Jarrod Washburn’s Improvement, Part 2

Part One

Last night, I looked at the macro level of Jarrod Washburn‘s improvement against left-handed batters and pointed out the dramatic increase in swinging strikes that he had achieved against them this season compared to years past. It seemed to be the result of his added two-seamer, as his missed bats on pitches in the 87.5 to 92.5 mph range had noticeably increased.

But is it just that? I wanted to see if there was also anything to be found in the location data and that investigation brought forth these charts, 2008 pitches to LHB on the left, 2009 on the right.

Well, aside from the fewer pitches in the 2009 dataset, these two pictures look near identical. There might be some trace differences, but I seriously doubt any of those could be playing much impact in the huge jump in effectiveness that Washburn has seen to date.

Location shot down, I took one more stab at pinpointing the cause. On the theory that with a new two-seamer in the mix, hitters were seeing a wider distribution of speeds and that was putting them on guard more, I expanded my first graph to include all pitches to left-handers this season compared to 2008

Nope, not really. It is a fairly close match in terms of speed distributions. What’s left? Movement. I do not have a good way to visualize movement data, so apologies for the lack of readability here. I broke pitches into four categories, whether they were swung and missed at or not and their starting speeds since the above histograms demonstrate that Washburn has a break between pitch types around the 85 mph mark. The first number is the average largest deviation, in inches, of the pitch from the straight trajectory*. The second number is the average angle of that break. To visualize that, draw a compass rose, 0 is south, +90 is west and -90 is east. The final number is the sample size.

Pitches missed, 2008, under 85: 10.1, 9.0, 41
Pitches missed, 2009, under 85: 09.9, 8.8, 25

Pitches !missed, 2008, under 85: 9.5, 6.2, 271
Pitches !missed, 2009, under 85: 9.9, 7.9, 94

Pitches missed, 2008, over 85: 4.7, -12.6, 11
Pitches missed, 2009, over 85: 4.7, -16.8, 23

Pitches !missed, 2008, over 85: 4.8, -15.6, 381
Pitches !missed, 2009, over 85: 4.7, -16.9, 172

Washburn’s slurvey breaking pitches certainly are not moving dramatically differently than before. Among his swinging strikes, the two rows are insignificantly different. There is a big more break and run shown overall as the second pairing shows, but nothing huge. Now granted we have a small sample, but we do have what looks like a potentially meaningful change in the third and fourth pairings. The group of fastballs and fastball-type pitches all seem to be moving more in on the left-handed hitters this season and particularly among pitches that are missed.

Still, I am not anxious to jump to any conclusions on that. We are talking about a four compass degree of difference in that third pairing, which does hint toward the increased use of a two-seamer, but interestingly enough, the length of the breaks are still similar when two-seamers are known for generating more movement.

*Straight trajectory is thus: take the two points where the pitch is released and where the pitch crossed the home plane. Connect the two points with a line.



Print This Post



Matthew Carruth is a software engineer who has been fascinated with baseball statistics since age five. When not dissecting baseball, he is watching hockey or playing soccer.



Sort by:   newest | oldest | most voted
Donovan
Guest
Donovan

As an M’s fan, I appreciate your delving deep into the reasons for Washburn’s success this year. But after reading the pieces, I’m left inferring that while he’s had greater success against LH in terms of swings & misses, there doesn’t appear to be a reason for it.

Is it possible this is a sample size issue as well? Looking at your data, he appears to have thrown about 275 pitches to LH this year, resulting in 17 more swinging strikes than last year’s average. While in percentage terms this is a notably higher result, it does translate into just one and a half more swinging strikes against LH per start.

Much like hitters start out hot/cold and then regress to the mean, does your analysis of the pitch data lead you to the conclusion that the most rational explanation is that Washburn will regress to the mean against LH?

Thanks again.

wpDiscuz