Tyler Cowen points to a Washington Post story about historical racing. It’s just like betting on a horse race, only the race has already happened.

Yes, you read that right. Critics of the proposal say that it really is a thinly-disguised (and have you noted that nothing is ever “thickly disguised”) form of slot machine. The key details:

In historical gambling, which is also called instant gaming, customers would put as little as a nickel and as much as $5 into a video terminal that resembles a slot machine. The terminal randomly selects a race from an archive of at least 10,000 previous horses races from tracks around the country. Customers review a graphic showing the odds and statistics for each horse before deciding which one to bet on.

The race appears on the monitor. If the chosen horse wins, the patron will receive a payout based on the odds, how much was bet and that day’s purse.

Actually, based on this description, there’s something much more wrong with historical racing. The obvious attack on any bet on anything that has already happened (instead of has yet to happen) is simply for a gambler to find out what the result was, said result now being a sure bet. This trick is the basis for a number of common cons; I arrange to get the results early or to delay your receipt of the results, and convince you to place a too-good-to-be-true bet in the interim.

Presumably, selecting from a large database of races and showing only horse odds and stats (rather than day-and-time and horse names) is to foil this form of attack. But that level of obfuscation, while it turns the game into a game of chance against the ignorant, doesn’t seem sufficient to foil the dedicated. I just need to compile a large database of odds, statistics, and results from historical races. Then, when the machine shows me the details of the randomly-chosen race, I punch enough of them into a handheld computing device with the database loaded on it. In a 10,000-race database, it should be pretty easy to zero in on the actual race. Armed with that information, I place a can’t-lose bet.

Since I don’t usually think of casinos, racetracks, and other professional gambling institutions as being that dumb, my assumption is that there’s some further angle. That angle, if it exists, is unlikely to increase the role of skill in the game. And if said angle doesn’t exist, I predict that historical racing will have a short and ignominious life.

Maybe a number of historical races have identical statistics? It seems unlikely, but it’s certainly possible.

The truth, though, is that the racetracks are probably betting on people not bringing around portable databases to beat the system. And if they catch you with one, they’ll kick you out. The precedent for this would be casinos’ reactions to card-counting during blackjack. Technically speaking, card-counting isn’t cheating, and so it isn’t illegal. But casinos are notorious for harassing gamblers who engage in it, simply because that technique gives counters an advantage over the house. I’d imagine a similar response to handheld databases or other methods of “gaming” historical races.

What’s really odd about the whole debate is that people are opposed to slot machines (and similar games) but are ok with betting on horse races. Now, that’s the real mystery.

There’s absolutely no way they could release a system that both provides enough information for you to look up the data and has a publicly-available data set like this. (Rapidly changing the data set wouldn’t help either.) All gaming devices generally go through thorough security audits, and this won’t get past the front door.

This isn’t just a slight advantage on the house - it’s a theoretically sure bet (plus is much less susceptible to player error). Additionally, it only takes “one” person to beat the system for everyone to do so, as I could sell my handheld system.

My hunch is that the player won’t get enough information from the console in order to narrow down the choices, or won’t actually get choices to narrow down at all. In fact, the console could create fake odds for all losers in the race, since it won’t have to pay at those levels. This makes gaming the system basically impossible (less impossible if they pay for three winners, aka win/place/show).