As those of you who have been following me since the inception of this site may have noticed, I have an allergy to low-priced horses. Those of you who know my friend Heckling Harry know that he has never seen an odds-on choice that he doesn’t like. In this week's column, I want to address a few important issues around this topic and explain some mathematical anomalies which you may not be aware of.
First off, what is the goal of handicapping? Most think that handicapping means attempting to pick the winner of a particular race. You cannot pick the winner of a particular race. Why? Simple – because there is no one winner of any race.
Let me explain. Racing is just like life. Consider it as a gigantic pill box with a million cubes in it. Each one of these cubes has a different result. If I reach my hand in and pick one, on any one pick, any cube could be selected. The cubes have different orders of finish on them…some have the #1 winning, some the #2, the #3, etc.
Every horse has some chance of winning. That is to say, that if we ran the race a million times over, every horse would win some percentage of those races. This means that every horse has his name on one of those cubes, and you don’t know which one cube will be selected tonight. Of course, if you were asked to wager on it, it would help if you knew exactly how many cubes had each horse’s name on it…then you would know exactly where you stand.
For example, if we had 100 cubes in the box, and I knew FOR SURE that 80 of them said "#1 wins" then I would know that he is the most likely winner. Surely though, one of the other 20 could be selected. Now if I were asked to bet on it, and each bet cost me $1, you would have to offer me odds - a payoff of at least $1.25 before I would be willing to play. If you offered me $1.10, for example, after 100 plays, I would bet $100, I would win 80 times at $1.10 each, or $88 - thereby losing $12, or 12%. On the other hand, if you offered me $1.40, I would win $112 for a net profit of $12, or 12%. Okay with me so far? Good.
Now here’s the kicker…you don’t know for sure how many cubes with each horse are in there. So the goal of handicapping is not to pick the winner, but to estimate the number of cubes in the box that have each horse'’ # on it. Once you do this, betting is easy. The better handicapper you become, the better able you are to estimate the true number of cubes in the box for each horse.
By and large, the handicapping public does a nice job of identifying the horse who has the most number of cubes, and they bet him down to favoritism accordingly. But more often, they bet more money on him than they should, based on the true number of cubes, or as we would say in statistics, based on the horses true probability of winning. These numbers aren’t exact, but an average favorite at Pocono I would guess pays about 7-5 (maybe less because Pocono tends to card more non competitive races then, say, the Meadowlands - by non competitive I mean races where one horse sticks out like a sore thumb). If the favorites win 35% of the time, let’s look at the math:
By the way, if the track’s take out of the win pool is 18%, and all the rest goes back to the bettors, this means that the crowd, as a whole, on win bets, loses 18% . In fact, this is what you would do if you picked totally at random. So in the above scenario, by betting on the "best horse" in each race (i.e. the horse with the best chance of winning), you would not be doing much better than you would with a dart board! In fact, what little difference there is would not be worth the time and trouble you go to in selecting these horses! This is REASON #1 that I have a bias against favorites - IN THE LONG RUN, YOU LOSE YOUR MONEY - and I don’t know about you, folks, but I’m there to win.
The second thing you need to understand about betting short-priced horses is the effect of the track’s "take". The track is fundamentally different than, say, a casino. In the casino, you play against the house, who sets the rules and the odds, and the odds say: YOU LOSE, WE WIN. It’s as simple as that. At the track, though, you are playing against the other bettors, not against the track. What the track basically does is it takes a certain percentage or commission of the top based on the size of the pool. These percentages vary among the different betting pools, and from state to state. Most states, win betting, the track’s "take" is 18% off of the top, the remaining 82% goes back to the bettors based on the odds that the players determine, not the house. So the track is really just a passive observer, basically charging a fee for the use of their facilities, and for helping to organize the game. Exotic wagering pools have a higher "take", like 22% for exactas, 28% for trifectas, etc. Here’s another reason why I don’t like playing trifectas at a place like Pocono Downs - first, there’s not that much money in the pool and second, the track’s "take" last time I checked was like 35%. Whoa Nelly – that’s one hell of a price of admission, don’t you think? Sounds like bad business to me, and that’s how you have to view this stuff, if you wanna make money - like a business.
Now what does the track’s "take" have to do with betting low-priced horses? Some more math: Lets say, hypothetically, that the track did this stuff for free, okay? No take. Further, let’s say there was a win pool with $1,000 in it. Further assume that $500 of this was bet on the #1. Okay, what should the odds be: well, even money, or 1-1. (Think about it, if he won, the bettors of the $500 would each get a $4.00 win payoff, or twice their investment - even money). So then, the whole pool would be distributed and the race over. Say another horse in that race had $100 bet on it. The odds would be 9-1. Now, next time your at the track and you see an even money shot, look at the tote board and see if he has half the pool bet on him. Answer - no! Why not? The track’s take, that’s why not. Here’s how it works.
In the above example, with no take, #1 would pay $4.00. But the track takes 18%, or .72, meaning that he really ends up paying $3.28 – (actually, he pays 3.20, because they truncate downward to the nearest 20 cent interval - called breakage - I’ll mention it in a moment). IN OTHER WORDS, HE DOESN’T GO OFF AT 1-1 LIKE HE SHOULD, BUT GOES OFF AT 3-5! But let’s say he paid $3.28. Think of the business aspect here. You bet $2.00, had a $2.00 profit and the track TAXES you 72 cents - 36% of your profit - THAT’S MORE THAN BILL CLINTON WANTS! Boy, sounds like bad business to me! What business do you know that could survive on a tax rate of 36% of it’s profit??
Now take the other horse, the 9-1 shot we computed above. He should pay $20, but he doesn’t - after the track’s take, his supporters end up with a mere $16.40. This horse does not go off at 9-1, but the tote board says "7-1". The "take" of $3.60 = 18% of your profit, versus 36% for the bettor of the short priced horse! This is reason number 2 for not betting short prices, regardless of how strongly you feel about the horse - THE EFFECT OF THE TRACK’S TAKE ABSORBS AND INORDINATELY LARGE PERCENTAGE OF YOUR PROFIT.
The final reason for my bias toward higher priced horses is something I hinted at above, the "breakage". Horses don’t pay $3.28. In Pennsylvania, they use what is called "dime breakage" which basically means that they truncate (not round) down to the next 20 cent interval. This means that a horse that should pay, even after the take, $3.39 ends up paying only $3.20. This sounds like stealing by the track, and it is. But it’s legal stealing! Again, 19 cents don’t sound like much, but if you’re betting a horse that should pay $3.39, and you only get $3.20, that 19 cents equals another 5.6% of your profits! Holy Cow! C’mon guys - get real! So overall, you can have a horse that should pay $4.14, then after take and breakage only pays $3.20! FOLKS - THAT’S 44% OF YOUR PROFITS DOWN THE TUBES - I HOPE YOU’RE FOLLOWING ME! On larger priced horses, the percentage effect of breakage is minimal. If you’re horse should pay $16.59, but you only get $16.40, well that’s only 1.3% of your profit…that I can live with.
For a long time Canada, and more recently New York, have adopted "nickel breakage" which means they only truncate down to the nearest 10 cent interval.. This is much fairer for the bettors. Of course, no breakage would be the fairest. What’s the matter, they’ve got computers? Well, I guess thinks could get a little hectic at the windows if they were giving out pennies, ya know?
Steve M.
