Can You Really Control Dice? Here's the Math We Use to Test It
Dice control — also called rhythm rolling or controlled shooting — is one of the oldest debates in craps. Some shooters swear a practiced grip and a consistent set can shift the odds. Skeptics say it's dice, not darts, and no amount of technique beats physics on a bounced, tumbling cube. Both sides usually argue from anecdote: a hot night here, a cold streak there.
We built PRO Sight, the analytics engine inside Craps@Home, to settle exactly this kind of argument with data instead of vibes. Here's what it actually measures, and why each metric matters.
SRR — the seven-to-roll ratio
SRR is the simplest and most-cited metric in the dice control world. It measures how many rolls typically pass between sevens. In a truly random game, a seven comes up on average once every 6 rolls (an SRR of 6.0). A shooter claiming real dice influence should, over enough rolls, post a meaningfully higher SRR than that random baseline — more rolls between sevens, which in craps math means longer hands and more points made.
The catch: a handful of good hands proves nothing. Randomness produces streaks constantly — that's what randomness looks like. SRR only becomes meaningful evidence over a large sample, tracked consistently, which is exactly what a session-by-session tracker like Craps@Home is built to accumulate.
Chi-squared distribution testing
SRR tells you about sevens specifically. Chi-squared testing goes further — it compares a shooter's entire distribution of outcomes (how often each number 2 through 12 actually came up) against the distribution true randomness predicts. Two dice don't produce every number equally often — a 7 should appear about six times more often than a 2 — and chi-squared testing checks whether a shooter's real numbers match that expected curve, or whether certain numbers are showing up more or less than chance alone would produce.
This is the more rigorous test, because a shooter could theoretically post a decent SRR by pure luck while their overall number distribution still looks perfectly random. Chi-squared testing is much harder to fake with a short lucky streak.
Survival curves
A survival curve plots the probability that a hand is still alive (no seven-out yet) after each additional roll. Random dice produce a well-known, predictable decay curve — the odds of surviving to roll 10, 20, or 30 drop in a specific, calculable way. If a shooter's actual survival curve consistently sits above the random-chance curve over enough hands, that's a meaningful signal. If it tracks the random curve closely, it isn't.
Why sample size is the whole ballgame
Every one of these tools shares the same requirement: enough rolls to separate signal from noise. A great single hand — even a 30-roll hand — is well within the range random chance produces regularly across enough sessions. Real evidence, one way or the other, comes from tracking dozens of hands and hundreds or thousands of rolls under consistent conditions.
That's the practical reason most dice control debates never get resolved: nobody's tracking enough rolls, consistently enough, to run the numbers properly. It's also exactly the gap Craps@Home's voice tracking and PRO Sight were built to close — every roll called out loud gets logged automatically, hand after hand, session after session, without anyone breaking rhythm to write it down.
Put it to the test
We're not here to tell any shooter their technique doesn't work. We're here to build the tool that can actually check. If you're a dice influencer, a sharpshooter, or just someone who's convinced their grip matters — we've got an open invitation for you.
Read the Dice Control Challenge →