Secretariat winning the Belmont Stakes by 31 lengths (photo via Secretariat.com)
Many racing fans believe that a horse that won its last start in dominant fashion is a good bet today. The argument is that such a horse was “never asked for its best” and, thus, is a prime candidate to repeat or even improve on that winning effort next time out.
Never one to accept beliefs — especially racetrack beliefs — as facts, I decided to test this theory on my database of nearly 117,300 last-out winners.
First, to establish a baseline, I simply looked at horses that had won their last race by any margin:
Number: 117,256
Winners: 20,508
Win Rate: 17.5%
$2 Net: $1.63
ROI: -18.35%
Then, I searched for steeds that had won their last race by five lengths or more:
Number: 16,790
Winners: 3,739
Win Rate: 22.3%
$2 Net: $1.66
ROI: -17.20%
Notice that, although the win rate is higher, the (negative) ROI is about the same.
This is a common phenomenon when it comes to horse racing stats. Despite what the tipsters and selection hawkers will tell you, pari-mutuel betting pools are generally efficient. Hence, any factor that wins more is likely to pay less.
One way that players can stem the flow of red ink, however, is by using mass public misconceptions to their advantage. For example, the vast majority of horseplayers believe in the “bounce theory,” the notion that a big effort, e.g. a dominant win, will cause a horse to regress, or “bounce,” in its next outing.
Despite my best efforts over the years to show that such regressions are generally statistical rather than physical in nature and that the bounce theory is largely bogus, both new and old racegoers cling to this mistaken belief like Browns’ fans waiting in line to buy playoff tickets.
Of course, the cure to these “bounces,” the myth goes, is to give the horse in question some time off to “recuperate.”
This is nonsense of the highest order.
In the years that I’ve done database testing, I’ve learned that for a form factor to be relevant, it must be recent. This is true of dominant wins as well, as the data below clearly demonstrates (or demonstrate for you Latin purists):
- Horse won its last race by at least five lengths and is returning to the races in two weeks or less.
Number: 3,271
Winners: 814
Win Rate: 24.9%
$2 Net: $1.85
ROI: -7.55%
- Horse won its last race by at least five lengths and is returning to the races after a break of two months or more.
Number: 1,581
Winners: 260
Win Rate: 16.4%
$2 Net: $1.31
ROI: -34.57%
Not only is the winning rate for big winners returning to the races quickly significantly better than the strike rate for their “rested” counterparts, but the ROI is much better too.
As bettors, we can use this information to improve our bottom line. Knowing that our fellow punters are overestimating the chances of dominant winners returning off a freshening can help us eliminate a lot of likely losers at a short price.
Likewise, knowing that honing in on big winners returning to the races within two weeks effectively cuts the track take in half, is a great starting point to the handicapping process.
