By David Renham April 2003

With the flat season just getting underway, I thought I would give you a look at short-term draw trends at a variety of courses. I wrote a similar piece 12 months ago, but this piece is more up to date and hence should be more accurate.

For simplicity I have split the draw into the top 40% and compared this with the bottom 40% of the draw (eg. in a 15 runner race, we would be comparing the low draws 1 to 6, with the high draws 10 to 15).

The results below are taken from handicap races with more than 10 runners between 1998 and 2002. The table is split into four columns - the first three are self-explanatory; the final column - "Advantage to top / bottom 40%" gives a figure to determine the strength of bias. For example, at Ascot over 5 & 6 furlongs, horses drawn the bottom 40% of the draw have won 1.04 more times than horses drawn in the top 40% of the draw. Hence at Ascot, there seems to be currently little or no draw bias. A figure of 2.00 or more would indicate significant draw bias.

Here is a selection of course and sprint distances I have studied. This is an extensive list, if not an exhaustive one :

Course and distance	Win % for all horses drawn in the bottom 40% of the draw	Win % for all horses drawn in the top 40% of the draw	Advantage to top / bottom 40%
Ascot 5 & 6f	6.12	5.90	1.04 bottom
Ayr 8f	11.89	3.25	3.66 bottom
Bath 5f 161 yds	6.11	7.38	1.21 top
Beverley 5f	1.18	12.63	10.70 top
Beverley 7f 100yds	5.59	8.58	1.53 top
Carlisle 5f	1.32	8.99	6.81 top
Catterick 5f	7.30	7.41	1.02 top
Catterick 7f	5.63	6.63	1.18 top
Chester 7f 122yds	8.39	1.89	4.44 bottom
Chester 10f	11.54	4.55	2.54 bottom
Doncaster 5 & 6f	3.97	6.90	1.74 top
Doncaster 8f (straight)	10.00	2.13	4.69 bottom
Epsom 7f	10.29	5.70	1.81 bottom
Goodwood 5 & 6f	3.28	7.58	2.31 top
Goodwood 7f	3.17	12.16	3.84 top
Hamilton 8f & 65yds	3.83	10.55	2.75 top
Haydock 5 & 6f	6.16	5.54	1.11 bottom
Kempton 5 & 6f	3.65	7.92	2.17 top
Lingfield 5f (turf)	1.79	9.23	5.16 top
Musselburgh 7f	4.17	9.41	2.26 top
Newbury 7f	7.05	6.15	1.15 bottom
Nottingham 5 & 6f	5.50	6.25	1.14 top
Nottingham 8f	7.73	6.09	1.27 bottom
Pontefract 5f	5.41	6.40	1.18 top
Redcar 5 & 6f	4.64	4.68	1.01 top
Salisbury 5 & 6f	5.56	7.56	1.36 top
Sandown 7f	5.52	7.69	1.39 top
Southwell 5f (aw)	7.94	6.58	1.21 bottom
Southwell 6f (aw)	6.71	7.19	1.07 top
Thirsk 7f	8.57	5.60	1.53 bottom
Warwick 7f	7.32	1.19	6.15 bottom
Wolverhampton 6f (aw)	6.45	8.73	1.35 top
York 5 & 6f	5.92	4.84	1.22 bottom
York 10f	8.43	3.74	2.25 bottom
York 12f	7.64	3.61	2.11 bottom

Some interesting findings - as you can see, draw bias even exists at 12 furlongs (see York stats). Indeed, York does show a low draw bias at 14 furlongs, but I have not included those figures. There are also a few surprises : for example, I had expected Southwell over 6f to show a significant low draw bias, but it does not. Indeed, high draws have a small edge according to the stats. Likewise, one would have expected Pontefract to have a low draw bias over 5 furlongs, but high have ended up showing a marginal edge. The other slight surprise was to see Newbury favour lower draws over the straight 7 furlongs.

Below I have listed the five most biased course and distances from the group I have looked at :

Course and distance	Advantage to top / bottom 40%
Beverley 5f	10.70 top
Carlisle 5f	6.81 top
Warwick 7f	6.15 bottom
Lingfield 5f (turf)	5.16 top
Doncaster 8f (straight)	4.69 bottom

From these five, the Beverley bias over 5 furlongs is no surprise to draw punters, as it is arguably the most draw biased course and distance in the country. However, I was surprised to see Carlisle (5 furlongs) and Warwick (7 furlongs) show such a strong bias. I think these are course and distances that we need to keep a close eye on in 2003. Lingfield sprint course is one that has often shown a high draw bias and clearly that has occurred over the past five seasons. The straight course at Doncaster is quite bizarre - over 8 furlongs there has been a strong low draw bias, while at 5 and 6 furlongs (on the same straight course), high numbers tend to dominate.

These findings illustrate that draw bias exists to a certain extent at many courses in Britain. However, although these findings are quite informative, we, as punters, need to be aware that draw bias can change and nothing is set in stone. Course officials are more aware than ever about trying to eliminate draw bias and so we need to keep abreast with any new developments. For example, at one meeting last season Beverley experimented by putting the stalls "low" over 5 furlongs, rather than their traditional position of "high". The draw bias was turned completely on its head with low draws clearly having the advantage over high draws. The race in question (the "Big Screen is Here Tonight Claiming Stakes" on July 22nd) saw Countrywide Dancer (drawn 1) win at 25-1, with Warren Place (drawn 2) second at 12-1. The forecast paid £306.13 to a £1 stake !!

What is clear, is that draw bias remains a key factor at several courses and 2003 will be no exception. I hope you, like me, are all looking forward to the new turf flat season. Let us hope draw bias will help us stay one step ahead of the bookies and the "crowd" this year.


I would like to thank David on behalf of Betting on Horse Racing for allowing us to use this article. He is seen as the primary expert on the effect of the draw. For more information about David and his work please visit his webiste Drawn To Win