Late Whistle

Findings

What the reports show

97,493 graded plays, Mar 1, 2015 to Nov 18, 2026. Incorrect calls and missed calls are always measured separately. Anything compared across many teams or officials is corrected for multiple comparisons.

Call types 6 vs 482

Defensive three seconds is almost never called late

6 correct calls since 2015, against 482 the league graded as missed. Traveling is missed 3 times for every correct call.

Missed-call rate is missed calls divided by all graded non-calls of that type. Take fouls, by contrast, had 2 errors in 809 calls.

Missed-call rate 41.7% (38.9–44.5%) All call types
Score state 60.5%

Late errors lean toward the team that is behind

Final 24 seconds of games decided by three or fewer: 525 errors helped the trailing team, 343 the leader.

Score before each play comes from the play-by-play running score, checked against the final score. Trailing teams defend more aggressively late, which creates more borderline contact; this shows who errors help, not why.

57.2–63.7% · describes who errors help, not intent
Home court 50.3%

No home-court tilt in the regular season

Of 6,120 regular-season errors, 50.3% helped the home team. In the playoffs it is 56.5%, mostly home court rather than seeding.

An incorrect call helps the team that was wrongly whistled against; a missed call helps the team that got away with it. Splitting playoff games by series game separates home court from seeding.

49.1–51.6%
Size 7.6%

Fouls on tall players are missed more often

Fouls on players 6-10 and up go uncalled 7.6% of the time, against 4.9–5.5% for shorter players. Big on big is the widest gap.

Heights are listed heights (nba.com for current players, Wikidata otherwise). Tall players are in different kinds of contact (post-ups, screens, box-outs), so this describes where misses happen, not why.

Odds ratio 1.08 per 3 inches, adjusted for foul type and season
Stars 1.00

No star treatment

A foul by an All-NBA player is missed as often as anyone else's, adjusted for foul type, season and height.

Stars are All-NBA first to third team that season or the season before. Models adjust for foul type, season and both players' heights.

Odds ratio 0.91–1.10
Close games 35.5%

A third of one-possession games had a late error that mattered

743 of 2,094 games decided by three or fewer had an error in the final 24 seconds that changed free throws, possession or points.

Only errors whose direct consequence is free throws, possession or points count. The share fell after 2022-23 mostly because the league lists fewer missed calls.

Lower since 2023-24, as the league lists fewer missed calls
Over time 1.9–5.0%

Incorrect calls have not improved

The share of late whistles graded incorrect moves between 1.9% (2016-17) and 5.0% (2020-21) with no trend.

Incorrect-call rate is incorrect calls divided by all graded whistles. Missed-call rates are not compared across seasons because the league changed what it grades.

Full seasons only; intervals on the chart below
Teams 0 of 30

No single team gets the calls

After correcting for 30 comparisons, no team's share of favorable errors stands out from an even split.

Each team's share of the errors that involved it and went its way, tested against an even split, with Benjamini-Hochberg correction across all teams.

Closest: q = 0.066 Every team
Officials p = 0.33

Officials' accuracy does not differ beyond chance

Adjusted for call type, no official's incorrect-call rate stands out. How many late fouls each official calls does vary, which may partly reflect court position and crew role.

Only fouls matched to the calling official with high confidence, from 2016-17 on. Officials are not named; see the Officials page for the spread of call volume.

66 officials with 100+ attributed calls Officials

Incorrect calls have not improved

Share of late whistles graded incorrect, by season, with 95% intervals

Seasons are labeled by the year they end. 2014-15 starts in March. The current season, in orange, is to date, so its interval is wide.

Claims that don't hold up

NO
The last 30 seconds are called tighter
The raw gap (2.6% vs 4.7%) is intentional fouls. Leave them out and both windows are 4.9%.
NO
National TV games are different
Incorrect calls 3.7% vs 3.6%. Missed calls 7.8% vs 8.3%.
NO
Playoff officiating is worse
Incorrect calls 4.2% vs 3.6% in the regular season; the intervals overlap.

Also tested and not supported: any single team getting the calls, star treatment, and differences in accuracy between officials.