clock menu more-arrow no yes

Filed under:

Astros Saber Update: Jarred Cosart & Brett Oberholtzer

New, 34 comments

Astros Sabermetrics: An Updated View of Young Astros' Starting Pitchers Cosart and Oberholtzer

J. Meric

Two weeks ago, we analyzed the sustainability of the young Astros' starting pitchers' performance. Obviously, this is a small sample, but we may identify some initial indications of likely positive or negative regression in the starting pitchers. Rookies Jarred Cosart and Brett Oberholtzer, in particular, exhibited amazing run prevention numbers. Both appeared to be prime candidates for regression which would increase their ERA.

At the time I concluded: Although both Cosart and Oberholtzer have gotten off to similar great beginnings to their careers as starters in the majors, Oberholtzer's performance seems more nearly sustainable based on the analysis of x-BABIP and plate discipline stats.

Checking in two weeks later, we can see that both Cosart and Oberholtzer have experienced some regression in key areas. (Here is a link to basic stats for Astros' starting pitchers.) For example, the very low HR/Fly rates have begun to regress in the direction of sustainability (Cosart from 5.2% to 7.4% and Obie from 2.4% to 3.2%). Cosart's HR/fly rate may be close to stabilizing, but Oberholtzer's HR/fly rate still is not sustainable. Cosart's BABIP, which was an unsustainable .231 has regressed to .253. It's interesting to note that both pitchers' ERA are still quite low, even with some regression in HR rates, peripherals, and BABIP.

One way to examine their performance is to compare their ranking with the 27 rookie starting pitchers in the majors this year. Oberholtzer and Cosart are No. 1 and 2 in ERA, at 1.98 and 2.13, respectively. While this is a credit to their run prevention results, the downside is that Cosart and Obie are 27th and 24th respectively in margin of ERA over FIP. This suggests the liklihood of continued upward regression in ERA. Oberholtzer's FIP of 3.23 still is excellent, ranking 4th among rookie starters, while Cosart's 4.24 FIP is just respectable, ranking only 18th. Cosart ranks last, and Oberholzter ranks 22d in K/9 among rookie starting pitchers. Oberholtzer compensates for the low strike outs with a No. 1 ranking in BB/9 at 1.54. Cosart's BB/9 has been gradually regressing to a lower number, but still ranks next to last (after Brad Peacock) at 4.75. Given Oberholtzer's solid results on advanced pitching metrics, FIP, x-FIP, and SIERA, the lefty appears better positioned than Cosart to maintain a superior ERA in the face of regression.


Which pitches are producing exceptional run prevention for Cosart and Oberholtzer? Cosart's cutter and Oberholtzer's change up are pointed to most often by broadcasters. But Oberholtzer's 2-seam fastball may be his most effective pitch, ranking as the best 2-seamer in run value among rookie starters and 20th among all starting pitchers. Cosart's cutter is the most effective cut fastball in run value (by a wide margin) among rookie starters and exhibits the 6th best run value among all starting pitchers' cutters.

Among all starting pitchers, Cosart has the highest velocity cutter (94.5 mph) and the 9th highest velocity on a 4 seam fastball (94.6 mph). Oberholtzer's change up is top 5 and top 25 in horizontal and vertical movement. Oberholtzer also ranks 8th in vertical movement on his two seam fastball.


In the early going, Cosart's cutter may be the most notable pitch among the young Astros' starters. After all, Evan Longoria said that Cosart had the best cutter he had seen since Mariano Rivera. In the earlier thread on this topic, the question aroise whether cutters might enable pitchers to produce a lower BABIP. Responding to that discussion, I examined some of the characteristics of starting pitchers who rely heavily on cutters.

36% of the pitches thrown by Cosart are cut fastballs, which is the 3d highest usage of the cutter among all starting pitchers (Brandon McCarthy and Dan Haren are the only pitchers with more reliance on the cutter). 21 starting pitchers threw the cutter with at least 20% frequency in 2013. My analysis is focused on these pitchers.













The table, above, compares the starting pitchers with high cutter usage to all starting pitchers in 2013. Because the comparisons are based only on one year, the sample size limitations make the results only suggestive in nature.

The average BABIP for cutter-reliant pitchers is lower than the average for the overall pool of starting pitchers. The difference in means is not large, but it does support the hypothesis that heavy reliance on the cut fastball may suppress BABIP. The high cutter percent group probably includes a wide range of pitch quality, which likely diminishes the consistency of the relationship between cutter usage and BABIP.

Similarly, the high cutter percent group produces a somewhat higher groundball rate. This result is expected and logical, since the cut fastball is intended to produce weak contact.

The final comparison is ERA minus FIP. The issue we address here is whether the weak contact from heavy cutter usage produces a greater liklihood of an ERA less than FIP. A negative E-FIP means that ERA is less than the FIP. The average for heavy cutter use pitchers is negative, while the average E-FIP for all starting pitchers is positive, meaning that FIP is ordinarily less than ERA. The results suggest tthat the cutter may help produce ERA results below FIP. Although this comparison does not meet standard tests for statistical confidence, the difference in means has a confidence level of approximately 68%.

Not surprisingly, the E-FIP for the cutter group is correlated with BABIP (R-squared of .32), implying that the differences in BABIP among cutter pitchers explains a substantial part of the difference between the pitchers' ERA and FIP. This is consistent with the notion that cutter usage enables weak contact by the batter, which in turn produces more outs on batted balls.

Have you been surprised by Oberholtzer and Cosart this year?

More from Crawfish Boxes: