Predicting Exit Velocity Using Prospect Power Grades

Publicly available Statcast data is just over two years old. Eric Logenhagen posted his first set of prospect grades before this season started. I have decided to say screw it to small samples and see how well Eric’s power grades match up with exit velocity number. Even with the limited sample, the results ended up fairly consistent.

For a little background, Eric’s power grades are based on the 20-80 scale and divided into Raw and Game power. These two categories each get a Present and Future grade. Kiley McDaniel explained the difference between the two:

Many organizations and I will split power into game power (predicting big league power stats) and raw power (how far he can hit the ball in batting practice) but they are often the same and it’s simply a way with numbers to better explain the components of power (and also comment on the hit tool).

Of the hitters Eric graded, 38 have spent time in the majors and put at least 50 balls in play. With the power grades being in increments of five, the correlation can’t be perfect so some error will exist. Here are how the power grades and exit velocity line up.

Grade: R-squared, Equation
Present Raw: 0.39, .167 * Grade +78.2
Future Raw: 0.32, .153 * Grade + 79.2
Present Game: 0.523, .227 * Grade + 78.7
Future Game: 0.567, .194 * Grade + 78.5

For reference, here is the graph of future game power and exit velocity which has the best r-squared:

The raw grades don’t correlate the best but the game grades line up great, especially considering the factors working against like small samples and jumps versus incremental grades.

Since I am head deep into the data, let me run another test. Here are the prospect grades compared to their HR/FB rates

Grade: R-squared, Equation
Present Raw: .381, .0066 * Grade -.139
Future Raw: .422, .0066* Grade -.143
Present Game: .301, .0064 * Grade – .043
Future Game: .332, .0055* Grade – .049

And here is a graph of the future raw power and HR/FB rate which has the highest r-square:

The change from game to raw power being the best fit doesn’t surprise me. Game power is looking at how often hitters transfer their power to the game. On the other hand, raw power points to the maximum power a prospect can deliver to the ball even if it is not consistent.

For convenience, here is a table with the best guess rookie performance knowing just their future game and raw power:

Some of these values, especially near the top end of HR/FB, seem a little high. First, few hitters will ever get 70 grades, let alone an 80. Additionally, the game is being played in the juiced ball era which has inflated home runs. Here is a yearly graph showing the rookie HR/FB rates.

The last three seasons really stick out. It works out to around a 3% to 4% point jump. This difference could change if the ball, or some other aspect of the game, is altered.

Overall, prospect power grades look to be a decent proxy for major league power with exit velocity and home run numbers lining up. The results can become more focused as Eric does more lists and exit velocity continues to be available.