Even at the quietest of times, there are few more emotive topics than the relative merits of jockeys. When there has been such a high profile booking as Tony McCoy on such a high profile horse as Denman, the discussion gets ever louder and the fall of Denman in the Aon Chase has turned a lively discussion into a volatile debate. Tom Segal opened this discussion last Thursday with a piece in the Guardian suggesting that McCoy is not the greatest jump jockey of all time or even this time, because he has not won enough big races recently.

“For me, he doesn’t win enough big races any more. Yes, McCoy is a wonderful talent and clearly a very, very good jockey, but if I wanted someone to ride my horse in a big race, I’d go elsewhere.”

This bold statement served as a fascinating precursor to Denman’s fall on Saturday and led many Racing Post readers and Betfair forumites to sound a similar note after the race.

“Just because AP has won the most races doesn’t make him the best jockey. Ruby Walsh is streets ahead. AP is the man on the bad horses in the bad races. He is not in the class of others in the big races – Walsh, Geraghty, Lee”
tmorris1

“As expected, McCoy was not up to the task…”
georgiofan

“McCoy deserves to lose the ride on Denman…”
rebels26

Predictably though, the Racing Post “expert jury” leapt to the defence of the champion suggesting that McCoy should keep the ride on Denman.

“I don’t think McCoy was to blame for either error…”
James Pyman

“The 14-time champion did little wrong”…”he rode the perfect race”
Graeme Rodway

As ever, the statements are more an emotional reaction to what the speakers intuitively feel, rather than the the result of analysing the evidence. I shall attempt to look at the facts more objectively. The specifics of whether Denman or McCoy (or maybe even the trainer) were to blame for Saturday’s demise are beyond my analytical capability and will remain a subjective discussion as far as I can see, but the general point about Tony McCoy’s record in big races and whether he is a superior jockey overall to Ruby Walsh is something that lends itself to analysis of the data.

I will start in this article by looking at the overall record of the two jockeys and follow up later this week with a look at the big race records of the two men. There are a number of ways of looking at the issue and all have some merit. Firstly, we can simply look at the overall wins to runs record of the two men since 2003.

A P McCoy R Walsh
Wins 1,366 1,274
Runs 6,006 5,412
Win% 22.7% 23.5%

Table 1: Wins to runs

Overall, their records are very similar, though a historical look at the numbers does suggest that McCoy is get leaner pickings since the split from the Pipe stable, whilst Walsh is picking up progressively better rides. Either that or McCoy is getting worse as Walsh gets better.

year A P McCoy R Walsh
2003 25.1% 20.3%
2004 24.8% 22.5%
2005 21.6% 22.3%
2006 23.4% 24.7%
2007 23.0% 21.5%
2008 18.6% 26.1%
2009 23.4% 26.2%
2010 16.2% 32.7%
Grand Total 22.7% 23.5%

Table 2: Jockey strike rate over time

There is some interest in these tables but it is difficult to draw many firm conclusions. A lot depends on the quality of the rides they were getting and what chance they had in the race. One way of trying to eliminate this is to look at how many winners you would have expected the jockeys to ride given the price of the horses they were riding. So, for example, if a horse was even money then you might expect it to win 50% of the time. If McCoy were to ride six even money shots, you might expect that he would ride three winners. The expected number of wins is derived by simply adding up the percentage chances of the horses winning, which I will approximate by using 1 / starting price. It should be noted that this will overestimate the chance of the horse winning as there is a margin built in to these prices. When I have the Betfair SP available it will be a more accurate guide as there is essentially no margin built in to those prices (before commission is taken out). However, for now, I will use the industry starting price as the measure for this. Provided we use the same methodology for both protagonists, we should get a fair guide.

    A P McCoy
year Expected Wins Actual Wins %
2003 213.6 179 83.8%
2004 230.3 218 94.6%
2005 235.3 199 84.6%
2006 217.3 200 92.1%
2007 214.4 196 91.4%
2008 168.5 140 83.1%
2009 238.5 217 91.0%
2010 24.5 17 69.4%
Grand Total 1,542.6 1,366 88.6%

Table 3a: McCoy expected vs actual wins

    R Walsh
year Expected Wins Actual Wins %
2003 162.9 147 90.2%
2004 189.4 177 93.4%
2005 188.6 176 93.3%
2006 197.9 183 92.4%
2007 192.0 163 84.9%
2008 214.1 204 95.3%
2009 202.6 189 93.3%
2010 34.2 35 102.3%
Grand Total 1,381.8 1,274 92.2%

