The draw and potential draw biases is where my interest in horse racing began, writes Dave Renham. Back in the late 1990's I remember reading some excellent draw articles by Russell Clarke in a magazine called Odds On and I was hooked. Within days I was doing my own research using my Superform Annuals and pen and paper. This progressed to putting data into computers using excel.
I dread to think how much time I spent collating data. My main memory is working on my computer from 10pm to 2am on a regular basis. However, in those days the hard work was worth it because it was still a very under-researched area and draw biases were quite strong at certain courses. In addition to that, it was at a time before racing computer programs were commercially available.
It is over 20 years since I wrote my first book on draw bias and how things have changed since those ‘good old days’. At this juncture, it needs to be pointed out that many of the draw biases that were around 15 or 20 years ago are either not as strong as they were, or have disappeared completely. For many years draw biases provided punters with money spinning opportunities, me included. Virtually all my decent winning bets from around 1997 to 2006 were influenced by the draw in some way.
However, as with most things, when a good source of highlighting winners is found, within a few years the edge starts to disappear. This is very much a horse racing trait - good ideas gain an initial edge because the majority of people do not use that winner finding approach. As time goes on however, the betting public and the bookmakers catch up, and as a result the prices tend to contract and the value begins to disappear. This has happened with the draw, and to confound the problem course officials started using other means of negating potential draw bias. Running rails are now moved in order to keep horses off the fastest strip of ground, and better watering and drainage systems mean that most straight courses are far more even than they were back then.
The draw has had massive exposure in the past, and with people realising the edge is disappearing, the subject is beginning to assume less importance. However, before we begin to write off the draw completely, I still believe there is an edge for the educated draw punter. I maintain that at certain tracks a poor draw can still all but wipe out the chance of a horse, while a good draw increases one’s chances considerably. The trick perhaps is to find biases that may be more subtle, or at least which most punters are less aware of.
During this period of racing inactivity I plan to look at a few individual courses in depth, focusing primarily on draw bias but looking at pace aspects as well. The first course that will be put under the microscope is Pontefract.
Pontefract is located in West Yorkshire and is a left-handed track that is undulating with a stiff uphill finish in the home straight. Indeed the lowest point on the track is around the six-furlong start while the finishing post is the highest point, meaning both the five- and six-furlong sprints are testing.
The course is around two miles in length and, something I didn’t realise, is that originally it was around four furlongs shorter. Being left-handed one would assume that lower draws may have the advantage over high drawn horses at some distances, but the proof of the pudding, as always, will be in the eating!
For this article I am using key tools on Geegeez: namely the Draw Analyser, Pace Analyser and Query Tool. The period of study is a long one – going back to 2009, but I will examine more recent data in detail too.
My draw research has always focused on handicap races only. My belief is that handicap races give a better and fairer data set as such races are generally competitive affairs. When analysing each handicap race, I divide the draw into thirds - those drawn in the bottom third (low), those drawn in the middle third, and those drawn in the top third.
It should also be noted that I also adjust the draw positions when there are non runners – for example if the horse drawn 3 is a non runner, then the horse drawn 4 becomes drawn 3, draw 5 becomes 4 and so on. On a completely fair course the winning percentages for each "third" of the draw should be around 33% each. The differences in the percentages will help to determine the strength of the bias. The good news is that the Draw Analyser on Geegeez makes exactly the same splits, and is also capable of calculating draw by the advertised stall in your racecard and the actual stall, accounting for non-runners.
In my experience, I consider there to be two types of draw bias. Firstly, clear bias towards one specific section of the draw; this is the strongest possible bias. Secondly, one can get a bias against one specific section of the draw.
Another key factor to take into account is field size: for potential draw bias to exist I maintain there needs to be a reasonable number of runners in the race, and eight or more runners is the figure I have chosen. Draw bias is far more likely to be prevalent in larger fields as horses will either be forced to run wide (hence having further to travel), or be forced to run on a part of the track where the ground may be slightly slower. If the data set is big enough I will look at bigger field data where I feel it is appropriate.
