Hit Tool Examination Pt 2: Necessary Changes

A couple of weeks ago, I examined the prospect Hit tool grade and how it provides useless information as it is currently being distributed. It’s time to dive back in. First, I am going to answer a couple questions which have come up on the topic and then get into my recommended changes.

Are there any systematic differences between Baseball America’s grades and those from MLB.com?

This study was easy. I grouped all players who had grades from both sources in the same season and I found the average differences.  The following table contain the averaged difference of the Baseball America grade minus the MLB.com grade for the 154 matched pairs.

Difference in Grades from Baseball America and MLB.com
Batting Power Speed Defense Arm
0.3 1.9 -0.9 -0.8 1.1

The final differences are small with Baseball American being higher on power while MLB.com is higher on Speed and Defense.

How well does the Power and Arm grade correlate?

This question popped up after enjoying some late night pizza at NYC’s Lunetta’s Pizza. The concept is simple; if a player has a strong arm, shouldn’t they be able to hit the ball hard? First, here is the arm and power grades plotted with a trend line.

Some correlation exists but nothing substantial. I examined the players with great arm grades and low power grades. Then, I went to see how they did in the majors. With the limited data, I ended up with a sample size of one: Christian Bethancourt, who has quit hitting and now pitches. Most the others with big differences (e.g.Francisco Mejia, Matt Chapman, Alex Verdugo, Brett Phillips, Drew Jackson, Jorge Alfaro, Raul Mondesi, and Victor Robles) haven’t played enough in the majors to draw a conclusion. It’s a decent idea but we will need to wait for the answer.

More on the Hit Tool

I pondered on where to go after proving the Hit tool is not even close to a predictive measure. I am back to believing the Hit tool consists of the following four components: Contact (contains any of the skills scouts use to evaluate a swing like bat control and speed), Plate Discipline (ability to not swing at pitches out of the strike zone), Power (enough power to hit line drives), and Speed (ability to infield groundballs into hits). With these components, a person should be able to get a good idea of hitters on-base rate (OBP) and additionally isolate power (ISO). These are the two desired hitting results, getting on-base and going as many bases as possible.

In the original article, I tried to split apart the power components. Instead, it can be an input for power and getting on base. The power’s speed component comes from the hitter’s ability to turn singles into doubles and doubles into triples as explained in this previous article.

To get an idea if the inputs would work, I took all the hitters from 2015 to 2016 and created a basic proxy for each of my four components.

Each of the above traits are scoutable and aren’t result driven. For each trait, I turned them into the 20-80 scale where 50 is average and each 10 point change is one standard deviation change. For example, a grade of 70 is two standard deviations above the mean. Here are how the grades correspond to the actually parameters.

Prospect Grades for Scoutable Inputs
Grade Contact% O-Swing% Speed Score AVG FB&LD Distance OBP ISO
80 97% 13% 9.2 333 .425 .322
70 91% 19% 7.4 321 .390 .267
60 85% 25% 5.6 308 .356 .213
50 79% 30% 3.8 295 .321 .158
40 73% 36% 2.0 282 .286 .103
30 67% 42% 0.2 269 .251 .048
20 61% 47% -1.6 257 .216 -.007

Then, I took each grade and ran a linear regression on the four components to predict the final grade. Here are the final weights:

OBP Grade = .32*Power Grade + .16 * Speed Grade + .30 * Plate Discipline Grade + .25 * Contact Grade (r-squared: 0.20)

ISO Grade = .87 * Power Grade + .13 * Speed Grade (r-squared: .60)

For the OBP grade, I could see using 30% for Power, Plate Discipline, and Contact and 10% for Speed for ease of grading. We aren’t aiming for perfection, just an improvement. Prefection later. Also, I am amazed the r-squared values ended up so high with the erratic nature of balls-in-play turning into hits. Time for a sample.

Let’s plug in Byron Buxton’s values and see how his projected and actual results compare. Just last year, Baseball America gave him the following grades:

Hit: 70
Power: 60
Speed: 80

With those values, he should be an elite hitter. He’s not close. Here are his 2016 grades, round to the nearest 5, using his actual results:

Contact: 30
Plate Discipline: 45
Speed: 75
Power: 45

The only value close is the Speed score. Using the weights found above, here are his estimate OBP and ISO grades and values.

OBP: 45 (.288)
ISO: 50 (.158)

Buxton actually posted a .284 OBP and .205 ISO. The actual ISO value was about a grade better than predicted but the OBP value was about dead on. How could the evaluators be so far off with him?

In the minors, he showed limited power, contact problems, and barely walked. Problems existed scouts must have hoped he would work out these issues. He didn’t (and most don’t but that research is for the next article).

Personally, I am going to move to the four components. I will could see the OBP value eventually being named the Hit tool and the ISO value can easily morph into Game Power. I love the framework and now it just needs to be refined.

The next major step is to determine what can be salvaged from existing prospect grades and what may need to be filled in from scouting the stat line. Until next time, happy prospecting.





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.

8 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
davels
6 years ago

great stuff!