For most of baseball history, the sacrifice bunt was simply what you did. Runner on first, nobody out, a competent contact hitter at the plate — you laid one down, moved the runner to second, took the out, and felt like a sound baseball man for doing it. Then somebody added up what actually happened next across tens of thousands of innings, and the verdict was brutal: most of the time, the bunt had been quietly handing runs back to the other team.
The weapon that killed the sacrifice bunt was not a rule change or a new pitch. It was a table — the run-expectancy matrix — and the simple, devastating comparison it makes possible. Once front offices internalized what that table said, the bunt began a long slide toward near-extinction, and the universal designated hitter finished off the last large population of bunters in 2022.
The run-expectancy argument
Start with the concept that doomed the play. Run expectancy asks: given a particular base-out state — which bases are occupied and how many outs — how many runs does the average team score from that point to the end of the inning? Decades of play-by-play data turn that question into a number for each of the 24 possible states, and those numbers are the yardstick for whether any in-game decision adds or subtracts expected runs.
The sacrifice bunt, in its classic form, trades a runner-on-first, nobody-out state for a runner-on-second, one-out state. And here is the punch line that overturned a century of orthodoxy: in the modern run environment, a runner on first with nobody out usually carries a higher run expectancy than a runner on second with one out. The bunt, executed exactly as drawn up, typically lowers the number of runs you can expect to score. You are paying an out — the scarcest resource in the inning — to move a runner one base, and the math says the out costs more than the base is worth.
Reading the matrix
A few representative numbers make the trade concrete. The table below shows illustrative run-expectancy values for the states most relevant to the bunt decision. Read each cell as “expected runs from here to the end of the inning.”
| Base state | 0 outs | 1 out | 2 outs |
|---|---|---|---|
| Runner on 1st | ~0.85 | ~0.50 | ~0.22 |
| Runner on 2nd | ~1.10 | ~0.65 | ~0.32 |
| Bases empty | ~0.48 | ~0.25 | ~0.10 |
Trace the standard sacrifice through those cells. You begin at runner-on-first, nobody out — roughly 0.85 expected runs in this representative set. A successful bunt lands you at runner-on-second, one out — roughly 0.65. You have spent an out to drop your expected runs by something like two tenths of a run. That is the whole argument in two numbers. (I have rounded these deliberately and labeled them representative because the precise figures shift season to season with the run environment; the relationship between the cells, not the exact decimals, is what matters.)
When the bunt still makes sense
Run expectancy is about average runs, and that is precisely where the honest exceptions live. Late in a tie game, you often do not care about scoring three runs — you care about scoring exactly one, right now. In that situation the right yardstick is not run expectancy but win expectancy, the probability of winning the game, and the two can disagree. Trading an out to move a runner into scoring position can raise your chance of pushing across the single decisive run even as it lowers your expected total runs. The bunt that looks foolish in the run-expectancy table can be correct on the win-expectancy table in the bottom of the ninth of a tie game.
The other classic exception is the hitter. The matrix above assumes a league-average batter; it does not hold for a genuinely punchless one. When the man at the plate is a weak-hitting number-nine hitter — or, before the universal DH, a pitcher who could barely make contact — his odds of doing something productive swinging away are so low that surrendering the out via bunt can actually be the higher-value play. A bunt is a bet that your hitter’s swing is worth less than a guaranteed advance, and for a pitcher holding a bat, that bet often paid.
The long decline
Armed with that understanding, teams did the rational thing: they bunted less, and less, and less. Sacrifice bunts per game have fallen over the long run as analytics spread from spreadsheets into dugouts, and the play that was once automatic became a deliberate, situational choice reserved for the narrow cases where it genuinely helps. The decline was gradual rather than overnight — orthodoxy dies slowly — but the direction has been unmistakable for years.
Then came a structural cliff. The universal designated hitter, adopted across both leagues in 2022, removed pitchers from the batting order entirely. Because pitchers had been one of the last reliable sources of sacrifice bunts — weak hitters for whom the bunt math actually worked — erasing them from the lineup erased a whole category of bunting at a stroke. The single largest population of bunters in the game simply stopped coming to the plate.
Through the analytics lens
The death of the sacrifice bunt is a clean case study in how analytics actually changes a sport — not by inventing exotic new tactics, but by auditing old ones against the data and discarding the ones that do not survive. The run-expectancy matrix did to the bunt what win-probability models do to a thousand other in-game decisions: it replaced “this is how it has always been done” with “here is what it costs.”
It also fits a larger pattern. The same analytical pressure that devalued the surrendered out helped drive the three-true-outcomes game and the modern strikeout era: when outs are precious and the home run is the most efficient run, small ball — bunts, slap hitting, giving yourself up — looks less and less like strategy and more like a tax. The bunt did not disappear because anyone banned it. It disappeared because the math stopped justifying it, and the dugout finally believed the math.
The bottom line
The sacrifice bunt died of arithmetic. In the modern run environment, a runner on first with nobody out is usually worth more expected runs than a runner on second with one out, so the classic sacrifice typically lowers expected scoring — you are spending your scarcest resource, an out, to buy a single base. The exceptions are real and worth knowing: late and close, when win expectancy and the play-for-one-run logic can favor it, and the genuinely overmatched hitter, the pitcher most of all. But those are narrow cases, the universal DH erased the biggest population of bunters in 2022, and the long decline tells the rest of the story. The bunt was not outlawed. It was simply found wanting.
Sources & Further Reading
- Tom Tango, Mitchel Lichtman & Andrew Dolphin, The Book: Playing the Percentages in Baseball — the foundational run-expectancy and bunt analysis.
- FanGraphs — run-expectancy and win-expectancy tools, plus sacrifice-bunt rate data.
- Retrosheet — the play-by-play records from which run-expectancy matrices are built.