OK time to crunch some numbers.
Pontefract 5 furlongs (8+ runner handicaps)
There have been 89 qualifying races - five-furlong handicaps with eight or more runners - during the period of study. Here are the overall draw splits:
Despite the track being left handed and the 5f distance having a bend to run round, low drawn horses do not dominate. The A/E values below suggest that the low drawn horses are overbet and are essentially poor value:
For the record, if you had bet every horse from the bottom third of the draw at £1 per bet you would have lost £136.34; backing all middle draws would have lost just £9.62 at starting price.
In the following table individual draw positions have been broken down for 5f 8+ runner handicaps at Ponte:
A few individual stalls made a profit but clearly there is no pattern to this so I would not be advocating backing certain draws in the future.
Field size seems to make no difference in the draw figures, but I was keen to look at whether the going made a difference. Back in the late 1990s and early 2000s when the going got testing in sprint races at Pontefract, horses tended to head towards the near rail in the straight giving higher draws an edge. Unfortunately for the minimum distance we only have 15 handicap races that have occurred on soft or heavy going; but, interestingly, lower draws have won 9 of the 15 (66.66%). That's far too small a sample from which to make any concrete conclusions; however, the 6f stats may give us more data to work with and may hopefully will show correlation.
Regarding 5f soft or heavy ground runners, you would make a very small profit backing lower drawn horses each way (£3.03 to £1 level stakes).
Let us look at pace and running style now. Here are the overall figures:
An notable edge for front runners can be observed. Moreover, better than 52% of horses that took the early lead went on to finish in the first three. This implies a strong front running bias.
On good ground or firmer the front running bias gets even stronger – early leaders win 20.48% of these races with an IV of 2.15. On good to soft or softer, conversely, front runners have failed to win any of the 22 races. It will be interesting to see if a similar pattern emerges over 6f.
Lastly for the five-furlong range, a look at draw / pace (running style) combinations for front runners in these 5f races:
Due to the left handed nature of the course/distance one might have expected more leaders to have come from the lowest draws. Interestingly, though, those horses that led from the bottom third of the draw (low) only managed to win three races from 39 attempts (SR 7.69%); A/E 0.51.
Horses that led early from middle draws went on to win over 25% of the time giving a positive A/E of 2.66. One additional stat is worth sharing: horses drawn in the bottom third of the draw (low) that were held up early have a dreadful record, winning just 2 races from 98 with an A/E of just 0.17.
Pontefract 5f Handicaps (8+ runners) Summary
The draw seems to be fair with no bias, while from a pace perspective front runners do have an edge.
Early pace is generally far more material than stall position.
Horses held up from a low draw have a terrible record.
Pontefract 6 furlongs (8+ runner handicaps)
There have been 153 qualifying races over six furlongs during the period of study. Here are the overall draw splits:
There seems to be a small advantage for lower draws here. It may not be hugely significant but is worth further investigation. The A/E values correlate to a certain extent as shown below:
A look again at individual draw positions and how they have fared over time:
Stalls 1 to 3 have decent individual A/E values and stall 2 has secured a long term profit. However, backing this draw blind in the future looks a less than robust way to produce a profit. I would be encouraged, however, if a horse I fancied was drawn in the bottom three stalls – this would be an extra tick in the box as it were.
This graph, which shows IV3 (the average Impact Value of a stall and its closest neighbours, e.g. 456), helps to visualise the table above from a 'likelihood of winning' perspective:
Looking at field size, low draws have the strongest edge in smaller fields (races of 8 or 9 runners). There have been a decent number of these races – 62 in total. The draw split for winners as follows:
The A/E value for low drawn horses edges up to 1.06 here. It seems therefore that a lower draw is more preferable in smaller fields. It is nothing to go ‘crazy’ about but a lower draw under these circumstances does look preferable.
