in 5f 8+ handicaps:

 

 

This generated the following results:

 

 

These are the figures seen at the start of this article is the very first table. From here I clicked on the qualifier tab, thus:

 

 

With 20 qualifiers per page, there were 3 pages of qualifiers in total. I copied each page of data and pasted into Microsoft Excel. I then created two new columns using a simple formula – ‘Rivals beaten’ and ‘Rivals Not Beaten’:

 

 

 

I then totalled up the Rivals Beaten column (286) and the Rivals Not Beaten column (161). To calculate the PRB figure you add together the two totals (447) and divide the Rivals Beaten number into this total (286 divided by 447). This gives us the 0.64 figure shown earlier.

I personally will do these run style PRB calculations for other courses now unless the data set is too big. Copying and pasting a few pages of qualifiers is relatively quick; however, doing it, say, for 100 pages is probably going above and beyond!

SIDEBAR ENDS

-------------------

 

I have also looked at the 5f non-handicaps stats – data is limited as there have only been 17 qualifying races (with 8+ runners). However, the win / win & placed (each way) stats suggest a strong front running edge here too:

 

 

All the stats I have shared do seem to point to a strong front running bias over 5f at Dundalk regardless of race type.

Below is a PRB heat map overlaying draw thirds and run styles for 8+ runner handicaps at Dundalk since 2020:

 

 

It is clear that low and led/prominent, or indeed almost any led/prominent position is hugely more beneficial than being waited with further back.

 

Dundalk Racecourse 6f Run Style Bias

Up a furlong now to six furlongs, and a look at the handicap run style stats over this distance since the resurfacing work; again 8+ runner races only to qualify:

 

 

At first glance these stats suggest that front runners have a slightly stronger edge over this extra furlong. The IV for front runners of 2.7 is high; as is the 1.79 A/E index.

Here are the win and placed percentages (each way):

 

 

There is a strong correlation here when comparing the each way stats with the win only stats – they paint the same picture.

Looking at the data from before the resurface (all the way back to 2009), front runners won just over 16% of races so the bias was evident but maybe not quite as strong. Hold up horses had a win percentage of 5% which is a bit below the more recent figure.

A look now at the PRB figures for front runners versus hold up horses since July 2020. Front runners have a huge PRB of 0.68, while hold up horses are down with a PRB of 0.43. These are similar to the 5f figures as one would probably expect. (These were calculated using Excel as outlined above).

It makes sense to share the non-handicap stats as I did over 5f. Here is a graphic of the win / win and placed (each way) run style percentages for non-handicap races over 6f:

 

 

In terms of win percentages, prominent racers have nudged ahead of front runners, but when we look at the win & placed (each way) figures, front runners lead once more. The strongest aspect of these data, it seems, is the quite dreadful win record for horses that race off the pace either in mid-division or held up: both run style win percentages loiter below 3%. This from 34 races giving us a small but probably satisfactory sample size. 

All the 6f handicap statistics point to the fact that front runners have a very strong edge. This advantage looks even more potent than over 5f. If this is the case, then perhaps it is down to the fact that over 6f the full turn comes into play; over 5f, because of the chute, there is effectively only a half turn (see first image at the top of this post). This is just conjecture, but it feels like a plausible suggestion. In non-handicaps, prominent racers and front runners combine to have a monster edge over those midfield and hold up horses in the first quarter mile or so.

When reviewing the draw / run style heat map we again see the difficulty slow starters face in getting competitive:

 

Dundalk Racecourse 7f Run Style Bias

Let's move on to 7f now (12/7/20 to 31/8/22):

 

 

Front runners are not the most successful group over this range; over seven furlongs, the accolade goes to prominent racers. There is still a run style bias in play here with runners close up or on the pace early having an edge over those further back. What is noticeable at most tracks is that the front-running win percentage tends to drop as the distance increases past the shortest races distances of five and six furlongs. This is mainly because the horses behind the early leader have more time to make their challenge and can eventually get to the front themselves.

Looking at the long term data going back to 2009 up to the time the course was resurfaced, front runners won 11.8% so they were slightly more successful during that period. However, a change of one percentage point is not statistically significant. It looks like the 7f run style picture has not really changed much since the renovation work.

7f non-handicaps have a similar spread of results since the resurfacing occurred – front runners have an edge with a 13% success rate; prominent racers 11%; midfield 9% and hold up horses were worse off again at just 4% (though there is some selection bias in the latter cohort due to horses of very limited ability, or those not yet ready or able to show their full ability, sometimes running in non-handicaps).

This time, the PRB heat map overlaying draw and run style is less conclusive, but it does suggest middle to high and not held up could be beneficial.

 

 

Dundalk Racecourse 1m Run Style Bias

The stats are similar over a mile with regards front runners – those most forwardly placed have won 11% of races in the past two years. Their A/E index stands at 1.13 and Win Impact Value at 1.48 suggesting a small edge still. Having said that, it is not a course/distance combination where I would looking to use run style bias as a key part in my selection process.

