Power Audit
A slugger's home runs dry up. Was it the park, bad luck, or the hitter? Split any power drop into those three pieces.
Builds on Answers for Coach
- Coach questions
- 3
- Asked on air
- 17×
- Top question
- Q-05
Coach questions it could answer
The playbook →- Is a power outage the park, the luck, or the hitter?
Split a hitter's home run change into three parts. Park is his expected home runs in a neutral park versus his own; luck is actual home runs versus what his contact deserved; hitter is the change in that contact itself.
- How much has the ball and the home run era shifted?
Run the same split league-wide by season to separate the ball, the parks and the hitters across the home run era.
- Why does offense shrink in October and in April?
Split league home runs by month to measure how much of the April and October drop comes from cold weather and the mix of parks, and how much is left for pitching and hitting.
Data it needs
The data lake →1 of 2 on hand; the rest arrives in phase 1, The spine.
Could use later: Weather
The question
“Power outage” is one of the hosts’ favorite puzzles: say a hitter who hit 35 home runs last year is on pace for 18. The usual suspects are a new home park, bad luck on fly balls, or a real change in the hitter, and the answer changes what you’d expect next year.
How it would work
- Park: estimate each batted ball’s home run chance in a neutral park and in the hitter’s actual parks.
- Luck: compare actual home runs with expected home runs from exit velocity, launch angle and direction.
- Hitter: compare the contact itself, how hard, how high and how often pulled, with his past seasons.
- The three parts add up to the total change, drawn as one stacked bar.
First version
- Any hitter, this season against last, with the three-part bar and the numbers behind it.
- A league leaderboard of the biggest power drops, labeled by which part dominates.
The data
Statcast batted balls with exit velocity, launch angle and spray direction, and park dimensions. Answers for Coach already keeps the Statcast cache for 2022–26.
What it needs
- A park model. Fence distance and height by direction, plus weather, turned into a home run probability for each batted ball.
- Clean parts. The three pieces overlap a little, so the method has to say how the overlap is split.
Status
An idea on the launchpad. Nothing is built yet.