What about the impact of the going in Ponte handicaps over six furlongs? It was noted above that, on soft or heavy ground in 5f handicaps, low draws seemed to have an edge albeit from limited data. In handicaps over a furlong further, the soft or heavy draw stats look as follows:
Again this data set is quite small (21 races), but a look at the win and placed data - table below - strongly suggests a lower draw is preferable:
For the record, backing all low-drawn horses EACH WAY on soft or heavy ground would have secured a profit of £19.57 to £1 level stakes.
Next follows a table illustrating the effect of pace and running style:
An edge for front runners again, while hold up horses have a relatively moderate record. When looking at 5f races earlier it was noted that front runners did better on firmer going and had struggled in testing ground. Unfortunately, from a statistical point of view at least, the complete reverse is the case here with front runners having performed far better on testing ground: indeed from the limited sample they have won over three times more than would be expected statistically. So one potential theory goes out of the window!
Again, we'll close out the distance review with a look at draw / pace (running style) combinations specifically for front runners in 6f handicaps:
As with the 5f range, horses which are drawn high are less likely to get to the early lead - in this case approximately half as likely as those drawn middle or low. There is little to choose between low and middle drawn horses in terms of getting to the early lead.
However, it should be noted that higher drawn horses that got to the lead have managed to go on to win almost 20% of the time.
Pontefract 6f Handicaps (8+ runners) Summary
To conclude, the 6f trip seems to offer low drawn horses an advantage which appears to increase in smaller fields.
The bias towards lower draws has been stronger on softer ground where, conversely, higher draws have struggled more.
Pontefract 1 mile (8+ runner handicaps)
Most people focus their draw attentions at sprint distances, but a mile for me has always been the key distance at Pontefract in terms of the draw. From my previous research, lower draws traditionally had a decent edge over a mile so let’s look at the current data. There have been 142 qualifying races which gives us a really good chunk of information:
As expected the low draw bias is strong, with the A/E values not surprisingly following a similar pattern:
And here is the performance of each individual draw since 2009:
Draw 2, as it did over 5f and 6f, shows a blind profit. The A/E values for draws 1 to 3 are good as one would expect. This table does show quite neatly the draw bias in operation – several columns show this such as the win% column, the ew % column and the A/E column.
Once more, the IV3 chart brings the point home:
As this mile trip indicates a strong bias it is worthwhile checking a more recent subset of the data to confirm the long-term perspective. Focusing on the last four seasons (2016 to 2019), during which time span there were 54 races, gives the following splits:
These are similar results albeit a slightly lower win percentage for the bottom third of the draw. However, it ratifies the bias which has been around for years remains alive and kicking.
A solid footnote is that in the past four seasons 23 of the 54 mile handicap races with eight or more runners were won by horses drawn 1 or 2 (SR 42.6%). Compare this with just eight wins achieved by the two highest drawn horses.
In addition, for those who like ‘exotic’ bets, you would have made a small profit if you had permed the lowest two drawn horses in every race in £1 reverse exactas: £14 profit from a £108 outlay. Of course an exacta is a pool bet so it is difficult to exploit potential draw biases in this way as such ideas, if overbet, would contract the returns. Having said that I have personally had much success in the past perming certain draws at certain tracks.
Back to the complete data set (going back to 2009) and a look at mile handicaps by number of runners - specifically looking at fields of 8 or 9 runners - there have been 53 races with the following draw splits:
A stronger bias it seems for lower drawn horses in small fields. The A/E values back this up as is shown below:
There also is a strengthening of the bias in bigger fields albeit from a relatively small sample. In races of 14 runners or more, 19 of the 30 races (SR 63.3%) have been won by the bottom (low) third of the draw.
Turning attention to the state of the turf, the win percentages for low drawn runners are extremely uniform and I have found nothing of note there.
However, with regard to pace and running styles, there are some factors to keep in mind. Here are the overall stats:
In racing in general, as the race distance increase so front running biases start to diminish. However, at Pontefract there is a stronger front running bias over a mile than at 6 furlongs. I found nothing of interest when delving into going considerations and field size, so nothing extra to report there.
