The anti-favorite bias in WNBA totals.
A pattern shows up in WNBA totals that does not show up the same way in NBA totals. When a heavy favorite plays a heavy underdog, the market sets the total lower than the actual scoring environment justifies. The favorite usually scores their normal output. The underdog scores less than their average, but not by enough to compensate. The total goes over.
This is the anti-favorite bias. It is not a tip. It is a structural feature of how WNBA totals get priced.
Why it happens.
When books set a WNBA total, they have to anchor it somewhere. The natural anchor is the average pace and efficiency of both teams. For competitive matchups, that anchor is reasonable.
For lopsided matchups, two adjustments get made. The first is correct: the underdog will probably score less than usual because they are facing a better defense. The second is overcorrection: the favorite gets blowout treatment, with garbage time minutes assumed to drag down their efficiency.
The blowout adjustment is too large. In the WNBA, blowouts do happen, but they happen less often than in the NBA. Rotations are tighter, benches are shorter, and starters tend to play deeper into games even when the score is decided. The favorite's scoring usually does not collapse the way the model assumes.
What the data shows.
In matchups where the spread is 8 points or larger, WNBA games go over the total at a measurably higher rate than they go under. The effect is not enormous, but it is consistent and it is not random.
What this is not.
This is not a system. A 54 percent rate against a 50 percent baseline does not mean blind betting on lopsided overs is profitable after the vig. It means the market is shading the price slightly wrong, and a model can use that as one input among many.
The bias also does not apply to every lopsided matchup. When the favorite is on the second night of a back-to-back, or playing without a star, or facing a team that pushes pace, the dynamics change. The bias lives specifically in the standard rested favorite versus standard underdog situation.
How we use it.
The WNBA model has a separate input for matchup spread that adjusts the total projection upward in lopsided situations. The adjustment is calibrated against actual results, not theory. When the model fires a total signal in a lopsided matchup, that input has already been applied. You are not getting a raw projection. You are getting a projection that knows the market's blind spot.