NBA Rematch Totals: Does Last Meeting's Over/Under Repeat?

Every bettor has seen it: two teams grind out a game far under the number, and the next time they meet the total drops and the public piles on the under again. I wanted to know whether that memory is worth anything. So I checked 6,759 regular-season rematches from 2017-18 to 2025-26, each with a posted total for both meetings.

The headline answer is that the trend does not repeat. After a 10+ point under, the rematch went over 52.4% of the time, against 49.9% for all rematches. After a 10+ point over, the rematch went over only 49.2%. If you want the daily card, my free basketball picks page carries that. This file is about the trigger and what follows it.

Aftermath · 01Big unders lead to more overs in the rematch

This chart splits every rematch by how the last meeting landed against its total. The two highlighted bars are the story. After an under by 20 or more, and after an under by 10-19.5, the rematch went over 52.4% in both bands. The lowest bar of the six sits after an over by 10-19.5, at 47.6%.

Taking every 10+ point under together gives 1,948 decided games at 52.4%. That is fair odds of about −110, and backing the over returned +0.1% at standard −110 pricing. Read that plainly: it covers the juice and not much more.

So I don't treat this as a money machine. I take it as a warning against the lazy read. A slow, ugly first meeting tells you very little about how many possessions the rematch will produce, and the over side is where the overreaction gets priced.

Aftermath · 02The line barely moves while scoring swings back

This is the chart I would pin above my desk. After a 20+ point under, the posted total for the rematch moved just −1.3 pts, while actual scoring changed by +28.4 pts. After a 20+ point over, the line moved +2.3 pts and scoring changed by −26.7 pts. The middle bands show the same mirror shape, with the line drifting slightly towards the last result while actual scoring swings well back the other way.

The market is not chasing last game's result. It shades the number a touch towards it, and the result reverts to normal anyway. Most of that reversion is shooting variance. One cold night from three can sink a total that the pace profile said should fly.

The useful part is the direction of the small line move. After a big over the book nudges the total up, scoring comes down, and the under returned −3.0%. The extra points are not enough to make the under a play, but they are enough to stop the over being one.

Aftermath · 03Longer gaps between meetings strengthen the under bounce

Schedule always matters to me, so I broke rematches down by days since the last meeting. With a gap of 45+ days, the rematch over rate was 50.6%, rising to 53.6% after a 10+ point under. The 15-44 days row tells a similar story at 53.8% after a big under. The 3-14 days row is the odd one out, with overs scarce across the board at 47.6% overall.

Two-game sets are the curiosity. Across 389 games the over hit 48.3%, but after a 10+ over in game one it hit 53.4%. That row comes mostly from 2020-21 onward and is thin. My honest caveat covers the whole file: the edge after a big under is a few points, well inside what one or two seasons of noise can produce.

Aftermath · 04Switching venues keeps the under effect slightly stronger

Finally, venue. With the same home team as last time, rematches after a 10+ under went over 51.1% across 455 games. With the venue switched, that rose to 52.8% across 1,522 games. A different floor, different crowd and different travel spot all help break the link with last game's flow.

After a 10+ over, venue barely registered: 49.8% at the same venue and 49.1% with it switched. The over-side reversion holds wherever the game is played.

Written byMarcus Sullivan

I write NBA previews with a strong focus on pace, matchup advantages, and how the modern game shifts betting markets.

View profile →