Betting On Baseball Over-Unders

Jonah Keri:

But a few years ago, a friend introduced me to the magical world of baseball over/unders. All I knew about baseball betting at the time was that teams were given certain odds of winning the pennant and/or the World Series. Always seemed like a sucker bet. The heavy favorites and fan favorites (often the same team, often the Yankees), would be given extremely unattractive odds to win. You could also put a C-note on a gigantic long shot, fantasize what you'd do with a hundred grand if your miracle bet paid off, and never see that picture of Ben Franklin again.
Over/unders were different. This was close to a straight-up bet, minus the vig the sports book or online gaming sites would usually take; if you looked hard enough, you could get even odds or positive odds on a bet. In essence, all that mattered was your ability to outsmart the line maker and, in a way, the general public. Casinos, like online betting sites, will set lines to attract maximum betting action. They're playing on everything from player and team reputations to the splashiness of offseason moves to the popularity of certain teams (often the Yankees and Red Sox, though Los Angeles-area teams seem to get weird betting lines often, possibly due to the ease with which you can hop over to Vegas from L.A.). For any baseball fan who thinks batting average was a quaint idea that stopped being useful 30 years ago, it's a golden opportunity to put your money where your stat-stuffed mouth is.
One of the most popular advances in the sabermetric community over the past decade has been the proliferation of projection systems. Baseball Prospectus's PECOTA, Dan Szymborski's ZiPS, Bill James's brand of projections, as well as CAIRO, Oliver, Marcel, and the no longer publicly available CHONE all crunched players' past results, combined those stats with age, injury data, and in some cases scouting information, and spit out forecasts for every player. (For more on how these systems work, see this writeup by FanGraphs' Steve Slowinski.) By combining all those player results on each team then adjusting for strength of schedule, you could build standings projections for all 30 teams. By combining those team win predictions with your own additional analysis, you could find and exploit inefficient betting lines, and make money off them.
I was hooked. And for the first couple years, I did really well. Last year … not so much. My highest-confidence pick, by far, was the Phillies under 97 wins. Chase Utley's injury frightened me, as did the team's aging core and the generally extreme nature of the line: Betting on a team to win damn close to 100 games (or lose that many) is usually asking for trouble. Philly won 102.

I go about the over-under bets more simply.  I throw a number out to Cubs Dad, or ask for one from him, then we negotiate a bit, then I usually take the under.  Last year, I took the over on the Reds for 89.5 wins, and was eliminated by about September 15.  This year, I'm taking the under on 87.5, so the Reds will probably win 90 or more.  That's fine with me.