This years Coventry stakes seems to be the cause of quite a bit of debate over the worth of the form. One camp is using the lack of winners to come out of it, as justification for saying it’s worked out poorly, while others led by Timeform’s David Johnson, and Kevin Blake in the Irish Field use a more sophisticated method to claim the form has a solid look to it.
In this weeks Handicappers Corner piece on Betfair, David claims ‘A lot has been said about the weakness of the Coventry Stakes form, but not much of it is borne out by the facts. Of the ten horses that contested the Coventry to have run subsequently, only two (War Command and Thunder Strike) ran below their Coventry figures next time, while seven of them ran to a higher figure, including the second Parbold and third Sir John Hawkins’.
In his Irish Field Blog Kevin Blake seems to be in broad agreement saying ‘ Firstly, far from working out poorly, the Coventry is working out reasonably well. Many seem to be falling into the trap of asking the overly simplistic question of “how many winners have come out of the race?”, but let’s not forget, War Command was SIX lengths the best on the day. This is a gaping chasm of superiority in the context of a six furlong race. It would be completely unreasonable to expect the horses that were beaten by six lengths and upwards to come out and win stakes races on their next starts. Yet, War Command’s four closest pursuers in the Coventry have all shown progressive form, a couple of them quite significantly, in Group 2 and Group 3 company since, giving the form a solid look’
Firstly I agree with Kevin in that using the amount of winners that come out of the race, is a flawed, and very basic method of evaluating whether the form of a race has worked out well or not. I would disagree with his assertion that his closest rivals have shown progressive form, and that it gives the form a solid look. Equally David’s claim that because 7 of the 10 horses who have run since the Coventry, beat the rating they achieved that day next time, somehow means the form worked out well, is slightly dubious. 5 of those 10 finished 9th or worse in the Coventry, so you would expect them to improve their rating in a more suitable grade, if they weren’t good enough, or if they were, then they just ran poorly that day, and thus an improved rating would be expected.
I think the best way of judging whether form worked out is, after evaluating the race on the day, and assigning ratings you feel are appropriate to the runners, you then look at how the horses do in their next runs compared to how you would have expected them to. I highlight expected as this is the key point. If horses were machines and always ran as good as they can without any other variables to consider, we could just look at the ratings we gave each horse in a race, then if they all achieved a higher rating next time, we could upgrade the form, and if they all ran below that rating next time we could downgrade it. We know that’s not the case so we need to use some intuition.
If the Coventry was a field of exposed handicappers, and all the placed contenders appeared to run to their optimum on the day, then using a method similar to what David and Kevin used, what be a very good starting point. It would follow that if the 2nd, 3rd and 5th improved their rating next time, that you might want to upgrade the rating you gave to the winner. This is exactly what happened with the Coventry runners, if using Racing Post Ratings anyway.
There are flaws with this way of doing things though, especially with a 2 year old race, which involved unexposed horses who would be expected to improve. Using the 2nd and 3rd as examples. Parbold recorded an RPR of 81 on his debut, 98 in the Coventry on his 2nd start, and then 102 at Goodwood. The 3rd horse home was Sir John Hawkins who recorded 89 in his maiden win, 95 in the Coventry and 102 and 101 in his next two starts. Now we have the 2nd horse improving his rating by 4lb next time, and the 3rd achieved 7lb better.
Does this mean, if this was the only evidence we had, we could upgrade War Commands RPR of 115 by 4lb+? I don’t think it does, as we go back to what would be expected. Both of these horses were having only their 2nd start in the Coventry, both improved a good bit from their debuts, and I’m sure both their trainers would have disagreed strongly if you suggested they weren’t capable of progressing again. That they achieved a slightly higher rating next time was thus to be expected, and just because they did so wasn’t reason to upgrade War Commands rating of 115.
To my eye the main reason for their low ratings in the Coventry is because, apart from the winner, very few, if any of the others ran their races optimally. Even the winner probably could have run faster had he delayed his effort longer, but given he strode clear in a part of the race that the others were going very slowly (only 2 other horses bar the winner, broke 13 seconds for the final furlong) means he was very likely flattered by his winning margin, and if the race had been run more optimally overall, its likely he wouldn’t have won by anymore than 4l. That means if his rating of 115 is justified, the others already shaped like capable of running 6lb or so better than the ratings they got, and it was these same ratings that were used to claim they showed improved form. I would suggest they just raced more optimally next time, and didn’t improve ability wise at all. Given they would have been expected to show normal progression from there 2nd to 3rd starts, I’d suggest the 2nd and 3rd were disappointing next time.
The 4th horse home in the Coventry was awarded a rating of 93, he has run twice since, to 76 and 101, he had achieved a 101 rating before the Coventry, and the fact he did it again after would just suggest he ran a little below par that day. The 5th Jallota ran to 90 in the Coventry, but to 105, 98 and 99 in his next 3 starts. Considering he walked the last furlong of the Coventry in 13.17, losing over 4 lengths to the winner, than its hardly surprising he did a bit better when paced more evenly.
To just use the subsequent ratings of the placed horses and conclude they back the form up with improved ratings, is ignoring expected improvement, and this is flawed. To give an extreme example lets say we have a horse, we’ll call him Horse A, that will always run as fast as he can, every time he runs, the only variable that changes his performance is the going, and he will not progress for fitness or any other reason. He’s having his first run in this country and while we know for a fact (lets presume its possible) that he’s 10lb better on soft, than fast ground, we don’t know what rating he should be on.
Lets say he runs 2nd, beaten a nose in a 2 horse race against another newcomer, who we’ll call Horse B, on fast ground, off levels. He gets a rating of 100, the winner gets the same. Now lets say he runs to a rating of 105 next time, on soft ground, in a big field handicap against exposed horses who looked to run their race. We know he can’t have improved, he never does, and we know he’s 10lb better on soft ground. Using a simplistic he ran to 5lb better next time method, would have you upgrading Horse B to 105 also. This clearly ignores the expected improvement (or in this case guaranteed improvement) from soft ground. This means we should actually be downgrading Horse B’s run to 95, rather than upgrading him.
We should have expected a certain amount of improvement for progression, from the placed horses in the Coventry, also for the fact they were paced more optimally next time. They appeared to have no excuses yet didn’t improve the expected amount, so whatever rating you gave War Command on the day of the Coventry, then I suggest the subsequent events should make you downgrade it by a few pounds. To further back that up, given how Aidan O’Brien’s horses generally improve, War Command himself certainly didn’t do anything for the form on Sunday.
Quite apart from whether the form worked out, based on the ratings they got on the day, the 2nd to 5th have run 8 times between them since the Coventry, and the best 5 racing post ratings they achieved since are 105-102-102-102-101, compare that to last seasons Coventry when by mid August the same sample of horses, had also run 8 times between them, this time receiving ratings of 112-110-110-108-106. No matter which way you look at this years Coventry, the form looks below recent standards for the race. It may look better by the years end, and just because the form is poor in behind him, doesn’t mean War Command can’t be top class, even if flattered by the margin. A 4l win on only his 2nd career start would still have been decent form, and a huge step up on his debut run. Given who trains him, more progress would be expected, it would be a worry that he didn’t appear to come on when beaten on Sunday, but the time figure for the race was good and he’ll surely be better over further.
If you find this post interesting, please share it via the social media buttons, you own blog, or online forums.