The bounce theory means different things to different racing nations. In the US they normally use the phrase ‘he bounced’ to describe a horse who put up a huge career best speed figure and then regressed badly next time. The theory being the effort took it’s toll on him, and he couldn’t reproduce it. This isn’t altogether surprising as when most horses put up a career best its because everything went their way, and it would be only natural that they’d be expected to regress back towards their mean performance next time, even without the big effort taking more out of them.
In the UK and Ireland however we use the phrase ‘I’m afraid he might bounce’ or ‘he bounced’ to describe a slightly different scenario. It would normally be when a horse is coming back from an extended break, say 300+ days, and then after running very well, returns to the track quick quickly, say within 21 days. I was watching ATR a few weeks ago when the pundit in the studio was very concerned about Llanarmon Lad bouncing, after running very well on his first start for almost 2 years, only 3 weeks earlier. ‘There was a massive gap between his 2nd run and the last day, and the worry is the bounce’. He went onto to repeat his bounce concerns twice more and said ‘it was a definite worry’.
Over the years you will have heard many pundits, both on TV and in print, express worries about a horse bouncing, so I decided to use Proform Software to see is there any truth in the theory. Proform is a very good database, and I use it often to find angles for both my Premium Advisory Service and Nap of the Day page. The good thing about it is the ability to export results into excel to do more advanced analysis. For the purposes of this research I’m going to split the database between flat and jumps, and I’m only going to use handicaps. I will include all horses who finished in the first 3 in a handicap last time, having been off the track for 300+ days, and now run in a handicap within 21 days of that last run. For comparison purposes it’s important to then take a similar sample of horses who weren’t coming back from a break last time, so for that I’ve used the same criteria of being in the first 3 in a handicap, except this time they will be returning from a break of not more than 100 days. This group will also be running within 21 days. I’ve used data from the last 4 years.
Starting with the Flat the above screenshot shows the results for horses who were in the first 3 last time, off a break of less than 100 days, and now return with 3 weeks.
Now we compare those results to the horses whose last run was after an extended break. The win strike rate drops from 15.9% to 12.56%. With a sample of just 199, using places will tell you much more than just wins, and again the horses on a 2nd run after a long break fared worse. With 35.68% of them placed compared to 39.45% from the other group. You will also notice the Actual/Expected figures of 1.01 for the 1st group, and 0.68 for the 2nd. That means if backing them at Betfair SP to return a set amount, you would only get 68 pence back, for each pound invested, on the horses considered most likely to bounce. The expected winners was 36.63, were as only 25 of them won. Again though the sample is small so the Betfair Place SP would tell us a lot more. Unfortunately this isn’t included in the Proform Database, but I do have my own database with these. It doesn’t include the last few weeks data, but of the ones I have the bounce theory group placed 71 times according to Betfair place rules. 80.25 places would have been needed to break even though, so the return on every pound invested was just 0.884. A loss of nearly 12%. It seems there may be something in this bounce theory, on the flat in any case, lets move onto the jumps action when the criteria was the same for each group.
Above is the results for the horses who were in the first 3 last time, having run within 100 days of their previous race, and are returning within 3 weeks. This sample includes chases, and hurdles.
Again I compare that sample with the horses who qualify using the bounce theory criteria. This time the win strike rate for the horses who were returning from a long break last time is higher, 17.28% versus 16.96%, the place strike rate is also slightly higher at 42.39% against 41.18%. Nothing in it really, but again it’s noticeable you’d have made a loss backing the ‘bounce theory’ horses with the return of only 85 pence for every pound invested. I can’t stress enough that the difference isn’t huge though, there was 42 winners, were as 49.7 would have been expected given the Betfair SP’s. I again checked how the horses in the bounce category fared to Betfair Place SP. This time the returns were a little better than the flat, 103 were placed, when 109.4 was expected, for a return of 0.941 for every pound invested. A loss of just under 6%.
The Bounce Theory – Horses are more likely to run poorly returning quickly 2nd run off a break – Hit or Myth?
The results suggest there is something in the theory, but not really what people who quote it suggest. The general insinuation is that the horses who had a big break between their 2nd last, and last run, and now return quickly, are likely to fare a good bit worse than horses who achieved the same last time, but run more recently on their 2nd last run. On the flat the horses who were returning from a shorter break last time did do a little better in both win and places strike rates, but not by much, were as over jumps the horses who were returning from a long absence last time actually did better in the strike rate departments. What was therefore a little unusual considering how many times you here the theory, is that while they didn’t perform much if any worse in a positional sense, the bounce theory horses returned a loss, both on the flat and over jumps, for wins and places. Normally if the general consensus is that a factor is a negative, that turns out not to be, the prices on this sample will be bigger, and thus they will prove profitable.
This isn’t the case here though as you’ll see if you take the 29,345 flat runners who were returning from a short break last time, and divide that by the expected winners of 4,667, you will see the average price was 6.35 in this sample. Using the same method the average price for the possible bounce horses was 5.43. It’s a similar story if you apply the same calculation to the jumps sample. The reason you made a loss backing these horses wasn’t because they performed any worse, it was because they went off at a shorter price on average. The most plausible reason for this could be that because these horses were returning from a break last time, the market expected them to improve for the run. They didn’t run any worse, but they didn’t run any better than horses returning from a shorter break either. This suggests the market pays little heed to the bounce theory, but does in fact overrate these horses, as it expects improvement that isn’t forthcoming. This isn’t that big a surprise, anybody can get a horse fit after a break these days, and the fact that the horse finished in the first 3 in a handicap after the break, suggest they ran right up to their best, thus instead of expecting improvement as the market does, it would seem more prudent to assume the horse was fit, and thus isn’t any more more likely than your average horse to improve next time. So the bounce theory boys come to the right conclusion that the horse is a poor bet, but with the wrong logic. Kinda like your girlfriend picking the grand national winner, after spending 20 minutes considering each horses name and colours!