Finally over this mile trip this is how the draw / pace (running style) combinations look for front runners in 1 mile handicaps:
These stats demonstrate that it is much easier - or at least more common - for a horse to lead from a low draw over a mile at Pontefract. Having said that, high drawn early leaders have gone on to win slightly more often in percentage terms. Horses that race mid division or are held up when drawn in the top third of the draw (high) have won just 7 races from 285 runners.
Geegeez Draw Analyser has a heat map to help visualise this, here displaying IV:
Pontefract 1 Mile Handicaps (8+ runners) Summary
The mile trip at Pontefract shows a significant draw bias to lower drawn horses. It is one of the strongest mile biases in the UK, if not the strongest.
From a pace angle, it is preferable for a horse to lead or track the pace.
Pontefract 1 mile 2 furlongs (8+ runner handicaps)
The final distance to be examined in this article is a mile and a quarter. The configuration of the track means that there is an extra bend at this distance as compared to the mile trip and hence one would expect low draws to again have a decent edge. There have been 107 qualifying races from which to find angles:
On first view this looks a very strong bias with lower draws dominating and higher draws seemingly at even more of a disadvantage than they were at a mile. The A/E values back up the raw win percentages as a measure of profitability:
Indeed backing every horse drawn in the lowest third over ten furlongs at Pontefract (8+ runner handicaps) would have returned £39.90 to a £1 level stake.
Individual draw data next, and can stall 2 make a blind profit yet again??!!
Yes! Stall 2 has made a blind profit again - meaning it has been profitable at every individual distance up to 1m2f - as have stalls 3 and 4. Again, this table helps one visualise the strength of the low draw bias. Would I consider backing draws 1 to 4 ‘blind’ in the future? No, but it is clear that these draws must be the primary focus when analysing these races. Here is the IV3 chart to bring that home:
Time to check out more recent data to see whether the bias has been as strong over the past four seasons (2016-2019). There have been 33 qualifying races during that time, giving these stats:
Whilst it is not quite as strong, that could simply be down to the smaller - less reliable - sample size. It still indicates that low draws have a substantial advantage over higher ones.
Moving back to the complete data set (2009-2019) the low draw bias seems to strengthen considerably as the field size grows. This makes sense as the extra bend potentially helps lower drawn runners and impedes higher drawn runners who have to race wider. In races of 12 runners or more, 20 of the 31 races (SR 64.52%) have been won by the bottom third of the draw (low). The A/E value stands at a very healthy 1.25.
Indeed moving the goalposts up further - to 13+ runners - low draws have totally dominated, winning a huge 17 of the 22 races (SR 77.27%). The A/E value for low drawn runners is an uber-impressive 1.53.
Looking at going data there is something which stands out albeit from a limited sample. Races on soft or heavy seems to increase the strength of the low draw bias. From 21 races 15 were won by a horse in the lowest drawn third of the field. That equates to over 70% and an A/E of 1.55. Of course with limited data one cannot be too dogmatic, but these figures are still highly promising.
A look at the pace / running styles figures next:
Front runners have a stronger edge than I had expected, winning twice as often as most other run styles: maybe that extra bend near the start helps.
And finally, the draw / pace (running style) combinations for front runners in 1m2f handicaps:
Lower drawn horses as expected lead more often and roughly four in seven of them go on to finish in the first three. High drawn horses tend to struggle when racing mid division or when held up. This was also the case over 1 mile as we saw; over 1m2f such runners have won only five races from 207 runners.
Pontefract 1m2f Handicaps (8+ runners) Summary
The 1 mile 2 furlong distance shows a similarly strong low draw bias to that at a mile, and it seems that bigger fields may accentuate this.
Soft or heavy going may also strengthen the bias but that notion is based on limited data and so a watching brief is recommended.
Fingers crossed, in the near future we will see race meetings start again at Pontefract and, when they do, I hope these stats will help point you in the right direction in the ‘fight’ against the bookmakers.