Still Looking for My Answer on the Horse Betting Quiz

Yesterday morning, I posed a challenging question for horse bettors:

Is there ever a situation in a parimutuel betting system in which you would want to bet on a horse to win, even though you knew for sure that the horse would lose the race?

Some clever folks came up with an answer involving “breakage.” That was not what I was looking for, but it is correct, meaning blog reader “dd” –the first one to mention it — will win a prize.

But I am still looking for the answer I have in mind, so keep on thinking. I knew this was a hard one.

Hint #1: It does not involve arbitrage between the parimutuel pool and any other odds.

Hint #2: It does not involve influencing other people’s perceptions of what the true odds should be.


I wonder if it has something to do with betting on all the horses a certain amount related to their odds. Obviously all the horses can't win so you are betting on some losers but if there was a certain wrench that was thrown in (ie: scratch horse) that messes up the oddsmakers and tips the winnings in your favor, you may end up guaranteed to win some money reagrdless of which horse wins.


Matthew J. Marshall:

Well, if we have no restrictions on the question, then there are no restrictions on the solution. Perhaps you have a perverse utility function and you like to lose money. Or maybe you value the color of the losing horse more than you value the money you will lose. Etc. Without restrictions, the question is not well defined.

"Foul play" and outside-of-the-parimutuel-system considerations are obvious answers to the problem. All of them are obviously correct: taxes may make you want to lose a bet; owning the horse may make you want to lose a bet; affecting odds on another bet may make you want to lose a bet. However, it is precisely because they are outside of the system that they are trivial answers and, I suspect, not at all what Dr. Levitt is asking for.


The only way you could know for sure a horse would lose a race is if you are involved or have knowledge of horse tampering. If you had tampered with horses in a given race, you may consider betting on a horse to win that to help cover your tracks.


You are the horses mother


How would you know for sure a horse would lose?
what Mike said ^



Thinking out loud, perhaps unwisely: The money that you bet on the horse that loses goes to the ones that bet on the horste that turns out winning. If you are the only one who bet on the winning horse, you get 80% of your losing bet back, even though the horse you put that money on lost. (The bookie keeps his 20%, after all.)

Presumably, that is the best you can do. If others, other than you, also bet on the winning horse, that 80% number drops fast. In any case, I don't see how you could get it up to 100%, where it would have to be for it to make sense betting on a losing horse.

Ken D.

I too am bothered by "you knew for sure that the horse would lose the race". I wonder if what is meant would more precisely stated as "you considered virtually certain to lose, so much so that the possibility of it's winning does not affect your calculations." (If I can get Old Dobbin into the Breeder's Cup Classic, it is mathematically possible that the rest will all go down in a pile-up and we will win in 5:00.) But if that is not what "sure to lose" really means, it may hold the key to the puzzle.


If you had won enough money gambling to move you just across the threshold of a higher income-tax bracket, you could make a losing bet to reduce your taxable income below that threshold.


acidtest's comments of yesterday already suggested several "non economic" options or, as Caleb Powers put it, the "non-betting reasons."

Of course, there is a classic example of seemingly "non-economic" behavior (i.e. a reversal of the apparent value of money) in Montgomery Brewster. He was only allowed to lose a certain amount to gambling in his quest to rid himself of a million dollars, but a can't-win horse would have been useful to him.

Since neither of the above posters were mentioned in this follow-up, it seems safe to suppose that the fundamental "more money is better" assumption is intended to be maintained in this exercise.

Mike D

If it's part of a coupled entry.

Frequently, trainers will pair horses in races as a single betting entry (1 & 1a, e.g.). You could be certain that the 1a horse will not win the race (he could be a "rabbit" to pace another horse), but you like the 1 horse to win, so you'll technically be betting on 1a to win.

Dustin Borklund

I just replied on the original post, and then I saw this post so I'm repeating my response.

