Coming out of the All-Star break, our board leaned under. Heavily. Six of ten plays. Before you assume the model ate something bad over the break, we did what we always do when a pattern shows up: we went and checked whether baseball agrees.
The question is simple. A starter gets four or five days between starts all season, then the break hands him eight or nine. Does the extra rest make him sharper, or rustier? Everyone has an opinion. We had 49,000 starts sitting in a database, which beats opinions.
What we did
For every season from 2015 to 2025 (2020 excluded, since that season was its own science experiment), we found the All-Star break automatically: the longest league-wide gap in games during the July window. Then, for every pitcher with at least five starts before the break, we took his own pre-break strikeout average as his personal baseline and asked one question: in his first start back, did he beat his own number or fall short of it?
That last part matters. We didn't compare Paul Skenes to a fifth starter. Every pitcher is measured against himself, which removes the "good pitchers are good" problem that ruins most stats you'll read this week.
What we found
Pitchers strike out about a fifth of a strikeout fewer in their first start after the break. Small? Yes. Real? Also yes. Across 1,619 first starts the effect is statistically solid (t = −3.3 for the three of you who wanted that), which means a run of luck this size shows up by chance roughly once in a thousand tries. And 54 percent of pitchers came in under their own baseline, when a coin flip says 50.
Then it disappears. By the second start back the gap is −0.03 K, which is a rounding error wearing a trench coat. The rust is a one-start phenomenon. He shakes hands with his slider again, and everything goes back to normal.
What it means if you bet
Here's the honest read, which is the only kind we do. A fifth of a strikeout does not win you a bet by itself. If a line is 6.5 and a model says 6.4, the rust does not turn that into a play. What it does is lean on close calls. Do the books already price the rust in? We can't tell from this data, and we won't pretend otherwise. What we can say: when our number already sat below the line, the first night back added a small, real tailwind, and that tipped our close calls under. That's the entire size of the claim.
It also means the reverse: if someone tells you to smash unders across the board every July because "pitchers are rusty," they've taken a fifth of a strikeout and marketed it into a strikeout and a half. The data doesn't support the yelling. It rarely does.
Why we're telling you
Because this is the whole product. We find small, real edges, we bet them when the price is right, we pass when it isn't, and every pick gets graded in public either way. The rust is one ingredient. The record is the receipt.