Skip to main content
BestSlip

Build smarter slips. Compare better odds. Track your performance.

Predictions

Bookmaker Track Record

The other two views are about prices on the board now. This one is about prices that have already closed, and whether the books and the market were any good.

152 closed games from Nov 12, 2026 to Nov 16, 2026. Every figure below is computed from prices BestSlip recorded and timestamped itself, and nothing is backfilled.

Moneyline — pricing ahead of the market

How far each book's first recorded price sat from where the market closed. Nearest the top is most accurate, which is also the worst price to bet into.

BookDistance (pp)Value (pp)95% intervalGames
BetMGM0.6-0.0±0.0*25·50
Pinnaclebenchmark0.8+0.0±0.0*10·20
ESPN BET0.9-0.0±0.0*12·24
DraftKings0.9+0.0±0.0*26·52
FanDuel1.0-0.0±0.0*26·52

Distance is how far a book's first recorded price sat from the closing consensus of the other books, unsigned and in points of implied probability. Smaller means the book was already pricing where the market ended up. It is the leadership figure and the table's sort order.

Value is the same comparison, signed: positive means the early price was better than the close. A book can lead the market and show no value — a price that is already right has nothing left in it — so these two columns tend to oppose each other, and neither is a recommendation. The books nearest the top are the ones worth benchmarking against and the worst ones to bet into.

The first recorded price is not an opening line. BestSlip saw it the first time it looked, which is a different claim; where a market was already trading, the figures here understate how far the book had moved before it was seen.

Marked * where the interval contains zero, meaning the book is not measurably different from the market over this many games. Intervals are computed per game rather than per selection, because the sides of a market move together and counting them separately would narrow every interval for no reason.

Point Spread — pricing ahead of the market

How far each book's first recorded price sat from where the market closed. Nearest the top is most accurate, which is also the worst price to bet into.

BookDistance (pts)Value (pts)95% intervalGames
DraftKings0.28+0.01±0.06*35·69
ESPN BET0.32+0.01±0.05*40·79
BetMGM0.350.00±0.00*35·70
Pinnaclebenchmark0.36-0.11±0.46*14·27
FanDuel0.44+0.01±0.05*41·81

Distance is how far a book's first recorded price sat from the closing consensus of the other books, unsigned and in line points. Smaller means the book was already pricing where the market ended up. It is the leadership figure and the table's sort order.

Value is the same comparison, signed: positive means the early price was better than the close. A book can lead the market and show no value — a price that is already right has nothing left in it — so these two columns tend to oppose each other, and neither is a recommendation. The books nearest the top are the ones worth benchmarking against and the worst ones to bet into.

The first recorded price is not an opening line. BestSlip saw it the first time it looked, which is a different claim; where a market was already trading, the figures here understate how far the book had moved before it was seen.

Marked * where the interval contains zero, meaning the book is not measurably different from the market over this many games. Intervals are computed per game rather than per selection, because the sides of a market move together and counting them separately would narrow every interval for no reason.

Totals — pricing ahead of the market

How far each book's first recorded price sat from where the market closed. Nearest the top is most accurate, which is also the worst price to bet into.

BookDistance (pts)Value (pts)95% intervalGames
BetMGM0.220.00±0.00*39·78
DraftKings0.280.00±0.00*34·68
ESPN BET0.350.00±0.00*39·78
FanDuel0.400.00±0.00*40·80
Pinnaclebenchmark0.440.00±0.00*13·26

Distance is how far a book's first recorded price sat from the closing consensus of the other books, unsigned and in line points. Smaller means the book was already pricing where the market ended up. It is the leadership figure and the table's sort order.

Value is the same comparison, signed: positive means the early price was better than the close. A book can lead the market and show no value — a price that is already right has nothing left in it — so these two columns tend to oppose each other, and neither is a recommendation. The books nearest the top are the ones worth benchmarking against and the worst ones to bet into.

The first recorded price is not an opening line. BestSlip saw it the first time it looked, which is a different claim; where a market was already trading, the figures here understate how far the book had moved before it was seen.

Marked * where the interval contains zero, meaning the book is not measurably different from the market over this many games. Intervals are computed per game rather than per selection, because the sides of a market move together and counting them separately would narrow every interval for no reason.

Was the market right?

The closing consensus with the margin removed, graded against the final score. Aggregated by league and market, and not any finer.

No settled game yet has both a final score and a closing market every book agreed the shape of. Grading needs the score, at least two books that priced the whole market, and — on a handicap or a total — one number they all closed on.

Why there is no accuracy ranking by team

Because it would be a ranking of noise that looked exactly like a ranking of skill. The arithmetic is the whole argument, and it is not close.

Grading a forecast against a result gives one score per game, and those scores are spread across most of the available range — a well-calibrated market still scores near zero on the games it called and near one on the games it did not. The interval on the mean shrinks with the square root of the count, so a cell needs 30 games before it can be bounded at all and several times that before the bound is narrower than the differences a reader would want to act on. A league accumulates a season of games. A team accumulates a thirty-eighth of one.

An accuracy percentage is worse than a Brier score on identical data, not merely different. It throws away the probability and keeps only whether the favourite won, which turns a graded forecast into a coin flip — the highest-variance measurement available. A book that says “home win, 95%” and a book that says “home win, 51%” score the same accuracy, and only one of them told you anything.

The team-level question does have an answer, and it is on the game pages rather than here: price lean, which measures how far a book's prices on a team sit from the rest of the market. It needs no results at all, it is a continuous paired measurement rather than a binary one, and it reaches a verdict in 10 games — the same reason the closing-line figures above need only 10 rather than hundreds.

Bet Slip

Loading slip…