Clutch NBA
CLE @ NYKWed, Nov 11, 7:10 PMBOS @ TORWed, Nov 11, 7:10 PMWAS @ INDWed, Nov 11, 7:10 PMDEN @ MEMWed, Nov 11, 8:10 PMLAL @ MILWed, Nov 11, 8:10 PMSAS @ CHIWed, Nov 11, 8:10 PMBKN @ NOPWed, Nov 11, 8:10 PMORL @ UTAWed, Nov 11, 9:10 PMGSW @ HOUWed, Nov 11, 9:40 PM
Tools / Is your hit rate real?
Free · no account

Is your hit rate real?

A player went over the line 7 times in the last 10 games. Is that a real edge or noise? Enter the record and the odds: the checker compares it with the market price. Free, no account.
Observed hit rate
70%
7/10
Market probability
52.4%
Plausible range of the true rate (95%)
40% - 89%
Chance of this streak with no edge
21.4%

Consistent with chance: a record like this happens often even when the market price is right. Games needed to tell this rate apart from the market price: about 31.

What the checker calculates

The market probability comes from the odds (at -110 it is 52.4%). The checker computes how often a record at least as good happens by pure chance if the market is right (binomial test), and the plausible range of the true rate (95% Wilson interval).

With 7/10 against a 52.4% price the chance of luck is about 21%: it happens all the time.

What the data says

In a test on 238,000 NBA lines, hit rates (last 5, last 10, season, vs opponent) added no predictive power beyond the price. Hit rate is a description of the past, not a signal. Use this tool to see how much a sample really tells you.

Frequently asked questions

How many games do I need for a hit rate to mean something?

Far more than most people think: to tell a true 60% from a 52.4% price with confidence you need around 170 games.

Why not just use the last 10 games?

Ten games swing a lot: a coin-flip outcome lands 7 or more out of 10 about 17% of the time.

The Clutch terminal does these calculations for every NBA player prop, using Pinnacle's no-vig price as the reference. Free with an account.

Analyze in the terminal

More free tools