Do Timeouts Actually Stop a Drought?

We checked more than a thousand timeouts called mid-drought. Calling one does not measurably break the cold streak faster. The idea that a timeout “stops the bleeding” is mostly an illusion baked into how the raw numbers get read.

The finding: a coin flip, not a circuit breaker

24.6%

of droughts break within one possession (~35s) after a timeout is called

23.5%

break in any equivalent 35-second window with no timeout — the natural baseline

+1.1

percentage-point lift from calling the timeout — statistically indistinguishable from zero

Across a random sample of 1,500 games and 1,014 timeout droughts, a timeout buys almost no extra chance of ending the drought on the next possession. Coaches aren't wrong to call them — a timeout resets the play, rests legs, and draws something up — but the specific claim that it breaks the scoring drought doesn't show up in the data.

Why the crude numbers lie

Look at this season's droughts and the surface story flips exactly backwards. Droughts where a timeout was called are longer and cost more points than droughts with no timeout:

Droughts (17,569 this season)	Avg length	Avg points allowed
Timeout called (4,185, 24%)	4:10	9.6
No timeout	3:50	6.9

Read naively, that says timeouts make droughts worse. They don't. This is reverse causality: coaches call a timeout precisely because the drought is already bad and getting worse. The timeout is a response to the length, not a cause of it. Comparing timeout droughts to non-timeout droughts compares a team's worst cold stretches against all the ordinary ones — the timeout never had a fair fight.

The only honest test controls for how long the drought had already run when the timeout was called, then asks: from that moment, does the drought end sooner than a drought that simply played on? That's the hazard test above — and the answer is no.

“But the bleeding slowed after the timeout”

It's true that the opponent's scoring rate drops after a timeout — from about 2.95 pts/min before to 1.3 pts/min after. But a timeout gets called at the peak of the bleeding, and extreme runs regress toward the mean whether or not you interrupt them. The slowdown would have largely happened anyway. It's regression to the mean, not a timeout effect — don't credit the whistle for the natural cooldown.

How we checked

The crude timeout-vs-no-timeout comparison in the table is computed live from every drought this season. The hazard finding comes from a separate offline analysis of a random 1,500-game sample (1,014 timeout droughts): for each timeout drought we measured whether it broke within one possession (~35 seconds) of the timeout, and compared that to the natural rate at which any drought breaks in an equivalent window. The two rates (24.6% vs 23.5%) are effectively the same.

One honest limit: this measures whether a timeout ends the drought faster, not every downstream effect (spacing, fatigue, a set play two possessions later). It says the direct “timeout breaks the drought” claim isn't supported — not that timeouts are worthless.