World Series Probability Added: Pitchers

Previous entries:
Primer
The Full Story
Batters

In a series with a 16-7 game and an 11-10 game, and one in which the big moments were dominated by the hitters, it can be easy to forget about the pitchers. And although the performances of Chris Carpenter and Derek Holland pale behind those of David Freese and Lance Berkman, they deserve to be remembered.

On the other side of the coin, Jason Motte, who was so excellent in the first two rounds of the playoffs, was bailed out by his teammates in the final two games. As poorly as Motte pitched, it was the entire Rangers bullpen that collapsed, with Scott Feldman, Neftali Feliz, Alexi Ogando, and Mark Lowe all coming up well negative for Ron Washington and the Rangers.

Observe, the leaders and trailers in pitcher series probability added for the 2011 World Series:

Click to see the fully functioning visualization and follow the jump for more.

The expected leader among pitchers was probably Chris Carpenter, but although he did have a good series, his three appearances couldn’t overtake combination of a brilliant Game Four start and a solid Game Six relief appearance from Derek Holland. As mentioned above, it really wasn’t a series with pitching to remember. Aside from Holland and Carpenter, only Jaime Garcia, whose excellent Game Two start was wasted thanks to a Jason Motte blown save, compiled a series probability added greater than 10%.

With great hitting performances must come poor pitching performances, and the largest impact of these poor performances came from Motte, who blew the lead in Game Two to cost the Cardinals 14.1% series win probability and allowed the go-ahead home run in Game Six to Josh Hamilton, costing the Cardinals 19.9% in series probability, until, of course, Lance Berkman and then David Freese came to the rescue.

It was Scott Feldman at the receiving end of Lance Berkman’s game-tying hit in extras of Game Six, costing the Rangers 11% of Feldman’s eventual -24% total series probability added, to go along with failings in Games Three and Seven. Coincidentally, these last two came in Matt Harrison starts, and Harrison was the worst starter in the series, allowing six earned runs in 7.2 key innings for a series probability added of -22.1%.

The real story has to be the Rangers’ bullpen, touted as a strength after trade deadline additions and through the first two playoff rounds. It utterly collapsed in the World Series, as Feldman along with Alexi Ogando, Mark Lowe, and Neftali Feliz combined for a shocking -69.45% playoff probability.

To see these and other pitchers’ contributions spread out by game, check out this alternate visualization of the pitcher series probability added:

Again, click to see a full-size interactive version.

I was also asked about the impact of errors on the pitcher totals. There were a whopping 14 errors committed in the seven games of this World Series, and combined they had an overall impact of about 30% series probability. However, the data we have doesn’t separate the impact of the hit from the error on plays like Elvis Andrus‘s single in the ninth inning of Game Two. It should be noted that a large majority of the SPA on this play (as well as the second largest, Yadier Molina’s single-and-error in the second inning of Game Five) comes from the hit, and not from the error. But for reference, here is a visualization of all 14 errors. By mousing over the bars, you can read the description of the play in question.