Friday, September 12, 2008

Why I Have More Money Than Vegas Watch


"New England was favored against Kansas City by 16 points last week, with the total sitting at 43. This would indicate an average final score of 29.5-13.5. Plugging this into the football Pythag formula, we get an 86.4% chance of the Patriots winning."

That's interesting. I would have used the money line to arrive at a similar conclusion much more quickly, but manipulating numbers for no reason makes you look smarter, I guess.

That's nothing, however, compared to this abortion of an argument:

"At Matchbook, you can currently get the Patriots winning over 12.5 games at +250, and the under at -490. This indicates that there's somewhere between a 16.9% and 28.6% chance of them winning at least 13 games. That's an unfortunately large gap, but it's the best thing we have, so we're forced to take the average, 22.8%."

Forced? Forced.

When I see a line at +250/-490, I think there are a few different explanations for why the line is set that way:

- Perhaps the oddsmaker isn't looking to actually book any action on this prop, so he makes a line that no one except a complete sucker will bet into. If you've ever checked the futures odds at the Stratosphere, you've seen an example of this style of bookmaking.
- In the case of Matchbook, perhaps the people who are offering bets aren't confident in their estimates, so they shade their odds accordingly. They might continue to offer slightly better odds until someone accepts the bet, which will give them a better idea of the true odds.
- Most likely, since there is next to no interest in betting NFL regular-season wins on Matchbook right now--just look at how few offers are out there for all the non-Patriots teams--the +250 and -490 are fairly arbitrary, making this much like an illiquid market at TradeSports: you can't glean any useful info from it.

Here's what I've never thought: Staring at two arbitrary numbers, the correct line must be at the exact midpoint of those numbers, and I'm forced--forced!--to make this the basis for an "analysis" of how that team is going to perform this year without its star QB. I've also never extrapolated this into an evaluation of what that QB is worth to his team, even when I admittedly have no clue how good his backup is.

I'm tempted to buy all the +250 Matchbook is offering--down to +240 now, by the way--just to see Vegas Watch have to make a hasty rewrite: Since the best offer on the Over would then be +123, my bet would have single-handedly shifted the Pats' odds of hitting their over all the way up to 30.9%:

+123: Equivalent to 44.8% chance
Midpoint between 44.8% and 16.9%: 30.9%

Go Math!

In fact, I would go do this right now, except I don't like the Patriots. I should, however, go bet all the Rays futures I can to artificially increase their chances of winning the division and World Series.

I conclude with a partial list of some other things you can learn by reading into current Matchbook lines:

- The Angels have yet to clinch the AL West, since that market is still open.
- The Angels are also the favorites to win the AL pennant, even though they'll be a dog to the Red Sox in their probable first-round matchup.
- Until just now, Toronto was being offered at -200 to go over 85.5 wins, meaning their chances of doing so must have been less than 66.7%. I personally believe they'll get there 80% of the time (there I go again, doing my own handicapping instead of letting the market do it for me) so I took the -200. Luckily, with no more offers on the Over, there's no longer a ceiling on Toronto's chances of winning 86 games, so it looks like this bet will roll home a winner!

Update: Keith Law not only linked to the article, he appears to have bought into its premise. For shame, Keith.


Garrett W. said...

It also ignores the fact he can't derive a vigorish free mid-market price. Something to strongly consider learning if you're going to use implied probability from betting lines.

brent said...
This comment has been removed by the author.