So I watched Redline last week, and it is one fine film, definitely worth checking out. But this post is specifically about one of the plot points in the movie: that protagonist JP and his partner Frisbee fix races for the mafia, which they do by having JP hang back until the last quarter of the race, suddenly take the lead, before ultimately losing the race. There will be some minor spoilers.

I got to thinking what kind of scheme the mafia had going that allowed them to profit from this situation, and what they needed JP to do. And it made me realize that, given JP’s capabilities, they could have fixed it a different way which would have allowed them to profit regardless of whether or not JP lost the race.

First of all, what was the mafia’s business model in Redline? At first, I thought it was by being the house and enticing people to make many losing bets on JP by making him look like a sure thing at the end of the race. But towards the end of the Yellowline race that started the movie, the mafia boss’s underling came to him and said that “all their positions on JP” had been “unloaded.” This implies that, in fact, the mafia no longer had any bets for or against JP at that point, so they had no reason to care whether or not JP won. It also implies that there is an after-market for these bets that they could sell already-made bets to.

So I’m not sure what scheme they had going on (if anyone is, please let me know in the comments!). But it seemed like they were pretty dependent on JP’s capabilities: JP needed to be someone capable of both (a) convincingly being an underdog for most of the race and (b) convincingly being the favorite for some of the last part of the race. And, of course, the mafia was quite insistent that (c) JP lose the race.

The fact that JP is conning the spectators twice, first by pretending to be worse than he is, then by pretending to be about to win before losing, should set off an alarm; you only need to con someone once in order to make profit. Indeed, if JP is capable of (a) and (b), there is a pretty safe way for the mafia to fix the race, one that doesn’t care if JP wins or loses.

So here’s the new scheme:

1. Keep JP at or close to last place for most of the race. Make bets for JP as late as possible while he’s still behind.
2. In the last segment of the race, have JP suddenly take the lead. As late as possible, take all those bets you bought for JP and re-sell them in the after-market.
3. Profit! It doesn’t matter who wins the race, because you’ve sold all your bets, and all you’re holding is cash.

How does this work? The key lies in the fact that the bookie must adjust the odds of a bet in a predictable way. And this affects the price at which one can sell bets in the after-market.

First, a quick review of how one accurately determines the value (expected value) of a bet. If you have a ticket representing a $1 bet on JP, its value V is determined pretty simply:

V = Y*Z

where:
Y = payout of the bet.
Z = probability that JP will win.

Betting odds during the end of the Yellowline race

One thing bookies have to do is to try to keep that expected value V fixed, and under 1. Consistency is important so that you get an even distribution of bets for the racers and under 1 is important so that the house is likely to make a profit. Everyone knows this, so the bookie must follow this behavior so that there is no suspicion of foul play. As Z fluctuates throughout the race, the bookie must change Y in the opposite direction to keep V as close to constant as possible (of course, there is flexibility in real life, since no one truly knows the value of Z).

Here’s an example using dummy numbers. Let’s say the bookie wants to keep V at 0.9 throughout the race. During phase 1, when JP is far back in the pack, Z is very low, making Y very high. Let’s say that, when the mafia bets on JP, Z(1) = 1/10,000 – probability of JP winning is 1/10,000. Then the bookie sets Y(1) = 9,000 – the $1 ticket that the mafia bought will pay out $9,000 if JP should win.

But then enter phase 2: JP surges to take the lead. Now, Z(2) = 1/10. The bookie adjusts the odds accordingly so that Y(2) = 9. But the mafia has already bought a lot of bets that will pay out 9,000 instead of 9. So the mafia turns around to the bettors and offers to sell these bets to them, for a premium, of course. These bets are now worth Y(1) * Z(2) = 9,000 * 1/10 = $900.

If bettors were willing to purchase bets worth $0.90 for $1, then they should be willing to purchase a bet worth $900 for $1,000. Let’s say the mafia offers it to them for $900, to entice them with a better deal compared to what the house is offering. In fact, the expected value of the bet is 1 at the price of $900, so ignoring risk aversion, there is no reason NOT to take it (plus, gamblers tend not to be very risk averse people).

And then, the mafia walks away. It has “unloaded” all its “positions on JP” and is sitting on a pile of cash. It bought bets for $1 that it sold for $900, a nice 89,900% profit. Of course, these are dummy numbers, but as long as the bookie follows this predictable behavior, and as long as the after-market is liquid enough for the mafia to resell all their bets to bettors, the mafia will profit. Using more reasonable numbers, even if the jump between phases 1 and 2 of Z was from 1/50 to 1/10, and V was fixed at 0.5, the mafia would make a 150% profit.

This fix is almost exactly what insider trading is in the stock market. In this case, JP (or more specifically, the probability that JP will win) is the stock, and the non-public information is the knowledge that (as well as of when) JP’s probability of winning will skyrocket. It would be like an insider knowing that a company will soon be bought up by another company and buying lots of shares in that company’s stock before it happens.

Here, the mafia is in an even better situation than insiders, because they are actively manipulating the stock instead of just knowing how the stock will move. The incentives line up so that the mafia wants Z(1) to be as low as possible and Z(2) to be as high as possible. Both are accomplished by having the phases 1 and 2 end as late as possible in the race; as it gets closer to the finish, the odds tend to get more extreme: the probability that someone who’s behind will win gets lower, and the probability that someone who’s in the lead will win gets higher.

It will look awfully suspicious both to bettors and to law enforcement if you see one player buy and sell huge bets at such times. So the mafia would want to hire many bettors to do this, perhaps during slightly staggered times, and on much smaller scales as not to arouse suspicion. So the profit would be slightly variable, and the bettors would have to get their cut. It would also create more possible holes, more people who could talk to the police. Still, don’t you think the mafia boss would’ve preferred to take on that extra risk given to what happened?

This wouldn't have happened if the mafia had adopted this business model.

Of course, if the mafia had done this and hadn’t cared about whether or not JP would win or lose, we wouldn’t have gotten the awesome moment when Frisbee was saved, nor would we have gotten the epic finish with the planted bomb. So all in all, I guess it was for the best.