Posted in Graphs, MLB, Philadelphia Phillies, Sabermetrics | Print | 4 Comments »
If it seems like the Phillies’ starting pitching hasn’t been as good this year as it was last year, it’s because it hasn’t. The fearsome foursome of Roy Halladay, Cliff Lee, Cole Hamels, and Roy Oswalt contributed to a league-best 2.86 ERA — nearly a half-run better than the next-best team, the San Francisco Giants. This year, that ERA rose to 3.81, only the sixth-best mark in the league. It’s been a rough year: Halladay had an injury problem, Lee has dealt with incredibly bad luck, and Vance Worley‘s season recently ended with elbow surgery. It hasn’t all been bad news, though, as Kyle Kendrick has had two incredibly good runs of pitching and Tyler Cloyd has looked mostly good since being called up recently.
It is generally difficult to compare something so broad as “starting pitching” from one year to the next, but we can get a rough idea using game score. While far from a perfect metric, it does give us an idea as to how the pitching has changed between 2011 and ’12. The following chart shows the frequency of Phillies starters’ game scores by buckets, with 50 being the average.
Percentage-wise, the 2012 Phillies had more “elite” pitching performances, game scores of 71 or higher. Meanwhile, the 2011 Phillies had more “slightly above average” performances, game scores between 51 and 70. The latter matters more because they occur more often: 64 of 140 games fell between 51-70 this year, and 70 of 162 occurred last year. Meanwhile, only 23 games reached 71 or higher this year, and only 46 did last year.
In terms of individual performances, all four of Halladay, Lee, Hamels, and Worley saw their average game score decline. The now-departed Joe Blanton saw a modest increase, Kendrick stayed about the same, and Cloyd has been about as good as Oswalt was last year, though in 20 fewer starts.
|2011||Average GS||Starts||St Dev|
|2012||Average GS||Starts||St Dev|
Another interesting item to look at is the standard deviation of each pitcher’s game score in both seasons. The standard deviation tells you the spread of data around the average — the larger the number, the more volatile the pitcher was overall. For instance, Halladay’s average game score in 2012 is 55 with a standard deviation of 17, so roughly 68% (why 68%?) of his starts fell between a game score of 38 and 72. Indeed, 21 of his 32 starts (66%) were between those two numbers.
From 2011 to ’12, Lee, Hamels, and Worley saw their standard deviation shrink along with their average game score, so not only were they worse on average, but their starts overall were more frequently mediocre, rather than sometimes elite. Put another way, Lee’s 2011 standard deviation of 19 is partially due to eight of his 32 starts producing a game score of 80 or better. This year, only one of his starts — a memorable one — was 80 or better.
I don’t mean to imply that more volatility in starters is always a good thing. Cy Young favorites in their respective leagues, R.A. Dickey has an average game score of 63 with a standard deviation of 19, while David Price has an average game score of 61 with a standard deviation of 16, for example. However, because runs cannot go below zero, it is more rewarding to post a game score 19 above your average rather than 19 below, since there’s almost no change in win expectancy if you allow five runs instead of six, as opposed to a huge swing in win expectancy if you allow one run rather than two.
The Phillies’ starting pitching problems have been rather easy to diagnose this year: age and injuries, mostly. But it’s also true that the staff as a whole declined and was, perhaps, too consistent.