Chapter 1 of 11
Run expectancy and the 24 base-out states
Every situation in an inning is worth a number of runs. Learn the table and most strategy arguments become arithmetic.
The short version
An out is the scarcest thing an offence owns; a decision is good when the state it leads to is worth more than the state it leaves.
KBO 2026 season data through Oct 13, 2026
01
The idea
Every moment of an inning can be described by two facts: which bases are occupied and how many are out. Eight arrangements of runners times three out counts makes twenty-four states, and every plate appearance moves the game from one of them to another — or ends the inning.
Ask of each state a simple question — from here, how many runs does a team score before the inning ends, on average? — and you have a run expectancy table. It is the most useful single object in baseball strategy, because it turns arguments about momentum and pressure into a subtraction: is the state after the decision worth more than the state before it?
The second number that matters is the chance of scoring at least once. Expected runs is the right yardstick early in a game, when every run counts the same. Late and close, a single run can be the whole game, and the chance of getting one becomes the thing to maximise. Most of the arguments in this guide come down to which of those two numbers a manager should be looking at.
02
Try it
Set the runners and outs. The numbers are this season's: half-innings played from each state on the KBO's own event rates.
Tap a base to add or clear a runner.
Expected runs
0.97
from KBO event rates
Chance of a run
44%
before the inning ends
Lab matrix
1.01
borrowed shape, KBO scale
This situation is worth 0.40 runs more than the start of an inning (0.57).
How many runs the rest of the inning produces
| Runners | 0 out | 1 out | 2 out |
|---|---|---|---|
| ___ | |||
| 1__ | |||
| _2_ | |||
| __3 | |||
| 12_ | |||
| 1_3 | |||
| _23 | |||
| 123 |
Table: the chance of scoring at least once from each state. Tap a cell to set the diamond.
03
The math
The value of any play is the change in run expectancy it causes, plus any runs that scored on it. Analysts call this RE24, after the twenty-four states.
value of a play = RE(after) − RE(before) + runs scored
Example: with a runner on first and nobody out, a single that moves him to second is worth 0.61 runs (1.58 − 0.97).
The table makes the price of an out visible. A leadoff out costs 0.27 runs; a leadoff walk adds 0.40. The first out of an inning is the most expensive thing an offence can give away, which is why nearly every chapter that follows begins by asking what an out buys.
04
This season in the KBO
Runs per team game
5.07
Runs per inning
0.56
From the empty state
0.57
simulated
Score at least once
29%
per inning
The KBO scored 0.56 runs per team per inning this season. Played out from the bases-empty, nobody-out state on the league's own rates of walks, singles, extra-base hits, home runs and strikeouts, the chain lands on 0.57 — close enough that the simulated table can be trusted in the cells the real one cannot be checked against.
The lab's matrix takes a different route: it borrows the shape of a large modern sample and rescales it to the KBO. The two disagree by at most 0.29 runs in any cell. Where they differ, the chain is the more Korean of the two, because it is built from this league's own mix of outcomes; the lab matrix is the more battle-tested, because its shape came from real innings.
05
When the conventional wisdom is wrong
01
“Get him into scoring position”
A runner on second with one out is worth 0.74 runs; a runner on first with nobody out is worth 0.97. Moving him up with an out costs 0.23 runs and changes the chance of scoring by 0.3 points. Scoring position is a place, not a prize.
02
“A productive out”
Runner on second, nobody out: 1.20. A grounder that moves him to third: 0.97. The out cost runs on average — though the chance of a run went from 64% to 67%. It is productive only when one run is all you need.
03
“Two out, the inning is over”
With the bases loaded and two out a team still expects 0.89 runs and scores 36% of the time. Two out is not over; it is expensive.
04
“Runs are runs”
They are not, late in a close game. The next chapters keep two ledgers — expected runs and the chance of scoring once — and a third, win expectancy, built for the KBO's eleven-inning game.
The KBO publishes no play-by-play, so neither table can be read straight off Korean innings. The simulated one uses this season's league-wide event rates and the lab simulator's baserunning assumptions (how often a runner scores from second on a single, and so on), which come from a large modern MLB sample. 40,000 half-innings were played from each state; the sampling error is about a hundredth of a run.