[washingtonismoney]

Is that right? If the casino correctly predicts the betting, does the actual result matter? It seems to me that if the money comes in 50-50, then regardless of the actual result, the losers will pay the winners, with the casino taking its cut. It seems to me that it shouldn't matter whether the betting line turns out to be close to the actual result or far off the mark, because that shouldn't affect the number of winners or losers.

That's correct. But you need to take the analysis one step further.

Let's imagine casino lines are systematically wrong. That is, the casino's line -- while intended to produce a fifty/fifty betting split -- can be predictably bet against (whether moneyline or spread). Well then, of course, you can and would bet against it. If you could predictably make money in that fashion, you would amass brobdingnagian stacks of capital and unbalance the spreads. That would result in the casinos losing money.

So there's good reason to suspect that casino lines are pretty good guides to the ultimate outcomes to games -- that variation around them is pretty random, for the same reason that stock prices' variation is unpredictable. (Which is why I have my money in index funds.) If variation were predictable, you'd see people destroying the casinos. But we haven't observed either a class of pro gamblers or devastated casinos. We know that sports information is widespread and moves quickly. It's an efficient market! As such, the price signals are generally pretty reliable.

Of course, I'd be willing to change my views if there were reliable empirical research that demonstrating otherwise. I haven't observed that, so based on my priors, I suspect that's not the case.

[washingtonismoney]

Is that right? If the casino correctly predicts the betting, does the actual result matter? It seems to me that if the money comes in 50-50, then regardless of the actual result, the losers will pay the winners, with the casino taking its cut. It seems to me that it shouldn't matter whether the betting line turns out to be close to the actual result or far off the mark, because that shouldn't affect the number of winners or losers.

That's correct. But you need to take the analysis one step further.

Let's imagine casino lines are systematically wrong. That is, the casino's line -- while intended to produce a fifty/fifty betting split -- can be predictably bet against (whether moneyline or spread). Well then, of course, you can and would bet against it. If you could predictably make money in that fashion, you would amass brobdingnagian stacks of capital and unbalance the spreads. That would result in the casinos losing money.

So there's good reason to suspect that casino lines are pretty good guides to the ultimate outcomes to games -- that variation around them is pretty random, for the same reason that stock prices' variation is unpredictable. (Which is why I have my money in index funds.) If variation were predictable, you'd see people destroying the casinos. But we haven't observed either a class of pro gamblers or devastated casinos. We know that sports information is widespread and moves quickly. It's an efficient market! As such, the price signals are generally pretty reliable.

Of course, I'd be willing to change my views if there were reliable empirical research that demonstrating otherwise. I haven't observed that, so based on my priors, I suspect that's not the case.