Wheel-Bias Detective
Inspect simulated logs from five wheels and look for a planted sector—or select the wheel generated with equal probabilities.

Each wheel has 3,000 simulated spins. Four models have a planted sector and one is uniform. A correct choice earns 2 points, or 1 if you used the five-pocket scan. This is a learning game, not a test that certifies a real wheel.
Wheel 1 of 5 · 3,000 spins
Scoring
Choose any pocketequipmentPocketA compartment on the rotor in which the ball can settle. The number of pockets depends on the wheel.Read full definition → in the planted sectorwheelSectorA continuous arc of neighboring pockets around the wheel, rather than a group defined by the betting layout.Read full definition →, or choose fair for the uniform model. A correct answer earns 2 points, reduced to 1 after using the sector-scan hint. The game scores against the model used to generate the data; it does not prove that the visible sample identifies that model with certainty.
Read the band carefully
Under the uniform model, each pocket has expected count 3,000/37, about 81.08, and standard deviationmathStandard deviationThe square root of variance: a measure of dispersion in the same units as the result.Read full definition → about 8.88. The two-standard-deviation band is a reference for an individual pocket. With 37 pockets on display, some can fall outside it by chance. Selecting the hottest pocket after looking at them all needs a different assessment from testing a pocket chosen in advance.
What the sector scan does
The hint checks all 37 overlapping five-pocket arcs and highlights the largest standardized excess. That selected maximum is not a preselected-sector test, and its displayed score is not a multiple-testing-adjusted probabilitymathProbabilityA measure of how likely an event is, from zero to one. Percentages express the same scale from 0% to 100%.Read full definition →. The planted sectors can contain three, four or five pockets, so the five-pocket scan is an aid rather than an exact template.
From a game to evidence
The global chi-square testlabChi-square testA test comparing observed category counts with those expected under a specified model, such as equal pocket probabilities.Read full definition → asks about agreement with equal pocket probabilities. A small p-valuelabp-valueUnder a specified null model, the probability of a test statistic at least as extreme as the observed one, using the chosen test.Read full definition → does not by itself locate a bias, explain its cause or establish a profitable wager. Failure to reject uniformity does not prove fairness. A real investigation needs a planned method, uncertainty estimates and new data for validation.
Questions
A planted pocket has that many times the probability of a non-planted pocket in the same model. Probabilities are then normalized across all 37 pockets. It is not the multiplier over 1/37.
No. Detection depends on the size and shape of the bias, the test, the error threshold and the desired power. This fixed sample size is a game setting, not a universal evidence requirement.
Comic relief

A high count is a clue to examine, not proof of a profitable bias.
Sources
- Barnhart, R. T. Beating the Wheel.
- Pearson, K. “On the criterion that a given system of deviations from the probable…”