 

*

 

In conclusion, looking at run style as a whole across the shorter distances, front runners have a clear advantage over sprint trips (5f-6f) in handicaps. In non-handicaps this is true over 5f and to a lesser extent over 6f. Having said that, hold up horses and horses that race mid-division early really struggle over five and six furlongs regardless of race type. Over 7f (both handicaps and non-handicaps), horses that front run or race prominently have had the edge over horses that race mid-pack or at the back early. Once we get to 1 mile and further it becomes a relatively even playing field.

 

Draw Bias at Dundalk Racecourse

With the draw I like to compare recent time frames of a similar length. This is because draw patterns can (and do!) change at any time. Also, at Dundalk they have adjusted the rails on the home turn, putting in a false rail which potentially has ramifications for the draw. This change occurred more than five years ago so it makes sense not to include races before the false rail was introduced. Hence I will be looking at 8+ runner handicap draw data from July 12^th 2020 to August 31^st 2022, which is clearly the key data, but then comparing it with the prior two years before the resurfacing and after the false rail addition (going back to the start of 2018). Still with me? Good!

When I look at the draw my first port of call is to split the starting stall numbers into three roughly equal thirds and compare win percentages; percentages of around 33% across the board give us a completely level playing field in terms of the draw, deviations from that may imply an advantage or disadvantage.

 

Dundalk Racecourse 5f Draw Bias

Since the resurfacing in 2020 here are the draw splits over 5f (8+ runner handicaps):

 

 

There looks to be a clear edge to lower drawn horses here with those drawn very high struggling commensurately. Let us look now at the data in terms of percentage of rivals beaten (PRB).

 

 

This backs up the win% breakdown showing good correlation, though perhaps not as striking as in pure win percentage terms. Another snippet worth sharing is that horses drawn five or lower have won 69% of the races from just 44.5% of the total runners.

Time to compare the win percentages from this post-renovation period with data from the two years or so prior, specifically 1st January 2018 to 31st March 2020:

 

 

The results are very similar and the sample sizes are virtually identical. It could be that the new surface has slightly strengthened the low draw bias; it will be interesting to monitor the results of the next 12 months to see if this is the case.

As far as the draw is concerned I rarely look at non-handicap data unless I have a huge sample size. Here we have less than 20 races and so it is probably not a worthwhile proposition to dig into the numbers.

 

Dundalk Racecourse 6f Draw Bias

Below are the  6f  draw data by thirds:

 

 

There have been nearly 60 races so we have a decent sample size with which to work; low draws seem to have a strong edge, but my initial enthusiasm is tempered by the PRB figures:

 

 

This is a good example of why digging deeper into data is important and why we need to be cautious of small samples, even when they're relatively large in the context of horseracing analysis. The figures here suggest a slight lower to middle bias exists but nothing more. We need to be wary of that 50.8 win% for the bottom third of the draw, which is almost certainly inflated. The each way figures for each third are also virtually identical, which lends more credence to the PRB figures.

Looking at the data from 1^st Jan 2018 to 31^st March 2020, I would suggest a strong low draw bias was in existence as the win percentages and the PRB data correlate strongly this time. Win percentages first:

 

 

These are the equivalent PRB figures:

 

 

0.59 is a huge PRB figure for the lowest section of the draw. Hence my reading of the draw at Dundalk over 6f is that the new surface has some an effect in nullifying a previous low draw bias. Given the choice I’d still prefer to be drawn low, but it is marginal. Essentially I would not see the draw as a major factor over 6f at Dundalk.

 

Dundalk Racecourse 7f Draw Bias

Below are the 7f data and the draw splits since the course was resurfaced are as follows:

 

 

It's relatively even this time, although middle draws have won a shade more often than the other two sections. The PRB figures are as follows:

 

 

These figures correlate with the win stats suggesting if there is any edge, it lies with the middle draws. Previous to this (1/1/18 to 31/3/20), the PRB split had been similar with 0.54 for middle draws once again, but the 0.47 and 0.50 were reversed with low draws at 0.50 and high at 0.47. It seems therefore that a middle draw looks to have a small advantage over 7f but it does not look strong enough to easily exploit.

Other Dundalk Draw Observations / Conclusions

I have looked at draw data for a mile and beyond but the thirds are very even and essentially it looks like only the 5f distance is of real interest here in terms of a usable draw bias. At the minimum range, a low draw does seem to have a tangible edge and is something we should factor into our selection process.

If you, like me, are interested in exotic ‘draw based’ bets, note that a profit could been secured in 8+ runner 5f handicaps (12/7/20 to 31/8/22) combining the five lowest drawn horses in full cover tricasts: this would have provided four winning payouts and returned a profit of 52p in the £. It should be noted that a five-horse combination tricast/trifecta bet involves 60 bets so it can become a costly wager during a long losing run. Of course the beauty of tricasts is that payouts can be massive, which is why punters like me are lured into them!

*

That concludes part one of this Dundalk analysis and it’s time to research part two. My study to date suggests that, in terms of run style and draw, we have the following biases to work with this winter:

Handicap biases

5f – front running bias; low draw bias

6f – front running bias

 

Non-handicap biases

5f – front running bias;

6f – front runners/prominent racer bias

 

Dundalk, as a track, does not offer as wide a range of biases as, say, Chester, but over five and six furlongs the run style edge in particular should give us a leg up over the uninformed.

PART TWO OF THIS ARTICLE CAN BE FOUND HERE >

- DR