Baseball futures can be found at books right now, and one tool to help when trying to get a good grip on win totals is the Pythagorean W-L Theory as it's applied to MLB teams.

Many books are offering lines on MLB season wins. Like all bets on futures and props, these are often easy pickings. There are two key reasons these can be grossly mispriced: public misperceptions and significant roster changes.

The normal public bettor believes that all else being equal, a team will win about the same number of games that it won the prior year. A better way to estimate how many wins a team will earn the following year with the same roster is using a variation of the “Pythagorean theorem.” In the course of a year, teams will win and lose many close games. For the most part, this averages out. However, some teams are exceptionally lucky or unlucky. These teams look better or worse than they actually due to luck.

Bill James’ “Pythagorean Theorem” of baseball suggests that if you know how many runs scored (RS) and runs allowed (RA) a team has, the odds of a team winning an individual game is:

RS ^ 1.8
----------
(RS^1.8) + (RA^1.8)

When the theoretical win rate is very different from the actual win rate, teams are likely to move towards the theoretical win rate.

Last season, the Los Angeles Angels won 100 games. Scoring 765 runs and allowing 697 runs, you would expect them to win about 88 games. Not surprisingly, they are currently priced at Under 89½ (-150) at TheGreek.  On the opposite extreme, Toronto won 86 games last year, versus a theoretical 94 games. They are currently trading at Under 80½ (-220), which is surprising unless you look closely at the roster changes.

In the days before the season opens, there are a lot of roster transactions and changes to the disabled list. These betting lines are usually left up overnight, rarely reflect recent transactions, and often fail to adjust for long-past trades. If you truly want to attack these lines, you need to have a way to predict season-long runs scored and runs allowed.

It is fairly easy to estimate runs scored. I use player projections from the Bill James handbook, and feed that data into a “run estimator;” there is an excellent formula in “The Hardball Times Baseball Annual,” page 214 in the article by Colin Wyers. It is a little trickier to estimate runs allowed. Again, I use the pitcher projections from Bill James, and use the run estimator after converting the pitcher stats into likely singles, doubles and other inputs into the run estimator.

Putting in this kind of work to attack season wins might seem daunting, but almost everything carries directly over into handicapping individual games as well. And, once in awhile you find some real gems as well – like Arizona Under 86½ wins (-120) (I think this is the strongest play in the NL).