Roulette Lab · Lab · Context · about 5 min read

Roulette Computers

How a computer times the ball, who built the first ones, and how accurate it has to be.

Short answer

A physical predictor estimates the ball and rotor’s motion. Thorp describes a 1961 device and casino trials; those historical results are not a guarantee for a current wheel. Betting deadlines and device laws are separate constraints.

first built
1961, Thorp and Shannon
what it measures
ball and rotor timing
model break-even error
about 17.2 pockets, single zero
banned in Nevada
since 1985

How the first roulette computer system was built

In his 1998 account, Edward Thorp dates the idea to 1955 and describes working with Claude Shannon at MIT in 1960–61. They studied a regulation wheel on a slate billiard table in Shannon’s basement, using a clock and strobe light to improve their timing.

By June 1961 they had a cigarette-pack-sized analog computer with twelve transistors. Toe-operated switches timed the ballequipmentBallThe small ball that settles in a numbered pocket to determine a physical roulette result.Read full definition → and rotorequipmentRotorThe rotating part of a physical roulette wheel, carrying the numbered pockets.Read full definition →; tones identified one of eight sectorswheelSectorA continuous arc of neighboring pockets around the wheel, rather than a group defined by the betting layout.Read full definition → to bet on. Thorp reports testing it in Las Vegas that summer with ten-cent chipsmoneyChipA token used to place a wager, with a printed or assigned value. Also called a check in some casino terminology.Read full definition →. Repeated breaks in the earpiece wires prevented sustained serious betting. Distance and competing work later kept the project from continuing.

This is the inventor’s retrospective account, not an independent casino trial report. His 1969 paper gives 1957 for the initial idea; the later account gives 1955. Both describe work completed by 1961.

An imagined reconstruction of the basement experiments described by Thorp, with a wheel, billiard table and strobe.

From digital computers to shoes

Bradley Rhodes’s MIT wearable-computing timeline credits Alan Lewis with a digital roulette computer carried in a camera case in 1972. It describes an observer using the computer and relaying predictions by radio to a bettor. That brief record does not establish the device’s casino return.

Doyne Farmer’s own account describes the Eudaemons’ move from a computer worn under the arm to one concealed in a shoe. He wrote its program in machine code. Toe switches supplied timing measurements, and vibrating signals passed the prediction to the player placing the bets.

Farmer reports an advantage of about 20%, alongside persistent hardware problems and small stakes. This is his retrospective estimate, not a published spin-by-spin trial. The Heinz Nixdorf MuseumsForum’s account describes radio interference disrupting the shoe system during its November 1981 test. Historical results do not establish what a device could achieve today.

An imagined observer-and-bettor arrangement; the balcony scene is not a verified reconstruction of Lewis’s trial.

Why late bets mattered

A strategy that measures the current spin cannot use those measurements if betting has already closed. Thorp’s account describes late betting in games he observed, not every modern studio. Nevada’s device provision dates to 1985; its current wording includes an approval exception.

Why operating conditions matter

A predictor fitted to one setup needs fresh validation when the wheel, ball or operating conditions change. A change does not by itself prove that prediction has become impossible; what matters is the resulting prediction error and whether it still supports the proposed bet.

In his 1998 account, Thorp identifies closing bets before launch as a way to prevent his method from using the current spin. This is a constraint on that method, not evidence that every casino follows the same procedure. Check the actual betting deadline rather than assuming that a historical account describes today’s game.

An imagined ball change at a crowded table; the scene does not establish a historical event or identify the spectators.

Can Roulette Dealers Hit Numbers? The Myth

How good the prediction must be

The table uses a mathematical model: prediction error is a centered normal distribution wrapped around 37 pocketsequipmentPocketA compartment on the rotor in which the ball can settle. The number of pockets depends on the wheel.Read full definition →, each one unitmoneyUnitA base amount used to express stakes or results.Read full definition → wide. Expected return follows by integrating the selected pocket interval and applying 35:1 payoutsmoneyPayoutThe amount paid on a winning wager. Check whether the quoted figure is net profit or total return including the stake.Read full definition →. This is a model illustration, not measured accuracy for a current device. At σ = 10 it gives about +43.9% for one pocket or +42.5% for five centered pockets. Break-evenmathBreak-evenA result with no net gain or loss, or a threshold at which expected profit is zero. Specify which meaning applies.Read full definition → is about 17.22 pockets for one and 17.16 for five. A different error distribution gives different values.

typical error σ, pocketsone numberfive neighbors
0+3,500%+620%
2+611%+468%
4+258%+237%
6+139%+133%
8+79%+77%
10+44%+42%
12+22%+21%
14+8.8%+8.5%
16+2.1%+2.0%
17+0.3%+0.2%
18−0.9%−0.9%
20−2.1%−2.1%
no prediction−2.70%−2.70%
Independently calculated for a centered wrapped-normal error on a 37-pocket wheel. Returns are per unit wagered; the model and its σ must be validated before any practical inference.

