Understanding the Double Zero Wheel: Layout and Probabilities
The American (double-zero) roulette wheel contains 38 slots: numbers 1–36, plus 0 and 00. That means any single-number bet (a straight-up) has a probability of 1/38 ≈ 2.6316% of winning; any event that covers n distinct slots has probability n/38. The presence of both 0 and 00 is the key structural difference relative to the European (single-zero) wheel, which has only 37 slots. This additional slot increases the casino’s edge because most payout schedules are defined as if there were fewer losing slots. Understanding the layout visually helps: inside the betting layout each chip placement corresponds to a set of numbers—single number squares, shared edges for splits, small boxes for corners, rows for streets, and outside boxes for dozens and even-money bets. When calculating risk, always count exact covered numbers. For example, a split covers 2 numbers → probability 2/38; a corner covers 4 numbers → 4/38; a dozen covers 12 numbers → 12/38. Because every spin is independent, probabilities remain constant across spins. Knowing the raw probabilities is the foundation for computing payouts, expected value, variance, and long-term outcomes for bankroll planning.
Odds and Payouts for Common Bets in Double-Zero Roulette
Roulette payout rules are standardized for common bet types, and in American roulette those payouts are paired with the 38-slot probability to produce the familiar house edges. Common payouts and their corresponding covered numbers:
- Straight-up (single number): covers 1 number, probability 1/38, payout 35:1.
- Split (two adjacent numbers): covers 2 numbers, probability 2/38, payout 17:1.
- Street (row of three): covers 3 numbers, probability 3/38, payout 11:1.
- Corner (square of four): covers 4 numbers, probability 4/38, payout 8:1.
- Line / Six-line (two adjacent streets): covers 6 numbers, probability 6/38, payout 5:1.
- Dozen / Column: covers 12 numbers, probability 12/38, payout 2:1.
- Even-money bets (Red/Black, Odd/Even, 1–18/19–36): cover 18 numbers, probability 18/38, payout 1:1.
There is also a special five-number bet specific to American roulette (0, 00, 1, 2, 3) that covers 5 numbers, probability 5/38, and usually pays 6:1 (this bet is notable because it yields a worse house edge than other bets). For each of these bets, you can compute the chance of winning and compare the expected return based on the payout. Most bets are intuitively “fair” in the sense that their payout scales roughly with the number of covered numbers, but because the payouts are slightly less generous than exact fair odds implied by 38 slots (e.g., straight-up pays 35:1 instead of 37:1), the casino retains an edge.
Calculating Expected Value and House Edge for Each Bet
Expected value (EV) per unit wagered on a roulette bet quantifies the long-term average outcome. For a $1 straight-up bet in American roulette: you win $35 with probability 1/38, and you lose $1 with probability 37/38. EV = (1/38)*35 + (37/38)*(-1) = (35 - 37)/38 = -2/38 ≈ -0.0526316, so you lose about $0.05263 per $1 bet on average — a house edge of 5.263%. The same EV applies to most standard bets (split, street, corner, line, dozens, even-money), because the casinos use payout multiples that create the same net shortfall: they pay as if there were 36 or 37 numbers, not 38. For a general bet covering n numbers paying at an agreed payout, the EV calculation is: EV = (n/38)*payout + (1 - n/38)*(-1). For example, a dozen (12 numbers) pays 2:1: EV = (12/38)*2 + (26/38)*(-1) = (24 - 26)/38 = -2/38. The five-number bet is an exception: with probability 5/38 you win 6:1, so EV = (5/38)*6 + (33/38)*(-1) = (30 - 33)/38 = -3/38 ≈ -0.078947, i.e., a house edge of about 7.895%. That makes the five-number bet the worst in terms of expected loss. Expressing house edge as a percentage of the wager, most bets show 5.263% loss per dollar on average. This constant negative EV makes roulette a negative-expectation game for players in the long run. Short-term variance can produce winning sessions, but over many spins the average result converges to EV by the law of large numbers.
Betting Strategy, Variance, and Bankroll Management
Because the EV of most roulette bets is negative and fixed (about -5.263% on an American wheel), no betting pattern can change the long-term expectation. Systems like Martingale or Fibonacci only change variance and risk profile: they can produce small, frequent wins but expose you to rare catastrophic losses (long losing streaks that exceed your bankroll or table limits). Variance per spin depends on the bet type—the larger the payout multiple when you win, the higher the variance. For single-number ($1 straight-up) bets, standard deviation per spin is high because wins are rare but large relative to the stake; for even-money bets, wins are frequent but small, so variance per spin is lower. A simple way to think about risk: if you want lower volatility, prefer bets that cover more numbers (dozens/columns/even-money), but your expected loss per dollar is the same, so you will still lose on average faster in dollar terms when wagering larger amounts. For bankroll management, common practical advice includes setting a session loss limit, limiting individual bet sizes (e.g., not more than 1–2% of bankroll), and treating gambling as entertainment with a predetermined cost. If you want a mathematical approach, Kelly criterion can size bets to maximize growth when you have an edge, but since roulette offers a negative edge, Kelly would recommend betting zero. In short: understand the odds and EV, choose bet types that match your risk tolerance, avoid chasing losses, and accept that no strategy can overcome the built-in house edge over the long run.
