MLB Postseason Predictions

Even the unluckiest baseball teams can win at the game’s roulette table on any given night. In baseball, luck exists and shows itself in many ways. A batter can be unlucky and tag a grooved pitch for a screaming line drive right at a fielder, for instance. A fielder can be lucky – like Josh Donaldson was in this video – by booting a ground ball directly into his throwing hand. There’s also a much broader form of luck that exists in baseball, a sort of fortunate series of events¹ known as cluster luck.

Make your postseason predictions on the BSL Forum.

Cluster luck has been recognized for years without having much made of it. Think of any time a pitcher was said to scatter hits and escape a game unscathed: if there are 9 or 10 hits in the box score, most would have expected a short outing by the pitcher giving them up. If they’re strewn across 7 or 8 innings, with just one single every three outs, the pitcher still could have left the game having given up 0 runs and with a win intact. This is lucky: clearly, the opposing batters were able to hit his pitches. The pitcher is simply fortunate that they failed to do so in consecutive at bats. Similarly, a team can notch just one hit in the first inning – say, a home run after an error – and win a game. That’s an unlikely and lucky occurrence for that offense.

This sort of luck might be sustainable for one game or even for most of a season, but it’s very difficult to predict the games in which it strikes and it’s even harder for teams to carry cluster luck on their side from year to year. To strip out the effects of cluster luck, David Smyth created Base Runs (BsR), a dynamic model of team run scoring that takes into account baserunners, advancement of baserunners, outs, and guaranteed runs on both the offensive and defensive sides of the ball.²

To predict the outcomes of each postseason series, including the wild card, I’m going to use Base Runs to strip the luckiness (or unluckiness) from each team’s offensive and pitching capabilities and see how many runs they would score or give up without any cluster luck. Then, I’ll use these new runs for and runs against numbers to define new pythagorean win percentages, now termed W%+. Since the best indicator of postseason success is regular season success, these can be used to predict every postseason series, all the way up to the World Series.

As an added bonus, I’ll be doing the same for individual games in the upcoming ALDS series between Baltimore and Detroit.

Using W%+ to fill in the bracket from here is pretty simple. Oakland would beat Kansas City in the AL Wild Card game, while Pittsburgh beat San Francisco in the NL Wild Card game. Don’t take these to the bank just yet, I’ll have more on the wild card in a minute.

In the AL, the Los Angeles Angels of Anaheim would take care of Oakland while the Tigers and the Orioles flipped a coin over and over until the Tigers just barely squeaked away with a series win. Then, the AL’s LA team would best the Tigers on the way to the World Series.

In the NL, the Washington Nationals would destroy whichever sorry Wild Card team made the mistake of winning just one extra game this season. The Los Angeles Dodgers would beat the Cardinals, then the NL’s LA team would lose to the Nationals in the ALCS, with the Nationals headed to the World Series.

In a cross-country matchup, the Nationals would beat the Los Angeles Angels in the World Series, bringing a big trophy back to the east coast (but just a little too far away from the Warehouse for my tastes). So there you have it – your eventual World Series winner is the Washington Nationals, who were somehow unlucky and still won 98 games this season.

Back to the Wild Card matchups: the W%+ you see above is based on team performance, which isn’t really what you’re getting in Game 163 the Wild Card game, at least not on the mound. Each team will trot out their best single-game pitchers in hopes of suppressing the other team’s offense and ensuring themselves a win. Oakland will start former Red Sock Jon lester, while Kansas City sends Big Game James Shields to the mound. According to Major League Baseball, Madison Bumgarner will take the hill for the Giants while Edinson Volquez will start for the Pirates. To get a more accurate prediction regarding the future of the Wild Card game, we have to factor in the skill of the one starting pitcher they’ll use. The best way to do so is to go through the same process as if that one starter pitched every game for their teams; certainly a Kansas City Royals team with 5 James Shields clones would be better than the real Royals team. For the purposes of this exercise, I adjusted all four pitchers’ stats to 1,450 innings, which is about the average that a full team pitches over the course of a season. Here’s the same table, but adjusted for the Wild Card starters:

Adjusting for luck and projected starters, Oakland still beats Kansas City, but putting Lester on the mound makes Oakland fare more likely to win than before. Baumgarner has been excellent this season, but it’s not enough to topple the Pirates, who have had horrendous luck this season. There’s no reason that their luck should follow them into October, so in the NL Wild Card game, I expect the Pirates to beat the Giants in a one-game playoff.

Enjoy the postseason, baseball fans!

1. Sorry, Lemony Snicket

2. Tom Tango considers Base Runs to model the run scoring process better than any other run estimator.

Patrick Dougherty

Patrick was the co-founder of Observational Studies, a blog which focused on the analysis and economics of professional sports. The native of Carroll County graduated with a Bachelor’s degree in Economics from Loyola University Maryland. Patrick works at a regional economic development and marketing firm in Baltimore, and in his free time plays lacrosse.

Twitter