Deceleration

Questions

Does timing the animation predict an RNG result?

A physical ball model does not apply to a software-generated outcome merely displayed by an animation.

Does a visible launch and late deadline guarantee vulnerability?

No. Those conditions may allow measurements, but do not establish sufficient predictive accuracy or permission to use a device.

Fun with Math

Try these yourself first. Each solution is folded; open it when you are ready.

  1. A neighborsbetNeighborsPockets adjacent to a selected number around the wheel. A neighbors bet includes a chosen number of them on each side.Read full definition → call puts one chipmoneyChipA token used to place a wager, with a printed or assigned value. Also called a check in some casino terminology.Read full definition → on a number and on the two pocketsequipmentPocketA compartment on the rotor in which the ball can settle. The number of pockets depends on the wheel.Read full definition → either side of it on the wheel: five chips. How often must a prediction land in those five pockets to break even, and what does it earn if it is right one time in five?

    Show the solution

    Answer It breaks even at 5 in 36, 13.9%, against 13.5% by chance alone. Right one time in five, it returns 44% more than it stakes.

    If the call hits, one chip wins 35 to 1 and comes back with its winnings, 36 chips, and the other four are lost. So a hit returns 36 for 5 staked, and the call breaks even if it hits 5 times in 36, 13.9%. By chance it hits 5 times in 37, 13.5%. That small gap, less than half a percentage point, is where the whole 2.70% comes from.

    At one hit in five, 0.2 × 36 = 7.2 chips come back for every 5 staked, +44%. At one in four, +80%.

    These are expected returns under the stated success probabilities and payoutsmoneyPayoutThe amount paid on a winning wager. Check whether the quoted figure is net profit or total return including the stake.Read full definition →. A prediction would need independently measured accuracy; this calculation does not establish that any device or player achieves it.

  2. Consider four independent, zerowheelZeroThe 0 outcome, conventionally shown in green. Ordinary color, parity, high/low, dozen and column bets do not include it.Read full definition →-mean prediction errors whose standard deviationsmathStandard deviationThe square root of variance: a measure of dispersion in the same units as the result.Read full definition →, measured in pockets, are 1.4, 5.5, 7 and 6. The first two represent timing errors. What is the combined standard deviation, and how much would eliminating both timing errors reduce it?

    Show the solution

    Answer About 10.8 pockets overall, versus 9.2 without the timing errors: a reduction of about 1.6 pockets.

    For independent errors, variancesmathVarianceThe expected squared distance of a result from its expected value. It measures dispersion in squared units.Read full definition → add. The combined standard deviation is √(1.4² + 5.5² + 7² + 6²) = √117.21 ≈ 10.83 pockets.

    Without the timing errors it is √(7² + 6²) = √85 ≈ 9.22 pockets. The reduction is about 1.61 pockets, not 1.4 + 5.5.

    This is a hypothetical error budget. Correlated errors require covariance terms, and a standard deviation alone does not determine a prediction’s hit rate or betting return.

Comic relief

A prediction cannot guide a bet after betting has closed.

A prediction cannot guide a bet after betting has closed.

Sources

  1. Thorp, E. O. “Optimal gambling systems for favorable games.” Review of the International Statistical Institute 37(3), 1969.
  2. Thorp, E. O. The Mathematics of Gambling. Gambling Times / Lyle Stuart, 1984.
  3. Thorp, E. O. “The invention of the first wearable computer.” Second International Symposium on Wearable Computers, IEEE, 1998.
  4. Bass, T. A. The Eudaemonic Pie. Houghton Mifflin, 1985.
  5. Small, M. & Tse, C. K. “Predicting the outcome of roulette.” Chaos 22(3), 2012.
  6. Goldman, N. Great Gambling Scams. John Blake, 2007.
  7. Nevada Revised Statutes, chapter 465. Section 465.075, use of a device to project the outcome of a game (1985).
  8. Edward O. Thorp, The Invention of the First Wearable Computer https://www.cs.virginia.edu/~evans/thorp.pdf
  9. Nevada Revised Statutes 465.075 — checked November 15, 2026 https://www.leg.state.nv.us/NRS/NRS-465.html#NRS465Sec075
  10. Bradley Rhodes, A brief history of wearable computing https://www.media.mit.edu/wearables/lizzy/timeline.html
  11. J. Doyne Farmer, Roulette https://www.doynefarmer.com/roulette
  12. Heinz Nixdorf MuseumsForum, Shoe computer https://www.hnf.de/en/permanent-exhibition/exhibition-areas/computers-for-everyone-1980-to-2000/hackers-reaching-out-for-forbidden-knowledge/shoe-computer.html
Default model: European, single zero; exceptions are stated in the articlemethodology ↗sources ↗