EXPLAINER Glossary Published · 5 min read · 1,193 words

What Is the House Edge? Covering 18 Numbers on Roulette: You Win 48.6% of Spins and Lose the Same 2.70%

A bet that wins almost every spin feels like a way around the wheel. We priced it exactly and ran 40,000 sessions: one miss repays the house for 1.0 wins.

What Is the House Edge? Covering 18 Numbers on Roulette: You Win 48.6% of Spins and Lose the Same 2.70%

Betting on 18 numbers in single-zero roulette doesn’t beat the house; the average loss is 0.486 units per spin, which is exactly 2.70% of the 18 units you stake each round. That loss rate is identical to a straight-up bet on a single number, despite covering half the wheel.

In a 100-spin session you lay down 1,800 units and the model predicts an average loss of 48.6 units. The median result is down 36 units; the worst 5% lose 360 units or more, while the top 5% win at least 252 units. Going 20 spins without a miss is under 0.001% likely.

WHAT WE FOUND
  • Each spin loses 0.486 units on average, which is exactly 2.70% of the 18 units wagered.
  • Over 100 spins you stake 1,800 units and the model predicts an average loss of 48.6 units.
  • In 40,000 simulated 100-spin sessions, 35.6% of the runs ended with a profit.
  • The median session finishes 36 units down; the worst 5% lose 360 units or more.
  • The probability of going 20, 50 or 100 spins without a single miss is less than 0.001%.

What the house edge is

The house edge is the share of every stake that a bet loses on average in the long run. It is set by the rules and the payouts of the game, not by the way the player bets.

Covering 18 of the 37 numbers on a single-zero wheel with 1 unit each puts 18 units on every spin and wins on 48.6% of spins. A winning spin returns 36 units for a net gain of 18; a losing spin costs all 18.

Each spin loses 0.486 units on average, which is 2.70% of the 18 units staked: the same house edge as a bet on a single number. The expected loss per unit staked is therefore 2.70% no matter how many numbers are covered.

Over 100 spins the player wagers 1,800 units and loses 48.6 units on average. In 100 spins the bet is expected to miss 51.4 times, and each miss costs 18 units. One losing spin wipes out the gain of 1.0 winning spins.

The chance of getting through 20 spins without a single miss is less than 0.001%; through 50 spins it is less than 0.001% and through 100 spins less than 0.001%.

Covering 18 numbers wins 48.6% of spins, yields a net gain of 18 units on a win, and costs 18 units on a loss.

One hundred spins of it

The median outcome is a loss of 36 units, while the bottom 5% of sessions lose 360 units or more and the top 5% win 252 units or more.

The profit-share looks low because every losing spin erases the gain of a single winning spin.

From the same simulation the bet is expected to miss 51.4 times in a 100-spin run, and each miss costs the full 18-unit stake.

The chance of avoiding any miss is vanishingly small. The model gives a probability of less than 0.001% to get through 20 spins without a miss, the same sub-0.001% figure for 50 spins, and again less than 0.001% for the full 100-spin stretch. In practice, a flawless streak is essentially impossible.

Covering more numbers merely rescales win frequency and payout: covering 18 numbers wins 48.6% of spins for 18 units, while covering 35 numbers wins 94.6% of spins for 1 unit.

The price per unit staked

Each spin of the 18-number cover loses 0.486 units on average, because the model records a loss of 0.486 units per spin.

Across a full hundred-spin run the same model reports a total loss of 48.6 units, while the player has laid down 1,800 units (18 units each spin).

The table rows illustrate that the share of stake lost never moves. A single-number wager loses 2.7 units over 100 spins; covering 18 numbers loses 48.6 units over 100 spins; covering 35 numbers loses 94.6 units over 100 spins.

Because the loss per unit staked is constant, the per-spin cost scales linearly with the amount placed.

The product is the expected loss per spin, and over any number of spins the cumulative loss will be 2.70% of the total amount you have risked.

Why covering more numbers changes nothing

Covering more numbers does not tilt the odds; it only reshapes how often you win and how much each win pays. When you place one unit on each of the numbers you choose, the chance of a winning spin rises in direct proportion to the count of numbers, while the payout per win falls by the same proportion because the roulette table always returns 36 units on a hit.

For example, covering twelve numbers wins on 32.4% of spins and gives a net gain of 24 units when it hits. Covering 18 numbers wins on 48.6% of spins and gives a net gain of 18 units when it hits.

Because the win probability and the net win amount scale together, the expected loss per unit staked stays at 2.70% in every case. Whether you bet a single number or the whole half-wheel, the house edge is calculated as a percentage of the total amount you lay on the table, not as a fixed number of units per spin.

What does change is the way a loss feels across an evening. A losing spin with 18 numbers costs the full 18 units, but such a loss is offset by wins on almost half the spins. With fewer numbers the loss per spin is only one unit, but wins are far rarer, so the same 2.70% edge is realized through many small defeats instead of a few large ones.

In the simulated data set of 40,000 hundred-spin sessions of this bet, 35.6% of the sessions ended in profit.

THE VERDICT

The house edge remains at 2.70% regardless of how many numbers you cover. That translates to an average loss of 48.6 units for every 1,800 units wagered in a 100-spin run.

The data behind this piece

Every figure in the text comes from the runs below. They are computer simulations and exact calculations, not real-money play, and they can be reproduced: the model parameters are stated and the random seed is fixed.

Covering 18 numbers on a single-zero wheel

How we ran it: Exact arithmetic for a bet of 1 unit on each of 18 numbers of a 37-pocket wheel (a straight-up win returns 36 units), plus 40,000 simulated sessions of 100 spins of that bet. Random seed: 39053

What covering more numbers changes
Numbers coveredSpins that winNet gain of a winning spinCost of a losing spinWins needed to repay one lossAverage loss per 100 spins
12.7%3510.032.7
1232.4%24120.5032.4
1848.6%18181.048.6
2464.9%12242.064.9
3081.1%6305.081.1
3594.6%13535.094.6

Questions players ask

How much does covering 18 numbers on roulette cost me on average?

A single spin with an 18-number cover loses 0.486 units on average, which is 2.70% of the 18 units you stake. Over a typical 100-spin stretch you would lay down 1,800 units and the model shows an expected loss of 48.6 units, roughly 2.7% of your total wager.

What is the probability of winning a spin when I bet on 18 numbers?

Covering 18 of the 37 numbers gives you a 48.6% chance that any given spin lands on one of your selections. The roulette wheel’s design fixes this proportion; each win still pays 36 units, so the higher hit rate is offset by a proportionally smaller payout per win.

Can I avoid losses by never missing a spin?

The model shows the odds of completing 20 consecutive spins without a single miss are below 0.001%; the same sub-one-in-a-hundred-thousand likelihood applies to 50 and 100 spins. In practice you will almost certainly encounter a losing spin well before those milestones.

How often will I be in profit over many 100-spin sessions?

In 40,000 simulated 100-spin sessions the bet ended in profit in 35.6% of cases. The median outcome was a loss of 36 units, with the bottom 5% losing 360 units or more and the top 5% gaining at least 252 units.

Check it yourself:Run the simulators yourself
Share: TelegramWhatsAppXReddit