Jordan’s Research Ramblings: Underrated Minor League Power Performances
Minor league data is not yet at the same quality of MLB data, but it’s catching up. In the past few years, key new data sources have become available to the prospect analyst: FanGraphs and Baseball America now offer average exit velocity, while Prospects Live offers fly ball distance, and Rotowire offers hard hit rate. This data can be incorporated to generate a Mike Podhorzer-style expected home run per fly ball (HR/FB) rate for minor leaguers. Expected home run per fly ball rate measures how many home runs can be expected for a player for a given set of characteristics–in our case, average exit velocity and fly ball distance. I’m currently working on incorporating this data into my peak wOBA projections for prospects, to more accurately capture prospect power performance. In the meantime, I wanted to highlight some exit velocity and fly ball distance power gainers in each full-season league.
I used 2016-2019 MLB data to build a model of HR/FB rate based on exit velocity and fly ball distance. Average fly ball distance in season one is correlated with average fly ball distance in season two at .39; average exit velocity is a bit stickier, and correlated with itself year-over-year at .59 (minimum 50 batted ball events in each case). I had to use MLB data as year-over-year exit velocity data for minor leaguers is not available; 2019 is the first season where it was made available to the public. Both of these metrics capture repeatable power skills to a substantial degree.
Using weighted regression, the r-squared for a (predictive) model of season two HR/FB rate from season one average exit velocity and average fly ball distance is 24%. 24% of the variation in season two home run per fly ball rate is explained by the combination of season one exit velocity and fly ball distance. A 1% increase in fly ball distance is associated with a 3% increase in HR/FB rate (p<.001); a 1% increase in exit velocity is associated with a 5% increase in HR/FB rate (p<.001). A 24% r-squared (or .49 correlation) is similar to the r-squared for HR/FB rate predicting itself one season ahead.
The r-squared for a descriptive model, with exit velocity and fly ball distance explaining same season HR/FB rate, is of course much higher, at 57% (.76 correlation). See the figure below.
For example, Jo Adell’s average fly ball distance was 1.04, 4% above league average (311 feet for Adell vs. 300 feet Triple-A average), while his average exit velocity was 1.03, 3% above league average (92 average EV for Adell vs. 88.9 league average EV at Triple-A). These numbers are associated with a 1.28 xHR/FB rate (28% above league average). League average HR/FB in Triple-A is 15.4%, so Adell’s xHR/FB rate at Triple-A is 20%. See the appendix for a table of league averages for each metric and also for Adell’s full calculation (in case you want to plug in numbers yourself for a different player).
The rest of this article focuses on minor league power underperformers that are unveiled once exit velocity and fly ball distance are accounted for. It leverages the MLB model of HR/FB rate, assuming the relationship between fly ball distance and exit velocity and HR/FB rate is similar in the minors and majors (after adjusting for league context; different leagues have very different league average HR/FB rates, fly ball distances, and exit velocities, so this is an important step).
- Malcolm Nunez. 7% xHR/FB vs. 0% HR/FB. 90 average exit velocity. 273 average fly ball distance.
- Additional home runs using xHR/FB rate: 1.3. New triple-slash incorporating additional home runs: .199/.262/.267
- Gerlado Perdomo. 6% xHR/FB vs 2% HR/FB. 87 average exit velocity. 274 average fly ball distance.
- Additional home runs using xHR/FB rate: 3.4. New triple-slash incorporating additional home runs: .277/.402/.398
Nunez had a historic season in the Dominican Summer League in 2018 before metaphorically and literally whiffing in 2019. It’s perhaps not all doom and gloom, however, as his exit velocity was quite strong at 90 MPH, and his fly ball distance was only slightly below the Class-A average. Do continue to use him as an archetypal reminder to be skeptical of DSL performances, but don’t fully abandon him yet. Note that I expected his new triple-slash would be considerably better than it actually turned out to be; it’s still pretty yucky even after granting him the additional home runs.
Perdomo has elite plate discipline and contact skills and is young for level. He’s also fast and defends a premium position. His power output is the only major question mark for him, typically determining whether a prospect analyst is able to contain their enthusiasm for him. His power measures are average-ish for his league, which is actually pretty exciting once you consider his age and bat-to-ball skills. Look for him to continue to rising up lists in 2020.
- Alejandro Kirk. 9% xHR/FB vs. 6% HR/FB. 91 average exit velocity. 292 average fly ball distance.