Table 3a: Walsh expected vs actual wins

Walsh is favoured by these figures as he rides closer to the number of winners that the market might expect. This indicates that his mounts are, at the very least, the better betting proposition. Whether that really means he is the better jockey is still open to question. It still depends a lot on what the market is saying. Another way of expressing the result is that McCoy’s mounts are shorter in price relative to their chances than are Walsh’s. This could be the result of McCoy’s mounts being significantly overbacked and would therefore not necessarily represent a disparity in ability between the two men.

A further indicator is our old friend consistency. Recall that this is the average of the difference between the best rating a horse has achieved and the rating it achieves on a specific day. Therefore, if a horse runs to a rating of 57 for McCoy and we know that it is capable of running to a rating of 74, the we can count 17 against McCoy for this run. By averaging across all rides a jockey has, we get a measure of how much a jockey gets, on average, from his mounts. Note that we cap the difference at 50 to ensure that no one result has too much influence on the number. This measure takes out of the equation both the quality of the rides that each jockey gets and the market expectations. It just looks at the quality of the performance of the horse for the rider relative to its overall capability. It is, however, dependent on the accuracy of the ratings used. I have used Raceform ratings for this analysis. Here are the consistencies for the two jockeys.

year A P McCoy R Walsh
2003 15.43 14.30
2004 18.98 16.08
2005 18.65 15.71
2006 19.65 15.48
2007 17.93 15.89
2008 16.85 14.68
2009 16.29 14.26
2010 12.60 9.58
Grand Total 17.66 15.11

Table 4: Jockey consistency over time

Walsh is strongly favoured by these figures, reliably scoring less than 16, whilst McCoy only manages to dip below this figure in his first year of the period. The evidence continues to favour the popular view that Walsh is the better jockey. It is possible to argue that McCoy takes more rides and is less selective and that this might adversely affect his figures. This is true, but only to a limited extent. Walsh has had plenty of rides and I don’t think it likely that this factor will have a material impact on the results. Walsh has the better strike rate, he is more likely to win on an equivalently priced horse and he gets his mounts to run closer to their best. These results are not some statistical aberration but can be seen right across the period. Specifically, he has had a lower consistency for every year in the period.

Before closing the article, I would like to examine one more facet of the jockeys’ figures, namely how often they fall or unseat. Much of the discussion on Saturday centred around the supposition that McCoy had put Denman into the fence wrong and that was later contrasted with Ruby Walsh’s sterling effort to remain on board Master Minded as he ploughed through the last in the Game Spirit Chase. Given our earlier results, the figures are surprising:

    A P McCoy
year Fall or Unseats Rides %Fall
2003 33 714 4.62%
2004 49 880 5.57%
2005 41 922 4.45%
2006 41 855 4.80%
2007 37 852 4.34%
2008 19 751 2.53%
2009 39 927 4.21%
2010 4 105 3.81%
Grand Total 263 6,006 4.38%

Table 5a: McCoy falls and unseats over time

    R Walsh
year Fall or Unseats Rides %Fall
2003 47 723 6.50%
2004 47 788 5.96%
2005 34 791 4.30%
2006 41 742 5.53%
2007 36 758 4.75%
2008 38 782 4.86%
2009 36 721 4.99%
2010 6 107 5.61%
Grand Total 285 5,412 5.27%

Table 5b: Walsh falls and unseats over time

So despite our conclusions on the relative merits of the jockeys, Walsh is the most likely to fall on his mounts. Again, this is a consistent result applying to all but one calendar year in the period. This is surprising for two reasons. Firstly, consistency is adversely affected by a preponderance of falls and unseats and therefore it is notable that the rider with the lower consistency has the higher number of falls. It strengthens our previous finding. Secondly, McCoy’s riding style is reckoned to be more aggressive but these figures suggest that he takes a more conservative approach at his fences; falling less often but getting less out of his mounts.

The main conclusion remains that Ruby Walsh is, according to the figures I have, the better jockey and the more reliable betting proposition. Tom Segal appears to be justified in claiming that McCoy is not the top jockey around at the moment. However, the question of the big race records of the two men remains open. McCoy may yet be the best choice for Denman in the big showdown. I will examine the relative big race records later in the week.