A Question From My Only Reader

In lieu of Trisomy 21, I figured it would be better to answer this question:

Moyer, question -

I've been casually reading your blog for a while and realize that most of your bets are placed on the team getting points. Is your mindset to pick against the pack? Is the success rate higher when you pick the underdogs? As you mention the other day your record isn't terrible (slightly under .500) but if you reversed your thought process would you be slightly over .500?

Absolutely, I bet on underdogs a lot more than I do favorites. I would guess the ratio is 10-1, or even 20-1. But the reason is because that is where the market is inefficient. People, in general, like betting on favorites. I wish I knew enough about stock market inefficiencies to frame it in those terms, but I know big time traders do the same thing on Wall Street (which is obviously working out well for them).

It's probably easier to put this in an example. Let's say the true line for the Penn State/Illinois game this week should be 10 points. But, since the bookmakers know that: it will be a popular game to bet on, Penn State is probably slightly overrated in the public's mind because of the blowouts, and that the average bettor likes gambling on favorites, they inflate the line.

To continue with the example, let's assume that if the line is set at 10 points, Penn State covers 50% of the time and Illinois covers 50% of the time. However, if Vegas sets the line at 10, 80% of the money will come in on Penn State. Before going forward, we need to specify some things. If 80% of the action is on Penn State, that means when Penn State covers, Vegas has to pay out $800 for every $1100 wagered, assuming -110 juice, while they get $220 for every $1100 wagered. Likewise, when Penn State covers, Vegas pays out $200 and receives $880. Here is the relevant math:

E(line) = P(PSU)*(Bets won/lost with PSU covering) + P(Ill)*(W/L Ill)

E(line) = .5 * (-800 + 220) + .5 * (880 - 200) = $50

So, because of the juice, Vegas still wins. Now, what if they set the line at 14 and only get 60% of the action on Penn State with P(PSU) down to 45%. When Penn State covers, Vegas has to pay out $600 while getting back $440. When Illinois covers Vegas pays out $400 and receives $660. Plugging back in:

E(line) = .45 * (-600 + 440) + .55 * (660 - 400) = $71

Notice, even though the line no longer has equal expectation for one side winning versus the other, Vegas has increased their profits, assuming an equal number of bets are placed on each line.

Also, on top of the mathematical theory above, you can play off public perception as well. For example, last week Tennessee got emasculated by Florida while Auburn won the first 58:57 against defending national champion LSU. In the public's mind, Auburn is much better than the Volunteers, yet they are only favored by six. You can figure this out intuitively by watching PTI, reading message boards, etc.

Vegas is taking on some risk in this case. If Auburn does indeed cover, the books are probably going to lose some money. That number is too low to get anywhere close to an equal amount of money down on both sides. However, Vegas also thinks that Tennessee covers a much larger amount of the time (this would be P(Ten) from above) than in the example I used. That is, in the books' minds, their expected value (E(line) from above) is huge by setting this number and they are willing to accept the increase in variance (aka risk). Just for S&G's, with P(Ten) = 80% and 90% of the action coming in on Auburn:

E(line) = .8 * (-100 + 990) + .2 * (-900 + 110) = $554

I'm exaggerating a little bit, but not a whole lot.

So, how does this all come into play for me? If I can find the lines where Vegas is taking a stand, like Auburn/Tennessee and Michigan/Wisconsin this week, then essentially, I am betting on Vegas to win. Obviously, the books' expected value is greater than mine because of juice, but you see my point. Vegas doesn't build $500 million casinos every year because they set losing lines.

The last thing to realize is that gamblers have become slightly (and I do mean slightly) more sophisticated in recent years. Most people are at least aware of the paper I linked to above, thanks to Bill Simmons. So now, the books will put out bait lines to get people to bet on underdogs. The Tampa Bay/Green Bay NFL line is an example. They could have set an equal action line like GB -3, but the game is in Tampa and these teams are really about equal, regardless of what the betting public thinks. Instead, they put out a bait line to attract a ton of Green Bay action. Any "sophisticated" gambler sees that line and takes the dog. The true contrarian knows the difference.

Anyway, so there is another look at contrarianism. I'm sure no one has read this far, and, frankly, I don't blame you. But it was at least cathartic to write the numbers out to show that I'm not completely retarded, the start of college football season notwithstanding.

Good luck this week.