- Additional home runs using xHR/FB rate: 2. New triple-slash incorporating additional home runs: .295/.401/.479
- Vidal Brujan. 6% xHR/FB vs. 3% HR/FB. 87 average exit velocity. 279 average fly ball distance.
- Additional home runs using xHR/FB rate: 1.3. New triple-slash incorporating additional home runs: .296/.363/.415
Kirk has completely dismantled the minors in 2018 and 2019; it’s perhaps harder to argue he’s not the games most underrated offensive prospect at this point. Few offer his elite combination of contact ability, patience, and power. Brujan, like Perdomo, is more known for his speed and contact skills. He also offers a bit of pop, though, and is relatively young for someone who reached Double-A later in the year. He should be able to offer enough over-the-fence power to keep defenses honest, maybe eclipsing the 10 home run per year mark in his prime.
- Royce Lewis. 9% xHR/FB vs. 5% HR/FB. 90 average exit velocity. 290 average fly ball distance.
- Additional home runs using xHR/FB rate: 1.7. New triple-slash incorporating additional home runs: .242/.300/.405
- Oneil Cruz. 10% xHR/FB vs. 3% HR/FB. 91 average exit velocity. 297 average fly ball distance.
- Additional home runs using xHR/FB rate: 2.2. New triple-slash incorporating additional home runs: .285/.360/.481
- Nico Hoerner. 9% xHR/FB vs. 4% HR/FB. 88 average exit velocity. 297 average fly ball distance.
- Additional home runs using xHR/FB rate: 4.5. New triple-slash incorporating additional home runs: .298/.357/.462
These prospects are highly rated everywhere, but rate less favorably by my stats-first approach (with the exception of Cruz, who ranks highly regardless of approach). My league and age-adjusted statistics do not yet adjust for power luck, though, and these three had quite unlucky seasons when it comes to power output. There’s a reason Lewis especially continues to rank so highly in the industry despite his lackluster performance in 2019; the tools consistently rate very strongly, and his power measures, e.g., a 90 average exit velocity, back that up. Look for his performance to bounce back in 2020, though he’ll likely continue to rate higher on real-life lists than fantasy lists thanks to his premium defensive position and less-than-elite-but-still-solid offensive skills. Hoerner showed underwhelming exit velocities in his brief MLB sample (around 86 MPH average exit velocity); the numbers here suggest he might be able to push 15-20 home runs in his peak. This would be a big improvement on the 7-12 home run power he showed in 2019. 15-20 home runs paired with his contact skills would make for an exciting offensive package. I’ve underrated him in the past, but after incorporating underlying power metrics into his performance, I’m feeling increasingly optimistic.
- Jo Adell. 20% xHR/FB vs. 0% HR/FB. 92 average exit velocity. 311 average fly ball distance.
- Additional home runs using xHR/FB rate: 6 (including 2 home runs that were erased in a rainout). New triple-slash incorporating additional home runs: .305/.357/.537.
- Ke’Bryan Hayes. 17% xHR/FB vs. 9% HR/FB. 92 average exit velocity. 294 average fly ball distance.
- Additional home runs using xHR/FB rate: 9.7. New triple-slash incorporating additional home runs: .285/.354/.500
- Cristian Pache. 16% xHR/FB vs. 6% HR/FB. 88 average exit velocity. 306 average fly ball distance.
- Additional home runs using xHR/FB rate: 1.9. New triple-slash incorporating additional home runs: .292/.353/.486
Adell was the biggest xHR/FB rate gainer in the minor leagues, with a 0% HR/FB rate, and 20% xHR/FB rate. He clearly has elite power and appears to just have had an unlucky streak in his brief Triple-A debut. Look for the home runs to come in bunches in Triple-A in early 2020 as he makes it hard for the Angels to resist giving him an MLB job. Hayes is another player whose overall performance was underwhelming in 2019, below average in Triple-A. It looks considerably better after luck-adjusting his HR/FB ratio–the underlying power measures are quite strong. Pache’s power took a major step forward in 2019; his xHR/FB rate backs that up and then some. He ranks highly on prospect lists mostly for his defensive skills, but he should be able to hit and run enough to offer usefulness with the bat as well.
Jo Adell xHR/FB rate calculation for Triple-A using predictive model: 5.1*1.03+3.2*1.04-7.3=1.28. This means a player with his exit velocity and fly ball distance is predicted to have a HR/FB rate 28% above league average in the following season.