The Workloads of UCLA Pitchers by Gerald Schifman July 17, 2017 There was a time when Griffin Canning looked like a sure-fire first-round pick in this year’s draft. The UCLA ace had a dominant 2017 season, ranking sixth in the nation in total strikeouts while exhibiting promising stuff, good command, and smooth mechanics. He seemed like the type of pitcher who could fly through a farm system and quickly make a big-league impact. Three days before the draft, Baseball America predicted that Canning would be selected by the Yankees with the 16th-overall pick. Yet when draft day came on June 12, dozens of picks passed by without his selection. Every team in the first round passed on him, as did every team in the competitive-balance round. In the latter stages of MLB Network’s draft telecast, Canning was chosen by the Angels with the 47th-overall pick. On June 9, he appeared destined for a $3,458,600 bonus — that is, the value MLB had assigned to the 16th pick. Instead, he took the $1,459,200 earmarked for the 47th selection. In a matter of days, Canning watched his expected price tag get slashed by 58%. The cause of Canning’s draft-stock plummet was an ominous MRI that revealed a vulnerable pitching elbow and shoulder. These injury concerns are not a surprise; last month, I examined the workloads of the draft’s top college pitchers and found that the star UCLA Bruin was used very heavily. His alarming usage rates and murky MRI warrant a deeper investigation of how longtime UCLA head coach John Savage manages his pitchers. Is Canning’s case emblematic of a culture of overuse in the program? Let’s check. First, using NCAA game logs, let’s examine the pitch counts accumulated by Bruins starters. The first four curves — regular-season-series games 1, 2, and 3, followed by midweek games — compose the top-to-bottom hierarchy in an NCAA rotation. Tournament play is held separate in a fifth curve, since the playoffs usher in greater workloads for pitchers across the college landscape. There’s a clear trend within the starter pecking order: from the midweek starter up through the series-opening ace, there’s a constant rightward shift wherein better starters are worked harder. Notice, too, that the series games 1 and 2 curves — along with the tournament curve — extend far over the 120-pitch threshold, signaling that high pitch counts are a regularity for Savage’s top starters. Percent of Starts Spanning 115+ Pitches Game Type UCLA Rate League-wide NCAA Rate Series Game 1 33.0% 14.4% Series Game 2 11.0% 8.6% Series Game 3 1.1% 6.3% Midweek 0.0% 1.8% Tournament 12.0% 15.8% All UCLA pitch-level results are for the 2012–2017 seasons, comprising the entirety of the NCAA’s game log data. All league-wide NCAA pitch-level results were computed using 2012–2015 data previously scraped by Bryan Cole. One third of regular-season starts from UCLA’s aces hit the 115-pitch mark. That massive clip more than doubles the NCAA’s average mark. The Bruins’ series game 2 and tournament starters each post double-digit rates that are closer to the high NCAA norms. The weaker links in UCLA rotations virtually never exceed the 115-pitch threshold. Savage deserves credit for handling his lesser arms with care, but his best pitchers should expect heavier usage than their peers face at many competing schools. That calls into question overall staff management at UCLA. How balanced are the workloads of Savage’s pitchers? I’ll assess this with entropy, which measures evenness in a set of percentages. After identifying the portion of the total staff workload that was handled by each pitcher, we can apply Claude Shannon’s formula to calculate entropies. A relatively low entropy (closer to zero) indicates that a coach concentrates matchups in a few select pitchers, while higher values are evidence of workloads that are more equally distributed across the entirety of a staff. I scraped seasonal lines for all 2013–2017 Division I pitchers with Bill Petti’s handy baseballr package and computed entropies for each team-season. Every bar in the chart below depicts a team’s entropy value. The bars are highly transparent, so the darker the shading, the more school-years that populate that area of the distribution. UCLA’s entropy values are highlighted in blue. See how UCLA’s entropy sits off to the left side of every distribution? It means that, year after year, Savage entrusts just a few pitchers with the lion’s share of matchups. Four of the five values are beneath the ninth percentile, with the other (2016) coming in at the 26th percentile. Phrased another way, UCLA’s entropies mean this: the five most heavily used Bruins handle 74% of batters the team faced, on average, every year. In contrast, a team that continually puts itself on the far right side of the chart is better at dispersing its matchups. For instance, Georgia Tech hit the 91st percentile three times, and the top five pitchers on coach Danny Hall’s squad averaged a slimmer 60% of all team batters faced in the five years. To incorporate rest into the analysis, let’s check the frequency at which Bruins pitchers follow the Pitch Smart college-age guidelines from MLB and USA Baseball. I’ll compute the rates of games in which UCLA starters and relievers were in violation of Pitch Smart’s rest and usage “rules” over the past six seasons. Pitch Smart Violation Rates Class Level UCLA Rate League-wide NCAA Rate Freshmen 7.4% 2.9% Sophomores 6.7% 4.2% Juniors 8.4% 5.0% Seniors 6.4% 5.8% No matter the class year, Bruins pitchers are worked more often than their similarly aged peers. Violation rates for the school’s freshmen and sophomores exceed the NCAA’s benchmarks by several percentage points. UCLA juniors defy Pitch Smart particularly often, just as they prepare to leave school for pro ball and their long-term health stops representing a prime concern for the program. Given that Bruins pitchers at all class levels are overworked and the top arms are most susceptible, you might wonder whether UCLA coaches have any trouble drawing top pitching talents to play for their program. Consider the next chart, where the quality of 2017 pitcher recruits for each college is represented with a point. The chart considers Baseball America’s top-200 draft eligibles for 2017, with the x-axis showing the recruited players’ average rank and the y-axis displaying the number of top hurlers committed. Being located up and to the left on the chart is ideal, as schools in this area are recruiting the most players who have the smallest (and best) numerical rankings. UCLA’s point, again shaded in blue, is in that top-left sweet spot, thanks to its recruitment of Hunter Greene (No. 1), Hagen Danner (No. 45), and Jeremy Estrada (No. 125). Although these three agreed to go pro ahead of the July 7 signing deadline, college came fairly close to becoming a necessary backup plan for two of them. For Estrada, a deal appeared to be reached mere hours before the deadline, while Greene and the Reds quite nearly failed to come to an agreement. UCLA is a school that carries prestige — particularly for Los Angeles locals — but choosing to attend would have positioned several of 2017’s top prep pitchers for overuse. As long as Savage continues recruiting top pitchers and producing winning seasons, he has little incentive to change the way he manages his staffs. But just as MLB clubs are growing wary of drafting graduates from reckless programs, high-school pitchers would be well served to look at UCLA’s history of excessive workloads before making a commitment. The nosedive of Canning’s draft stock is a cautionary tale, one that epitomizes the overuse issues at the prominent school.