Learn · momentum
Does a lucky break create momentum?
No. Across 19,080 pure-luck plays since 2019, the team that got the break did no better on its next drive than the team that got the bad one: −0.07 expected points against average for the lucky side, −0.05 for the unlucky side.
The test
We took every play since 2019 where chance decided something worth at least a point: a fumble recovery, a field goal a kicker should have made (or missed), a judgment call, an interception-worthy throw that was or wasn’t caught. Then we looked at each team’s next drive, in expected points added, against the average drive.
If momentum were real, the lucky team would play better right after the break and the unlucky team worse. Neither happened. Both drives came in slightly below the average drive, and the difference between them, −0.02 points, is well inside the noise. Per point of luck, the carryover is -0.005 ± 0.016: zero.
| Kind of break | Lucky minus unlucky next drive |
|---|---|
| Fumble bounces | −0.05 pts |
| Interceptions | +0.01 pts |
| Kicks | +0.02 pts |
| Judgment calls | −0.02 pts |
Why it feels real
Runs happen. A team can string together stops and scores, and win probability can swing hard in one direction for a stretch; researchers at the University of Wisconsin–Milwaukee found such runs more often than chance would produce (UWM). That is a different question from ours. Runs can come from one team simply playing better for a while. What we tested is whether a lucky break itself changes what happens next, and it doesn’t.
What we do with it
Momentum is off by default in every “adjust the assumptions” panel on this site. If you believe a break lifts the next drive, set it: the panel scales each break by the carryover you choose and replays the game again.