If the horse you bet on gets scratched your money moves to the favorite automatically. So, if you know a long-shot horse if going to get scratched, you can bet on him and get better odds than if you had bet the favorite, even though you win if the favorite wins. This may only apply to parlay bets - it's been a while since I've been to a racetrack.


I have never bet on horses, and I am poor at math, but maybe this info will help someone else. Isn't it true that when you place a bet a bet on a horse, that horse becomes less of an underdog? For example, In the most recent Kentucky Derby 'Street Sense' was rated at 4-1, whereas 'Sedgefield' was rated at 50-1. Obviously Street Sense is the favored horse, it only pays $4 for each $1 bet, whereas Sedgefield pays $50. But if for some reason people started betting on 'Sedgefield' in great numbers, that horses 50-1 odds would start to shift, and the payoff would decrease (say to 30-1). Now the question is why would someone do such a thing?


The tax bracket has been along the lines I've been thinking too. However, I've been thinking rather than one bet putting you into the other tax bracket, it would be a series of bets. It would be a mean person that would rather loose money gambling to change tax brackets than just give it to charity.

Just thinking out loud: With the tax system, say you are a successful better and can make 75k in a year, but you claim 25k in gambling losses from picking the losing horse. You get a tax deduction for the 25k then. The trick is that if you simultaneously bet on the winning and losing horse, you have made most of the 25k back in your winnings, but you still get the deduction.

Mike B.

If you bet the horse to show, you get paid if the horse comes in 1st, second or third. So technically, you are still betting on the horse to win the race. The track (by law, I believe) has to pay you at least 10 cents on the dollar. Therefore, even if it loses (comes in 2nd or third) you will get paid a minimum of 10% no matter what the money pool is.


Let's say that you have a time travel device that has allowed you to go back in time with the intent of amassing a fortune on the races.

You are winning race after race. But just as the owners get suspicious of someone who is "too lucky" in Las Vegas, the powers note the substantial bets being placed--especially that $100,000 bet on that 30-1 horse that surprised the world by winning.

And so, to keep them from arresting you--which would mean the lose of your fortune and the inability to get back to your time machine--you start strategically losing. All of which is calculated to make them think you are a rube who had a tremendous streak of luck, thereby throwing them off of your track.

I know this is surely what you had in mind, Steve. Please send my prize to me now.


A quinella

Kilted Raven

Where you'd bet someone else a larger amount of money that you could always pick the losing horse.


OK, another one....

Your eccentric billionaire uncle just died and left you his vast estate. But there is a stipulation: You have to continue his tradition of going to a certain race track and betting on the horse with the worst odds. If the horse loses, you will inherit billions. If the horse wins, it will be an omen that you, too, against all odds, will turn out alright without his billions, and his wealth will go to his pet cat.

So, in fulfillment of the stipulation in the will, you go to the track and bet on a horse that not only has the worse odds, but one that you have ensured will LOSE. (You would prefer the billions, would you not?)

You may have drugged the horse. You may have hypnotised it. Or, far more likely, you have offered the rider $10 million to MAKE SURE the horse does not win. Or perhaps you have kidnapped the rider's bride and will release her when he loses.

In any case, just as stipulated, you bet on the horse to win. Yet all the time, you know it will lose.

And so, when the horse loses, you pay off the rider, or release his bride, or unhypnotize the horse, etc., and go and collect your BILLIONS.

It makes all the sense in the world!



In parimutuel betting don't you normally know who you're betting against, so potentially you could bet to lose in one race, not to change the horses odds, but to change other people's perception of your own betting abilities?

That way, by betting to lose on a Monday, on the Tuesday you could bet against the same person but this time to win, and your opposite number may think you are a foolish bettor and take your bet despite poor odds, thinking it was easy money.


You can "lay a horse to lose" meaning you bet on a horse you think will lose through a betting exchange where someone else bets the horse will win. If the horse loses, you keep the stake. If the horse wins you pay the other party.