Pitch Arsenal Score Part Deux by Daniel Schwartz December 19, 2014 This week, Eno presented Pitch Arsenal scores. There were a sundry of impactful responses which will result in a more comprehensive score in the future, such as incorporating pitch usage; a dominating pitch used more should be scored higher than a pitch with the same results that is used less. The other clear requirement is associating weight to the outcomes in which Eno evaluated: whiff per swings and grounder to fly ratios. That is all this post will get at. It sounds like follow-up scores may incorporate pop-up% and zone% as well. Again, we are looking for undervalued pitchers, but if we want to get more descriptive (simply presenting the results from last year), we could look at HR/FB and LD/BIP on each pitch type as well. I did something similar in the past, and there are different approaches, but here goes. Using BP’s PITCHf/x Leaderboard (1oo-pitch qualifier) I first found each pitch type’s Whiff/Swing and GB/FB z-score. I then weighed each z-score by swinging-strike’s (.62) and GB/FB’s (.27) average correlation to ERA, SIERA and xFIP for all pitchers over 70 innings pitched (my lazy attempt to omit relievers). What is not incorporated and should be in the future, is differentiating weights for each pitch type. For example, groundball induction is a more important requirement for Sinkers; Sliders and Change-ups should induce more whiffs as we know from these Pitch Type Benchmarks. The seven pitch types incorporated are Fourseamers, Sinkers, Cutters, Curves, Sliders, Change-ups and Splitters. For Trevor Bauer‘s Screwball and R.A. Dickey’s Knuckleball, I went to Brooks Baseball’s player cards and used their sabermetric outcome z-scores for these two pitch types and combined the weighted z-score to the rest of their repertoire score. The list will be distinguishable from Eno’s initial list based on the weights, but the names within each cluster will look similar. When you get Felix Hernandez and Corey Kluber in the top 5 and Kevin Correia dead last, you know we got it right: Notes: Pitchers with outcomes missing threw less than 70 IP. Pitch types were included only if pitches were thrown 100+ times. For example, Trevor Cahill wound up at #8 on this list. He has excellent breaking and off-speed pitches (all 1+ standard deviations more than average from a whiff/swing perspective), but his Fourseam is awful, and BP’s PITCHf/x leaderboard did not present his Fourseam outcomes based on a sub-100 pitch count as a starting pitcher so it was not included in his full arsenal/repertoire score. Results: Think the Indians aren’t looking on the pitch-outcome level? They have 5 pitchers in the top 55 Repertoire Scores: Carlos Carrasco and Corey Kluber are numbers 3 and 4 respectively! Newcomer Gavin Floyd is #15 on the list thanks to great breaking balls. Danny Salazar is #43 and his per-pitch outcomes were even better coming into 2014. T.J. House is the real deal at #54. And not far off is Trevor Bauer at #75 overall. I cannot not mention that the Mets have two guys in the top 21: Zack Wheeler (#9) and Jacob deGrom #21). Harvey if healthy should join them. As it relates to Cahill, for some value-add, I highlighted (red/green) those .50 more or less standard deviations from the mean in SwStr% and GB/FB, but also the three “luck” categories: LOB%, HR/FB and of course BABIP. Cahill ranks well in both SwStr% and GB/FB and was also relatively unlucky when it came to his BABIP and LOB% (about 65 points and 9% respectively from his career average). Cahill is a deep-league…even standard-league asset with some returned control and an ERA closer to expected ERAs below 4.00. I am thinking the magical-mystery-kind age 27 season for Cahill. Attend to the luck stats. Carlos Martinez has one of the best repertoire scores yet had a 4.03 ERA in part due to a .333 BABIP. His xFIP (3.45) and SIERA (3.54) in conjunction with the #1 overall arsenal score should ensure he is significantly better next year. The list verifies breakouts: Carlos Carrasco, Garrett Richards, Yordano Ventura and Tyson Ross (all top15). Ross has a devastating Slider (8th best; 100-pitch qualifier) with a solid Change from a whiff/swing perspective, but he induces a ton of grounders with his Change and Sinker (both have top 11 GB/FB rates). The list also provides hope for Justin Masterson who wound up at #7 on the list with 2.72 Sinker score and .74 Slider score. That 2.72 Sinker score out-summed 205 other full repertoire scores, but again keep in mind that his Sinker whiff/swing of 22.88% might be 3 standard deviations better than the average Sinker whiff/swing, but that rate is under the average whiff/swing rate for Curves, Sliders, Change-ups and Splitters. This is why it is important to differentiate weights for each pitch type. If there is a pitcher’s Repertoire score that baffles you, check out the “RepScores” tab to see which pitches were omitted and the individual pitch-type