Minors To Majors: Hitter Grades & Minor League Results
I’ve made it almost a month since I made the following declaration on investigating prospect Hit grades:
I am going to stay away from more Hit tool predictions until I have collected every one of MLB.com’s prospect grades from 2013 and 2014, not just the top 100. I probably will not be able to compare many to their major league stats but I can with Triple-A.
I broke my position after collecting MLB.com’s 2013 grades. I ventured forward without the 2014 grades. With this larger and more diverse dataset, I compared the hitters’ grades to their batting average, home runs, and stolen bases in both AA and AAA.
Trying to better understand the Hit tool stems from finding it doesn’t contain any predictive power. When looking at players with different grades, major leaguers ended up posting similar batting averages. I concluded two issues were causing the production to level out.
First, hitters needed a talent and/or production baseline to get into the majors. Some hitters with below average grades were under-graded and produced up to the MLB baseline. Additionally, “better” hitters were over-graded but still had just enough talent to make the majors. This talent convergence tends to average out the grades.
Second, I was only looking at players who made the majors. These are good players and have a baseline. Not every prospect has MLB talent. More bad prospects needed to be included in the dataset. I was needing more failures.
To solve these two issues, I compared the 2013 position players’ MLB.com prospect grades to their career AA and AAA production. About 300 players with grades were used. And in 2013, players received integer based grades (2 to 8). For ease and consistency, I’ve added the zero to each grade.
To start with, here are the median batting averages for each Hit grade at AA (283 samples) and AAA (240 samples).
| Grade | AA | AAA |
|---|---|---|
| 70 | 0.276 | 0.289 |
| 60 | 0.277 | 0.288 |
| 50 | 0.266 | 0.269 |
| 40 | 0.266 | 0.262 |
Even in the minors, there is not much range in batting average with the range being just 10 points in AA. Sadly, this is just another instance pointing to grading the hit tool has hard and/or not being done correctly. For now, I would group hitters as average (40-50) or above average (60-70). There doesn’t seem to be a third tier.
While my goal was to examine the Hit tool, I decided to find out how the Speed and Power grades held up. For Speed, I compared the grade to stolen bases per chance using Tom Tango’s formula.
| Grade | AA | AAA |
|---|---|---|
| 80 | 54% | 36% |
| 70 | 26% | 24% |
| 60 | 23% | 8% |
| 50 | 12% | 10% |
| 40 | 8% | 6% |
| 30 | 3% | 2% |
| 20 | 1% | 2% |
The numbers don’t surprise me with faster runners getting more stolen bases but the success rate dropping with the transition to AAA.
Finally, here is how the Power grades compare to HR/600 PA.
| Grade | AA | AAA |
|---|---|---|
| 70 | 36 | 29 |
| 60 | 17 | 19 |
| 50 | 13 | 15 |
| 40 | 7 | 10 |
| 30 | 5 | 6 |
| 20 | 1 | 5 |
Home runs steadily increase for each higher grade as expected. For hitters making it to AAA, they generally show some power increase from AA.
I continue to make little headway on unraveling the Hit tool. Maybe there is nothing to unravel. As for a hitter’s Speed and Power grades, they’re a recent proxy for future production.
Jeff, one of the authors of the fantasy baseball guide,The Process, writes for RotoGraphs, The Hardball Times, Rotowire, Baseball America, and BaseballHQ. He has been nominated for two SABR Analytics Research Award for Contemporary Analysis and won it in 2013 in tandem with Bill Petti. He has won four FSWA Awards including on for his Mining the News series. He's won Tout Wars three times, LABR twice, and got his first NFBC Main Event win in 2021. Follow him on Twitter @jeffwzimmerman.
First, I think you are addressing an important question.
Second, I suspect much of the problem has to do with the way scouting categories, especially the hit tool, have lagged far behind our understanding of the components that contribute to the success or failure of a hitter.
We all know that a batter’s offensive contributions revolve around two abilities: Getting on base and enabling base runners to advance (I am for convenience sake not mentioning the ability to advance one’s self as a baserunner through stolen bases or taking extra bases). It seems to me the “hit tool” is a poor proxy:
Here’s the variables as Kiley McDaniel described it on Fangraphs: 1) athleticism/looseness 2) bat speed 3) some feel for the bat head 4) some sense of a plan at the plate, to recognize pitches/adjust and other plate discipline type things. Kiley defended the approach, writing, “It may seem too subjective for that sort of thing an integral part of an evaluation, but it’s amazing to me how often a quick observation like that will be backed up by a hitting coach, the stat line, later at-bats and often the player himself telling you he was out of sorts,” but it seems to me his explanation raised as many concerns as answers: It seems scouts reach an initial belief based on a very small sample, then look for other means to confirm their bias.
There are too many variable and no uniform way scouts measure those individual variables or weigh their importance in creating a hit tool number. At least I am unaware of such an effort. Perhaps some scouts are more swayed than others by bat speed; some scouts may weigh plate discipline in a substantial manner and some may neglect it entirely. And I strongly suspect that with the absence of more objective and better defined measures, scouts tend to let variables that really have little to do with hitting affect their grading of the hit tool. Did scouts give Byron Buxton a high hit tool despite his poor plate discipline because he is such an explosive and talented athlete in the broadest sense? Do scouts discount the hit tools of players who appear to be not in shape?
The approach Kiley described may work well at the extremes — those with elite hit tools or horrible hit tools. But I would suspect they have very little value closer to the middle, as your analysis seems to show.
Thanks for your work!
Thanks for the reply. I know some organizations add a plate discipline value to their scouting profile. I just wish the answer was easier but the lack of differentiation in hit tool is disturbing. I’m working on a difficult and time-consuming solution but I hope to finally meld some of the stats and scouting reports. I really thought I would find a better answer with this analysis. Oh well.