Kendall Graveman was the Astros’ most valuable addition to the bullpen at the trade deadline. He gave up a grand slam in the recent series in Seattle, and in his next game in San Diego he came close to disaster, loading the bases and leaving the game after two outs and 26 pitches. Fortunately, the outcome turned out fine in the San Diego game, but the Astros lost the game in Seattle. Graveman threw 56 pitches in those two outings.
Maybe these are just hiccups. Every pitcher has a bad inning on occasion. But let’s take a closer look at Graveman’s performance after he was traded to Houston.
Graveman’s Performance With Seattle and Houston
As Seattle’s closer, Graveman’s pitching line was perhaps the best among baseball’s relievers. However, Graveman’s performance since joining the Astros has declined from that high water mark. In first two weeks following the trade, Graveman’s pitching was pretty much in line with his Seattle work. Since August 22, his performance has been more problematic. The basic pitching stats for Graveman’s performance in Seattle, Houston, and the specific period August 22 - Sept. 3 are shown below:
(Seattle / Houston / 8-22 thru 9-3)
- ERA: 0.82 / 4.15/ 9.00
- BB/9: 2.18/ 3.46/ 9.0
- K/9: 9.27/ 11.08/ 10.4
- BABIP: .176/ .291 / .308
- SIERA: 2.94 / 3.10 / (not available)
- RE24: 10.36 RS above avg./ 0.6 RS above avg./ -2.35 RS below average
Clearly, Graveman’s performance has declined since the trade from Seattle. Although an ERA above 4 is not great, the difference between the overall Seattle and Houston pitching lines, so far, is not that concerning. Moreover, as shown by the SIERA stat, the ERA differential between the Seattle and Houston stints probably overstates any true difference in performance. On the surface though, the stats covering the six games starting on August 22 are of greater concern. Still, to keep it in context, this period covers only five innings.
The prime suspect for this decline is regression to the mean. Batting average on balls in play (BABIP) is often a sign that regression is, or is not, expected. Statistically, pitchers cannot sustain a BABIP above or below a “normal” level. Graveman’s .176 BABIP during his relief appearances in Seattle is not sustainable. In other words, Graveman would be expected to give up hits at a higher rate than he did with the Mariners this year. The “normal” expected BABIP is not always clear. Generally a .290 - .300 BABIP is expected, on average, for starting pitchers. There is evidence that elite closers can sustain a somewhat lower BABIP. Some of the best closers currently have a BABIP in the .250 - .290 range. (As an illustration, Mariano Rivera and Billy Wagner have career BABIP of .259 and .261, respectively.) So, it’s not surprising to see Graveman’s BABIP regress to .291 in Houston. If “normal’ for Graveman is similar to Ryan Pressly’s current .250 BABIP, perhaps he will see downward regression in the future.
The regression notion is also supported by Graveman’s miniscule 0.82 ERA in Seattle. All of the current relievers in the majors with an ERA below 2 have an expected ERA (x-ERA) higher than their actual ERA. Graveman’s current season ERA of 1.76 is below his x-ERA of 3.44. Again, this implies that maybe this is just the bumpy ride of regression.
Examining Graveman’s peripheral stats, we can see that Graveman actually has struck out batters at a higher rate for Houston than he did in Seattle. On the other side, Graveman’s walk rate is worse with the Astros than his Seattle stint. Narrowing our concern to the period starting on August 22, it is evident that walks have been a significant issue in recent games. A 10.6 K/9 is fine, but the 9.0 BB/9 is going to create some difficulty. Command, in particular, was an issue in the most recent two games.
An interesting question is whether the Astros took into account potential regression when they valued Graveman for trade purposes. Did they trade for the pitcher with a sub-1 ERA, believing that represented his value going forward? Graveman is a good relief pitcher, but it was unrealistic to expect him to sustain a sub-.200 BABIP.
The risk of regression is not limited to the Astros. For example, Craig Kimbrel was an even more highly valued relief pitcher at the trade deadline. Kimbrel also carried a sub-1 ERA and a low .203 BABIP before the White Sox acquired him. Kimbrel has a 6.57 ERA with the White Sox, compared to his previous 0.49 ERA with the Cubs. Kimbrel’s BABIP with the White Sox is .308.
In some sense, regression is an inherent risk in trading for the highest level performers. It is probable that while the Astros (and White Sox) did expect some regression when they pulled the trigger on the deadline deals, these recent results were still even more than they (and Astros fans) braced themselves for.
The encouraging point is that Graveman’s velocity is fine. In fact, his fastball velocity is marginally higher in Houston than Seattle (96.8 vs. 96.3). The same is true for his sinker. His slider is about 1 mph slower in Houston than Seattle, but (given the pitch) I doubt that difference is significant. The sinker is Graveman’s main pitch, and he increased his usage of the pitch from 67% to 70% in Houston.
The downside is that his sinker has been less effective in Houston—particularly in recent games. The pitch value per 100 is +3.45 for Seattle and -0.15 for Houston. The slider, though, has been more effective in Houston. The wOBA on the sinker in the San Diego game was .465, due to poor control and a walk.
Batters are swinging outside the zone less in Houston (25% vs. 39% for Seattle). It’s possible that diminished command of the sinker has made hitters less likely to chase.
One question lurking below the surface is whether Graveman has any physical issues which caused the recent problems. In discussing his decisions on using Graveman, Dusty Baker’s comments to the media last week were vague enough to make one wonder. Baker has pointed out that Graveman is still getting accustomed to the work load of a relief pitcher, instead of his prior role as a starting pitcher. Graveman’s velocity is a positive sign, since injury is often associated with a loss of velocity. Command problems can be associated with physical issues, but it can also be due to delivery mechanics or fatigue.
Most likely the reductions in Graveman performance for Houston are reflective of regression. Hopefully, the recent high pitch innings are isolated instances. If Graveman is experiencing late season fatigue, perhaps he will need more rest. But Graveman probably will rebound as the effective set up man who will bolster the bullpen. However, we shouldn’t expect his run prevention to revert to the 0.81 ERA he posted in Seattle during the first half of the season. He was a good relief pitcher then, but the shiny sub-1 ERA over stated how good he is. He can revert to that same pitching skill level, but with a more realistic expectation.