The Closing Line Already Knows
I pulled both the opening and closing moneyline for 22,854 NBA games and checked them against who actually won. The headline isn't that the market is beatable. It's that the line gets smarter the longer you wait — and when the money moves it, it moves toward the team that wins.
1. The market is just… honest
Start with the boring, important part. If you take every game, sort it by the price the market gave each team, and check how often those teams actually won, you get a near-perfect match. A team the market made a 30% shot wins about 30% of the time. A 70% favorite wins about 70%. No level is systematically off.
No-vig closing price vs. real win rate, both sides of every game (about 45,000 data points). Hover a dot for the exact numbers.
This is what a sharp market looks like. There's no price tier where the public is getting fleeced or where easy money is sitting. Betting all favorites, all home teams, all dogs — none of it beats a line this well-calibrated. So if the price is already honest, where's the edge? In when you get it.
2. The line gets sharper as tip-off nears
The opening line and the closing line are both honest on average — but they are not equally accurate. There's a standard score for "how good were these probabilities," called the Brier score, where lower is better. The closing line beats the opening line: 0.202 at the close vs 0.210 at the open. Same story on log-loss (0.606 → 0.589).
Translation: as tip-off approaches and the sharp money, injury news, and confirmed lineups come in, the price tightens onto the truth. And here's the kicker — it gets cheaper too. The book's built-in margin (the vig) shrinks from about 4.0% at the open to 3.1% at the close. You wait, you get a sharper number AND you pay less for it.
No-vig de-vigging of the open and close moneyline pairs. Lower bar = less house edge.
3. When the line moves, it moves toward the winner
Lines don't sit still. About 60% of games saw the no-vig price move at least 2 points before tip-off, with the average game moving almost 5 points. The interesting question is: when the line moves, is it chasing noise, or is it learning something?
It's learning. Take the games where money came in on the home team (the line moved toward them). The opening price said those home teams were a 58% bet. They went on to win 65% of the time. And the closing price? It had already moved to 65% — right on the nose. The opening number was stale. The move corrected it. The close knew.
Home teams the line moved toward. The opening price under-rated them; the close caught up.
The mirror image holds on the other side: in games where the line drifted away from the home team, the road side won 52%, versus the 46% the open implied. Either way, the direction of the move was the direction of the truth.
4. This is why we grade against the close
Put it together. The closing line is the sharpest, cheapest, most honest number the market produces. It beats the open. The moves that get it there point at the actual winners. So there is exactly one clean way to prove you're ahead of the market: get a better price than the close. That's closing line value, and it's the one number that predicts whether you're actually good or just running hot.
Win rate can lie — you can go 6–4 on coin flips and feel like a genius. But if you keep beating the closing line, you were holding a better number than the sharpest snapshot the market ever made. Over time, that's the thing that pays. It's why our whole scoreboard is built around CLV, not the win-loss record.
The honest caveats
- The steam stat has hindsight in it. We only know which way the line moved after it moved. So "the move pointed at the winner" is a fact about the past, not a system — you can't know at the open which way the close will land. It proves the close carries real information; it does not mean chasing steam prints money after vig.
- Single open and single close. One opening price, one closing price per game. Shopping the best number across books is its own separate (real) edge on top of this.
- It's descriptive. "The close is sharper" is a measurement, not a promise. The books know all of this too — that's exactly why the close is so hard to beat.