History & Culture · History · about 6 min read

History of Roulette Systems

Two centuries of people writing down rules for changing their stake.

Short answer

Roulette systems have old names and many variations. Their histories are less certain than their arithmetic: on independent uniform spins, changing stakes or following past results does not remove the house edge.

Martingale in the Académie dictionary
1762
doubling definition
1878 edition
common feature
rules for choosing stakes
key distinction
a win is not proof of an edge

Martingale

The 1762 Académie française dictionary records martingalestrategyMartingaleA progression that doubles after each loss and usually returns to the base stake after a win.Read full definition → as a gambling term for staking what has been lost. Its 1878 edition explicitly describes doubling after a loss and also uses the word more broadly for betting schemes. These entries document usage; they do not identify a single inventor.

Today the familiar doubling progressionstrategyProgressionA rule that changes stakes over a sequence of bets, often using previous wins or losses.Read full definition → aims to recover previous losses and gain one starting unitmoneyUnitA base amount used to express stakes or results.Read full definition → when a win arrives. A finite bankrollmoneyBankrollThe amount set aside for play. Bankroll management concerns stakes and limits within that amount.Read full definition → or table limittableTable limitA permitted minimum or maximum stake, applied per bet, bet category or total wager according to the table.Read full definition → can stop the sequence first. The Strategy section shows how those limits affect a session.

An imagined eighteenth-century card room, evoking the gambling culture behind early betting systems.

Roulette Table Limits

D’Alembert

The progression bearing d’Alembert’s name raises the stake by one unit after a loss and lowers it after a win. The name alone does not prove who devised or endorsed the betting rule.

Its mathematical issue is straightforward: on independent spinsmathIndependent spinsSpins for which learning an earlier outcome does not change the probabilities assigned to a later one.Read full definition →, a run of one outcome does not make the opposite outcome due. A famous namesake does not change that assumption or the payoutmoneyPayoutThe amount paid on a winning wager. Check whether the quoted figure is net profit or total return including the stake.Read full definition →.

D'Alembert Roulette Strategy

Labouchère

The cancellation system uses a list of numbers to determine the next stake. A win removes entries and a loss adds an entry; the Strategy section sets out the rules and examples. Historical stories about its namesake should be distinguished from evidence that the method creates an advantage.

A sequence may clear with a profit, but a losing run can leave a growing list and larger required stakes. How often it clears depends on the starting list, bankroll, limits and stopping rule. There is no universal success percentage.

Monte Carlo RouletteLabouchère Roulette Strategy

Mathematics and simulation

A simulation can show how a specified system behaves under a chosen wheel model. It cannot test every possible rule, and a finite run need not return exactly the theoretical house edgemathHouse edgeThe casino’s expected net gain expressed as a fraction of the stake, under specified probabilities and payouts.Read full definition →.

For independent uniform single-zerowheelZeroThe 0 outcome, conventionally shown in green. Ordinary color, parity, high/low, dozen and column bets do not include it.Read full definition → spins with standard full-loss-on-zero payouts, each unit staked has expected net return −1/37. For a bounded session whose stakes are chosen using information available before each spin, adding the conditional expectations gives expected lossmathExpected lossThe probability-weighted average net loss under a stated model. With a constant edge, it equals edge times expected total stakes.Read full definition → equal to 1/37 of expected turnovermoneyTurnoverThe sum of amounts wagered over a session or period, including money wagered again after a return. Also called handle or action.Read full definition →. Stopping or changing stakes can change the distribution of results, but not create positive expectation under those assumptions. Other payouts, such as la partageruleLa partageA rule returning half of a qualifying even-money stake when zero lands. On a fair single-zero wheel it gives an edge of 1/74, about 1.35%.Read full definition →, need their own calculation.

An imagined system test with a stylized losing graph. Actual simulated sessions can rise and fall.

