Chapter 6 of 11
The times-through-the-order penalty
Hitters get better each time they see the same starter in a game. The hook should come from that curve, not from the pitch count alone.
The short version
Pull the starter when his third-time-through self is worse than the reliever you would bring in, not when he reaches a round number.
KBO 2026 season data through Oct 13, 2026
01
The idea
The first time a hitter faces a starter in a game, he is seeing the pitcher's release, his speed and his mix for the first time that day. The second time, he has a picture. The third time, he has seen most of the repertoire and the pitcher has started to tire. Hitters get better each time through the order, and the effect is large enough to measure.
That turns the hook into a comparison rather than a feeling. The question is not “is he tired?” or “has he thrown a hundred pitches?” but “is the version of him facing this lineup a third time worse than the reliever I would bring in fresh?”
02
The math
starter wOBA against, pass n ≈ talent + (n − 2) × penalty; pull him when it exceeds the reliever's
The Book measured the penalty in MLB at roughly ten points of wOBA per pass, partly familiarity and partly fatigue; this guide uses 9. It is an MLB-derived figure: the KBO publishes no batter-by-batter record that would let it be measured here.
03
Try it
The reliever starts 22 points worse than the starter: this season's gap between KBO relievers' and starters' FIP, converted to wOBA.
On these settings the starter stays better than the reliever even a fourth time through — an ace, or a thin bullpen.
04
This season in the KBO
Innings per start
5.0
Pitches per start
84
Batters per start
22.3
approx.
Quality start rate
35%
| Role | IP | ERA | FIP | K% | BB% |
|---|---|---|---|---|---|
| Starters | 6,824 | 4.25 | 4.26 | 20.1% | 7.6% |
| Relievers | 5,936 | 5.17 | 5.11 | 18.9% | 11.3% |
KBO starters averaged 5.0 innings and 84 pitches a start this season — about 22.3 batters, which is two full trips through the order and part of a third. Relievers posted a worse FIP (5.11 against 4.26), striking out fewer hitters. On those averages the tool above never pulls the starter: the typical KBO reliever is worse than the typical starter even facing the order a fourth time. The case for the hook rests on a bullpen's best arms, not its average — slide the reliever to the starter's level, a good set-up man, and the hook arrives at the start of the third pass.
The pitcher fatigue index →Every pitcher's pitches per inning →
05
When the conventional wisdom is wrong
01
The hundred-pitch line
One hundred is a round number, not a threshold. An efficient starter at 85 pitches facing the order a third time may be done; a laboured one at 95 in the fifth may have a pass left against the bottom of the order.
02
“Let him get the win”
Leaving a starter in for a fifth inning so he qualifies for a win is a decision made for a box score. The third-time penalty does not care what the scorer will write.
03
The ace is the exception
Set the starter twenty points better than the league in the tool. Even a third-time ace beats most relievers. The penalty is a reason to manage ordinary starters, not to take the ball from great ones.
04
Bullpens are finite
Every hook spends reliever innings, and a tired reliever is not the one in the chart. Over a 144-game season the hook has to be priced against tomorrow's bullpen as well as tonight's lineup.
Starts are approximated by the appearances of pitchers working as starters, so a starter's occasional relief outing is counted in his per-start averages. The batters-per-start figure uses the league's plate appearances per inning. The reliever default converts a season-long FIP gap into wOBA and ignores leverage and handedness.