The projection gap: a line is a price, your view is an opinion

Two numbers sit in front of you on any bet. One is the line, which is a price the market has settled on. The other is your expectation, which is an opinion about how likely the outcome really is. A literate bettor treats these as different kinds of things, and acts only when the distance between them, the projection gap, is large enough to be a real signal. The hard discipline is deciding what "large enough" means before you look at any specific bet.

The line is a price, not the truth

A line is the market's price for an outcome, and prices carry information. They already fold in public opinion, sharp money, and the built-in vig. Your job is not to accept that price and not to reject it on reflex, but to compare it against a view you formed independently.

That view has to come from somewhere honest. It might come from a model, from long-run averages, from structured reasoning about the matchup, or from some blend. What matters is that it is an actual estimate, not a feeling dressed up as one, and that you can trace where it came from. The gap between your estimate and the price is only as trustworthy as the estimate itself.

Tiny gaps are just noise

Suppose the price implies a team wins about 48 percent of the time, and your estimate lands at 49 percent. That one-point gap is almost certainly noise. Your estimate is not precise enough, and neither is the market's, for a sliver that small to mean anything real.

This is why the required gap should scale with how volatile the market is. A stable, heavily traded market prices efficiently, so a real edge there shows up as a small but meaningful gap. A thin, chaotic, low-information market swings around far more, so a much larger gap is needed before you can be confident it is signal rather than the market simply being loose. The number you demand should be bigger where the noise is bigger. What that number is, exactly, is yours to set, and you set it before you see the bet, not after.

The anchoring trap

The most common way this discipline fails is anchoring: you look at the price first, and your "independent" estimate quietly drifts toward it. You see -150, and your brain, wanting to feel reasonable, produces an estimate suspiciously close to -150. Now the gap looks small and safe, but only because you never formed a real opinion to measure against.

The defense is order. Form your estimate before you look at the line whenever you can. When you cannot, be honest that the anchor is there and treat a suspiciously tight agreement with the price as a warning sign, not a comfort. A projection gap is only meaningful if the two numbers were formed independently. If one number was born from the other, there is no gap to measure, only an echo.