Myth: A Roulette Strategy That Always WinsFibonacci Roulette Strategy

Why a winning session can mislead

A demonstration can end in profit without showing that the method has an edge. Some rules favor frequent small wins at the cost of occasional large losses; others do the reverse. The proportions depend on the exact rules and limits.

When comparing systems, look at all completed sessions, total stakes, losses and stopping conditions. A selected winning example leaves out the information needed to assess the claim.

systemstake rulewhat to check
Martingaledouble after a lossmaximum stake and loss when the sequence cannot continue
D’Alembertadd after a loss; subtract after a winstake growth and the specified lower limit
Labouchèrecancel or extend a liststarting list, large stakes and unfinished sequences
Fibonaccimove through a number sequencereset rules and bankroll limits
Paroliincrease after winshow many wins trigger a reset
Sector systemscover a group of wheel pocketsactual probabilities and payout for the complete bet

Systems for sale

Systems have also been sold with elaborate names and entry rules. The examples below illustrate the arithmetic of specific rules; they are not evidence of who first invented or marketed them. Waiting for a pattern does not alter the next independent spin’s probabilities.

systemthe ideawhat actually happens
Phase Oneafter five reds, bet black on a scale of 5, 25, 70, 200, 500 and 1,000 unitsOn a uniform double-zero wheel, black still wins with probability 18/38. Six losses have probability (20/38)^6 ≈ 2.13% and cost 1,800 units.
Watching for zeroswait for 40 spins without 0 or 00, then bet both, adding a unit after each missOn independent uniform double-zero spins, both zeros miss 40 times with probability (36/38)^40 ≈ 11.50%. The next hit probability remains 2/38. With one additional unit on each zero after each miss, ten losses cost 110 units; a hit on spin eleven nets 264 units overall.
Favored 5wait until five numbers repeat, then bet thempast repeats do not predict the next independent uniform spin
ROL / BEH rotationflat even-money bets rotating through red, odd, low and black, even, higheach bet retains its stated house edge under the independent uniform model
Bartschelly’s “Guaranteed” systema Labouchère that spreads each loss over the existing numbers instead of adding it at the endthe line never grows longer; the stakes grow anyway

Stopping rules

A limit can help control the time or money spent. If it stops bets that would otherwise have been made, it reduces exposure to the house edge. It does not guarantee that the player will reach a profit target or turn a negative-expectation bet into a positive one.

Questions

How old is the word Martingale in gambling?

It appears as a gambling term in the Académie française dictionary of 1762. That record is not proof of the method’s first use.

Does a mathematician’s name validate a system?

No. The rules, probabilities and payouts determine its behavior. A namesake is not evidence of a mathematical advantage.

Can a losing-expectation system have a winning session?

Yes. Expected return describes an average under a model, not the outcome of every session.

Comic relief

A system’s name does not change the odds.

A system’s name does not change the odds.

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. Doob, J. L. “Note on probability.” Annals of Mathematics 37(2), 1936.
  4. Halmos, P. R. “Invariants of certain stochastic transformations: the mathematical theory of gambling systems.” Duke Mathematical Journal 5(2), 1939.
  5. d’Alembert, J. “Croix ou pile.” Encyclopédie, vol. 4, 1754.
  6. Schwartz, D. G. Roll the Bones: The History of Gambling. Gotham, 2006.
  7. Kavanagh, T. M. Dice, Cards, Wheels: A Different History of French Culture. University of Pennsylvania Press, 2005.
  8. Bollman, M. Mathematics of the Big Four Casino Table Games: Blackjack, Baccarat, Craps, & Roulette. CRC Press, 2021. Chapter 2, roulette.
  9. Académie française, martingale, 1762 https://www.dictionnaire-academie.fr/article/A4M0448
  10. Académie française, martingale, 1878 https://www.dictionnaire-academie.fr/article/A7M0543
Default model: European, single zero; exceptions are stated in the articlemethodology ↗sources ↗