Tuesday, July 19, 2011

Living Up To Expectations

Why Detroit is primed for a fade and Milwaukee won't make the playoffs

Imagine a hypothetical 3-game series between two baseball teams. Team A wins the first game in a rout, 13-0. Team B recovers to win each of the next two games 1-0.

At that point, Team B obviously has the better record at 2-1. Their run differential, however, is -11; they've scored two, and allowed 13. Team A is behind in the standings, but boasts a run differential of +11.

The sample size is too small to draw any reasonable conclusions, but it's safe to say that in this short series, Team B has maximized their run scoring. If they maintained this exact pattern over a 162-game season, they would score 108 runs and allow 702, for a run differential of 594, but their record would be an excellent 108-54.

That's an extreme example that could never happen in major league baseball, but we see the phenomenon on a smaller scale every year. This season, the Detroit Tigers have been among the best in baseball at distributing their runs. With 419 runs scored and 437 runs allowed (-18 differential), you'd expect them to have won roughly 46.5 of their 95 games.* Instead, they've won 50. For the purposes of this post, we'll say Detroit's Wins Above Expectation (WAE) is +3.5.

*I'm using a simple formula that shows expected wins in proportion total runs scored and total runs allowed. The formula is (Runs scored x games played)/(Runs scored + Runs allowed).

On the other side of the table, the Toronto Blue Jays are two games below .500 at 47-49. Their run differential, though, is a respectable +17. Based on their runs scored, they should have won 48.9 games. Toronto's WAE is -1.9.

What conclusions can we draw from these numbers? The most obvious is that Detroit is winning close games, while Toronto is losing them. The stats prove the point- Detroit is 32-19 in games decided by three runs or less, but only 18-26 when games are decided by three runs or more. Toronto, on the other hand, generally wins the blowouts but loses the tight ones. They're 30-35 in three-runs-or-less affairs, and 17-14 when it gets lopsided.

More broadly, we can make some educated guesses. The two factors which logically seem most important in winning close games are opportunistic hitting, and a strong bullpen. We know Toronto is a power hitting team, so it makes sense that they beat up on bad pitchers and win some games by a lot of runs. Their pitching, on the other hand, is 50 runs worse than the best teams in the AL East, so they're vulnerable in games that are close near the end, especially in the division (though the Toronto bullpen is average- 8th in the American League in ERA).

With Detroit, conclusions are more difficult to draw. Their bullpen is the second-worst in the AL by ERA, and they actually score more runs than anyone in the AL Central. They also have the best power numbers in the Central, which means their profile is more fitting of a team like Toronto. So why, with a bad bullpen, are they winning 63% of their close games? There are two possible explanations.

1 - They're getting lucky.
2 - There's something inherent about the Tigers- call it clutch play, or opportunism, or whatever- that allows them to thrive in pressure situations, and which conversely makes them perform poorly when the pressure's off.

If you believe they're getting lucky, Wins Above Expectation can work as a predicting mechanism. If everything gravitates toward a statistical norm, Detroit is due for a correction- they should start losing games at a faster clip, and possibly end the season below .500 and out of playoff contention. Toronto, meanwhile, should start winning, though the pace won't be enough to catch their AL East brethren.

But using WAE to predict future outcomes is an exercise fraught with peril. While no team since the 2007 Arizona Diamondbacks has made the playoffs with a negative run differential, as Detroit is threatening this year, almost every playoff team finishes with a positive WAE. This makes sense, since in theory WAE should depend on opportunistic hitting and a strong bullpen, two critical components of a successful team. Here are the WAE totals for last year's playoff teams:

Tampa Bay: +6.5
New York: +5.3
Minnesota: +6.9
Texas: +5.3
Philadelphia: +8.4
Atlanta: +3.5
Cincinnati: +3.2
San Francisco: +3.8

You have to go back to the 2008 Los Angeles Dodgers to find a playoff team with a negative WAE- the Dodgers registered a -0.1 in a down year for the division. That same year, in Anaheim, the Los Angeles Angels posted a +15.3 WAE, the highest total for any team in the past four seasons.

Let's take a look at this year's WAE numbers by team, from best to worst. Bold italics denote a team currently in playoff position, and all numbers are rounded to the nearest tenth. Run differentials are in parentheses.

1. PHI: +6.2 (+88)
2. SFG: +6.2 (+12)
3. BOS: +4.5 (+113)
4. ATL: +4.4 (+61)
5. DET: +3.5 (-18)
6. TEX: +3.3

7. ARZ: +2.9 (+11)
8. NYY: +2.7 (+114)
9. CLE: +2.3 (+6)
10. LAA: +1.9 (+16)
11. PIT: +1.7 (+12)
12. TBR: +1.3 (+35)
13. STL: +1.1 (+26)
14. MIN: +0.8 (-65)
15. WAS: -0.1 (-10)
16. FLA: -0.3 (-30)
17. NYM: -0.6 (+12)
18. CWS: -1.1 (-7)
19. TOR: -1.9 (+17)
20. SEA: -2.0 (-33)
21. CIN: -2.1 (+29)
22. COL: -2.4 (-1)
23. BAL: -2.7 (-98)
24. LAD: -3.2 (-35)
25. MIL: -3.7 (-12)
26. SDP: -4.5 (-36)
27. CHC: -4.6 (-98)
28. OAK: -4.9 (-15)
29. KCR: -6.7 (-52)
30. HOU: -10.3 (-110)

As expected, WAE correlates pretty closely with actual winning- excelling in close games generally means you'll have a good record. The anomalies are where things get interesting. Detroit's success in spite of a bad bullpen and a negative run differential has already been mentioned. But the real strange bird here is Milwaukee. We already knew the Brewers were an odd team, but now they're threatening to make a playoff run with a negative WAE. That would be a rare feat, last accomplished by the narrowest of margins in 2008. The Brewers, though, are doing it in style, with a drastic -3.7 WAE.

Despite the difficulty in using these numbers to predict how the season will end, I'll go out on a limb and draw two conclusions:

1 - With a power lineup and a weak bullpen (third worst in the NL), it's no surprise the Brewers lose a lot of close games. That trend should continue as the season plays out. What is surprising is that they're atop the NL Central. With just under 70 games remaining, it's doubtful they can sustain their position. If you're looking for a dark horse candidate in that division, check out the Cincinnati Reds. They score the most runs in the NL, and their bullpen is 6th-best in the league. Sure, their starters are pretty bad, but their second half schedule is great- outside of the division, they only play 10 games against teams with a record above .500, and 26 against teams at .500 or worse. Their -2.1 WAE seems like a fluke of bad luck, while Milwaukee figures to continue losing the close ones.

2 - Detroit, with their second-worst bullpen ERA, can't possible continue to win 63% of three-runs-or-less games. In their division, Cleveland's offense scores almost as many runs, and their pitchers are better across the board (with a bullpen 3rd best in the AL). The Indians' +2.3 WAE seems a lot more realistic than Detroit's +3.5, and they're a good bet to outpace their rivals in the final two months.


  1. http://www.youtube.com/watch?v=AJeIc0F2zPY

    Seth Curry drops 35 at NC Pro-Am! 7/19/11

    Feel free to post this to your blog.

  2. Reason Tigers are winning close games: Victor Martinez. Look at his numbers with two outs, two strikes, or RISP--all clutch.