z-scores and weights under the “zWh/Sw” and “wzWh/Sw” columns. Each pitch-type score and repertoire score is in the last column, titled “Wh/Sw_Gb/Fb_Scr.” Back to the actual value of the content though – we are looking for interesting names and patterns, which I have highlighted in Red or Yellow based on the content below: Trevor Cahill interests me if he never uses his Fourseamer again and can induce more grounders with his Sinker. Cahill has a nice Cutter/Slider that he used toward the end of last year too. There is more swing and miss to his game. Dan Straily and the Sinker Pattern: Dan Straily might not have a spot in the Cubs rotation at this point, but he has a solid Slider and induces a surprising number of pop-ups. If he can keep his HR/FB ratio down (probably more of an issue in Wrigley than the Coliseum though) and have better luck in the left-on-base department, he could become effective again. His ranking in this list though is because of an elevated Sinker value (similar to Masterson). Look out for the Sinker and take the associated rankings with a handful of salt. James Paxton: By name and the Mariners un-willingness to move him, you know that he is good. It’s nice that our list backs him up. While his expected ERAs were 3.50+ (about 50 points higher than his 3.04 actual ERA), he should get even better; mainly his Change-ups, which should induce more whiffs. Even if we don’t see an impressive strikeout rate next year, I think he will be able to prevent hard contact with his CCC (Change-Curve-Cutter) repertoire and Fastball velocity (17th at 95.51 MPH last year; 100-pitch qualifier). Shane Greene: He’s no Max Scherzer, but the Tigers finagle themselves into another Yankees-Diamondbacks swap and again may have acquired the best piece. Another potent righty, Greene had the 10th best Slider Whiff/Swing rate and 5th best Cutter Whiff/Swing rate (enough velocity differential where they are distinguishable). Unlike Masterson and Straily, his Sinker does not inflate his repertoire score. Losing Rick Porcello hurts, but Shane Greene could be even better. Jesse Hahn and Pitch Omissions: Similar to Masterson and Straily, Hahn’s repertoire score is inflated by his Sinker although he had a great Curve (31st best Whiff/Swing rate). His Slider and Change were omitted from his arsenal in this evaluation based on pitch-counts under 100. Both would impede on his overall score. Through this approach, Hahn’s Changeup would be valued at -.15, which would drag his repertoire score from 1.59 to 1.44. The Slider would be far worse, but he used it less than 30 times, so I won’t subtract from his overall score in case he decides to swap the pitch out altogether next year. Wily Peralta. This is more about confirmation than highlighting a surprising name. Peralta won’t provide haughty strikeout rates, but his 18.4% rate from last year could jump a bit based on the per-pitch outcomes/more breaking-ball usage. In a future FanGraphs+ post, his release point consistency (within-game but especially game-to-game) will be highlighted, which could be a source of deception. While the Sinker does inflate his value some here (almost a standard deviation more whiff/swing than the average Sinker), it does its job inducing grounders. His Changeup and Slider are also .55+ standard deviations better than average at inducing whiffs. Eno already emphasized Carlos Martinez and Carlos Carrasco – both have glorious peripherals within each pitch-type beyond their Fourseams. Every pitch of theirs outside the Fourseam induced 1.5 standard deviations more whiffs/swing than average! Watch out for Danny Salazar who (by classification) introduced a devastating Changeup in September and should output value similar to or better than Carlos Carrasco based on a haughtier strikeout rate. Salazar was probably a luckier BABIP away from a 3.5 ERA last year. In fact, unsurprisingly, each Indians starter outside of Carrasco had a BABIP .5 standard deviations over the average (70IP qualifier). Review: Look for sleepers/surprises based on “Rep.Score” in column two of the embedded file as presented. Check out the “RepScores” tab to see if any pitch-types were removed based on pitch-counts and keep pitch-type usage in mind (counts in the 5th column). Remember, differentiating weights for each pitch type is not incorporated. Scroll to the right within the “RepScores” tab to view individual pitch-type scores as well as the complete repertoire/arsenal score. Watch for pitchers with Sinkers as their highest score. Sinkers whiff rates aren’t as impactful. Allen Webster has a great score, but it is inflated because of its Whiff/Swing rate. As mentioned above, the same holds true for Justin Masterson and Dan Straily among others. The Diamondbacks and Yankees won’t let the Tigers finagle themselves into another deal after Shane Greene enjoys success similar to the Scherzer application. And finally, try to own the entire Indians rotation in your fantasy leagues.