Often times, when looking at a pitcher’s peripherals, we’ll focus on pitch outcomes. How often do they get swinging strikes, or looking strikes, or strikes overall? Metrics such as these will usually correlate with a pitcher’s strikeout and walk rates, which in turn will generally predict their production as a whole. In some cases, though, we need to look at how well a hurler performs in certain situations. If a pitcher can’t sequence correctly, all the peripheral strength in the world won’t matter.
Matt Garza’s 2015 struggles are no secret. He utterly collapsed, posting the worst ERA (5.63) and DRA (5.33) in his career as a full-time starter. I’ve argued that he shouldn’t start for the Brewers in 2016, and while some may reasonably disagree with that #hottake, Garza undeniably has a shaky future at the major-league level. A large part of that stems from his exploding walk rate: After walking 7.3 percent of the batters he faced from 2010 to 2014, Garza issued free passes at an 8.6 percent clip in 2015, which would have been the 11th-highest in the majors if he’d qualified for the ERA title. Without control on his side, Garza looks pretty hopeless.
But beneath the surface, Garza didn’t get that much worse last season, or worse at all. According to BP’s PITCHf/x metrics, he threw 47.2 percent of his pitches in the strike zone, compared to a lifetime Zone rate of 46.3 percent. Even though he didn’t amass many chases — his O-Swing rate fell to 29.7 percent, from 30.9 percent for his career — he still managed to avoid balls. Per Baseball-Reference, 65.3 percent of his 2015 pitches went for strikes, a rate that topped his career mark of 64.0 percent. Garza also had a higher in-play-strike rate (32.4 of his 2015 strikes, compared to 29.5 percent for his career), which helped to keep plate appearances curt.
So what should Garza’s walk rate have been? In 2013, Rotograph’s Mike Podhorzer created an expected walk rate equation. It uses three statistics — strike rate, in-play-strike rate, and strikeout rate — to tell how many bases on balls a pitcher theoretically should have had. With an adjusted r^2 of .752 in the sample Podhorzer analyzed, this equation seemed to excel at that task. And for Garza, well, it paints a pretty clear picture:
| Rank | Name | BB% | xBB% | Diff |
|---|---|---|---|---|
| 1 | Matt Garza | 8.6% | 6.2% | 2.4% |
| 2 | John Lackey | 5.9% | 3.5% | 2.4% |
| 3 | Bartolo Colon | 2.9% | 1.0% | 1.9% |
| 4 | Shelby Miller | 8.5% | 6.6% | 1.9% |
| 5 | Mike Leake | 6.3% | 4.8% | 1.5% |
| 6 | Charlie Morton | 7.3% | 5.8% | 1.5% |
| 7 | Tim Hudson | 7.1% | 5.7% | 1.4% |
| 8 | Williams Perez | 9.9% | 8.7% | 1.2% |
| 9 | Matt Wisler | 8.4% | 7.2% | 1.2% |
| 10 | Carlos Martinez | 8.3% | 7.2% | 1.1% |
141 pitchers accumulated at least 100 innings last season. These are the top 10 in bad luck, as measured by difference between walk rate and expected walk rate. Among all of them, Garza tops the list; this metric suggests he should have walked a mere 6.2 percent of his opponents. So this means that, if his poor luck turns around in 2016, he’ll limit his free passes and improve overall — right?
Well, no. Misfortune does bear some of the blame for Garza’s inflated walk rate last season, but it can’t account for everything. We need to look a little bit deeper, because for all the benefits of that expected walk rate equation, it ignores one crucial factor: sequencing. That’s why, in 2015, Alex Chamberlain followed up on Podhorzer’s work, creating a new expected walk rate equation.
How did Chamberlain attempt to correct the flaw? In addition to strike rate, in-play-strike rate, and strikeout rate, he added 3-0 count rate: the percentage of all plate appearances that, at some point, contain a count of three balls and no strikes. While this doesn’t perfectly model sequencing — as Chamberlain explained, its multicollinearity isn’t especially strong — it does pump up the adjusted r^2 to .821. 3-0 count percentage clearly has some predictive value, and for Garza, that isn’t a good thing.
Here, again, are the ten biggest underperformers. Garza appears on this list too, but this time with considerably more luck:
| Rank | Name | BB% | xBB% | Diff |
|---|---|---|---|---|
| 1 | John Lackey | 5.9% | 3.7% | 2.2% |
| 2 | Bartolo Colon | 2.9% | 1.1% | 1.8% |
| 3 | Shelby Miller | 8.5% | 7.0% | 1.5% |
| 4 | Jorge De La Rosa* | 10.2% | 8.8% | 1.4% |
| 5 | Jon Niese* | 7.1% | 5.9% | 1.2% |
| 6 | Tim Hudson | 7.1% | 5.9% | 1.2% |
| 7 | Garrett Richards | 8.8% | 7.7% | 1.1% |
| 8 | Matt Garza | 8.6% | 7.5% | 1.1% |
| 9 | Trevor Bauer | 10.6% | 9.6% | 1.0% |
| 10 | Jimmy Nelson | 8.6% | 7.6% | 1.0% |
The new expected walk rate equation dings Garza a lot more, because 3-0 counts defined him last season: He began 6.4 percent of his plate appearances with three straight balls. That’s much higher than his 4.7 percent lifetime 3-0 count rate, as well as the 2015 MLB average of 4.4 percent. Garza didn’t really struggle in 3-0 counts themselves — 60.2 percent of them ended in a free pass, which was below the 62.4 percent major-league baseline — but the fact that he got in so many of them held him back.
For 2016, Garza has a number of things to work on if he wants to earn his keep. He’ll have to arrest his declining strikeout rate, which really hurt him last year, and he’ll need to keep the ball in the yard. Making progress on his walk rate would also aid his case, and to do that, he must stop himself from falling behind hitters so often. Garza still throws strikes and keeps at-bats short; if he can just sequence things correctly, perhaps his walk rate will improve to its earlier form.
1 comment on “Matt Garza and the Importance of Sequencing”