….also comparing the young starters, Hunter Brown and Spencer Arrighetti
TThe Astros’ bullpen has experienced a few—shall we say—hiccups over the last few days. (Thankfully, the relief corps came through with flying colors in Tuesday night’s game.) While we know the results have sometimes been disappointing, it may be helpful to look below the surface. Let’s turn to some of the Statcast metrics from Baseball Savant.
In addition to looking at the relievers, I’m also interested in the development of the two young starting pitchers in the rotation, Hunter Brown and Spencer Arrighetti. The Astros’ rotation has been battered by injuries and TJ surgery. The continued improvement of these two pitchers is essential to the Astros in both the short and long term.
Therefore, I’ve selected some of the more incisive Statcast metrics for both the relief staff, as well as Brown and Arrighetti. These metrics are expressed in Baseball Savant’s familar percentile of MLB players. This shows where the players rank among all players for each measure, with higher percentiles equating to higher level performance. In a separate calculation, I evaluated the relationships among the metrics by examining the R-square or coefficient of determination for the sample of all Astros’ pitchers.
Baseball Savant
Relief Staff
- The ability to strike out batters is probably what we want the most from a bullpen. The relievers’ job is situational in nature, and the situation sometimes requires a K. The more granular “whiff rate” is more predictive of strike outs than the simple K rate. Called strikes are not the most predictive part of strike out rates, because it involves a larger dose of random variation. Therefore, the swings and misses may provide the more stable and predictive measure of the ability to strike out batters. Statistically, the whiff rate explains about 30% of the Astros pitchers’ x-ERA.
- Josh Hader is in rarified air—the 99th percentile of whiff rate. Bryan Abreu follows closely in the 92nd percentile for whiff rate. Ryan Pressly’s 79 percentile whiff rate is also very good. These high whiff rankings are why these three pitchers are considered the high leverage relievers who cover the 7th – 9th inning. Unheralded Tayler Scott is approaching the same whiff territory with a 64 percentile ranking. Rafael Montero, however, is at the opposite end, falling into the bottom 10% of pitcher whiff rates. This is rather concerning, because it could be indicative of a significant decline in skill. Seth Martinez’s whiff percent falls in the lower mid-ranking among MLB pitchers with a 42 percentile. The middle ranking is fine for a middle reliever but probably means Martinez is less well suited for high leverage situations.
- Expected ERA (x-ERA) should be more reflective of run prevention skill than simple ERA, because it strips out some of the luck aspects of ERA. Hader again leads the way with his x-ERA among the top 13% of pitchers. Scott, Abreu, and Martinez also have fine x-ERA results. But here we start to see some concern for Pressly’s results, which are at the 51 percentile of x-ERA. That is a middling x-ERA. Again, we see a red flag for Montero, whose 22 percentile x-ERA ranking places him in the lower quartile of pitchers.
- Barrel% reflects a batted ball with optimal combination of launch angle and exit velocity. You would expect a lot of barrels become home runs, or at least extra base hits. This is where Pressly and Hader have run into late inning damage, with barrel percentiles of 21% and 43%, respectively. Montero has also been barreled up quite a bit. Scott and Martinez have been very effective at avoiding barrels. The good news for Pressly and Hader is that barrel percent tells us what happened, but is not predictive of future performance by the pitcher. That is to say, barrels are driven by the batter’s skill and can be predictive on the batter’s side, but quite volatile and random for the pitcher. The R-square analysis indicates that barrel percent has minimal explanation of x-ERA.
- The expected xwOBA and xSLG stats are reflective of the quality of contact and removes some of the luck context from wOBA and SLG. Statistically, x-SLG explains 65% of the Astros pitchers’ x-ERA. The x-SLG percentile for all the relievers except Montero and Pressly is fine—and better than fine for Scott and Abreu. Montero and Pressly rank in the 30 and 38 percentile, respectively, for x-SLG. And the xwOBA is pretty good for all of the relievers except Montero.
- The expected OBP (x-obp) can be viewed as an expected stat which provides information similar to WHIP. Statistically, the x-OBP explains 56% of the Astros pitchers’ x-ERA. The x-OBP is far more predictive of their x-ERA than either BB% or x-BA. (The Astros pitchers’ x-ERA is not meaningfully related to the pitchers’ BB%.) The relievers’ rankings for x-OBP are quite good—and in the case of Hader, excellent—except for Scott, Montero, and Abreu. Abreu is generally good on the Statcast metrics, but he tends to allow too many base runners.
- Based on this review, I have some concerns about Montero’s likely performance in the future. He ranks at the low end for all of the metrics. There may be some warning signs for Pressly, but his percentile rankings for the more predictive metrics (x-OBP and Whiff%) are not bad—and his x-SLG and barrel percent probably will regress to improved levels.
Hunter Brown and Spencer Arrighetti
Hunter Brown and Spencer Arrighetti appear to be fixtures in the Astros’ rotation for the indefinite future—or at least until Lance McCullers Jr. and Luis Garcia return from surgery. Both had a rocky April, but have been continuously improving through May and June. Over the last 30 days, Arrighetti’s ERA is 3.12 and Brown’s ERA is 3.10. Brown’s WHIP over that period is below 1, and Arrighetti’s WHIP of 1.35 is down from his seasonal WHIP of 1.62.
- Examining the table of Statcast metrics, above, Arrighetti’s Statcast percentile is worse than Brown’s for all but one metric. But that one measure, Whiff%, is fairly important. Arrighetti is ranked in the top one-third of pitchers for Whiff%. Generally the percentile rankings for Brown and Arrighetti on most of the measures are still showing the effect of early season meltdowns.
- Purely from a stuff standpoint, Brown currently is more highly ranked than Arrighetti. For example, Brown Stuff+ ranking is 104 compared to Arrighetti at 83. Despite the Stuff+ difference, Arrighetti has a higher spin rate on his fastball and curveball (60 and 81 percentile, respectively).
- According to Stuff+, Brown’s two highest ranked pitches are his slider and knuckle curve, while Arrighetti’s two highest ranked pitches are the change up and curve ball. This is based on the physical characteristics of the pitches.
Hopefully Brown and Arrighetti can continue the performance they have shown over the last 30 days. If they can, the rotation will be in good hands.