Fibonacci System in Roulette: How Often It Wins and What It Costs

The Fibonacci system is the gentle cousin of the Martingale. You bet on red or black and follow the sequence 1, 1, 2, 3, 5, 8, 13 …: one step up after a loss, two steps back after a win. The stakes grow much more slowly than with doubling, and that is why many players see it as the safer choice. We simulated one million Fibonacci sessions to check how much safer it really is.

The short answer: with a bankroll of 200 units and a target of +50, Fibonacci ended 66% of sessions in profit and lost the whole bankroll in 12%. That is fewer lost bankrolls than the Martingale (16%), but it also reached the target less often, needed far more spins and lost more money per session on average. Like every other way of betting on roulette, it lost 2.7% of what it staked. Try your own settings in the simulator right below.

66%
Of sessions ended in profit (bankroll 200, target +50)
12%
Of sessions lost the whole bankroll
247 spins
Median time to reach +50 units (Martingale: 99)
2.7%
Lost on average per unit staked

1. Try the Fibonacci system in the simulator

The simulator starts with the Fibonacci system and our standard settings. Change the bankroll, the base bet, the table limit or the target and see how the share of good and bad sessions moves. Switch to the Martingale or any other system to compare them on the same settings.

2. How the Fibonacci system works

Like the Martingale, the Fibonacci system uses an even-money bet: red or black, odd or even, or low or high (1 to 18, 19 to 36). All three pay 1:1 and lose on zero. Your stakes follow the Fibonacci sequence, in which every number is the sum of the two before it:

  • Start with 1 unit, the first number of the sequence.
  • After a loss, move one step forward and bet the next number.
  • After a win, move two steps back. At the start of the sequence you simply stay at 1 unit.
Bet number after a run of lossesStake (units)Total staked so farChance of getting this far
111100%
21251.4%
32426.4%
43713.5%
55127.0%
68203.6%
713331.8%
821540.9%
934880.5%
10551430.2%
11892320.1%
The Fibonacci ladder on a single-zero wheel. The last column is the chance that all bets before this one lost (zero counts as a loss).

Compare that with the Martingale, where the eighth bet in a row is already 128 units. With Fibonacci the eighth bet is 21 units, and after ten losses you have staked 143 units instead of 1,023. That slow growth is the whole appeal of the system. The catch is in the second rule: a win only takes back the last two losses, and the series goes on.

A typical series

SpinStakeResultRunning totalNext stake
11Loss−11
21Loss−22
32Loss−43
43Loss−75
55Win−22
62Win01
71Win+11
Four losses and three wins with the Fibonacci. With the Martingale, the first win after four losses would already have ended the series at +1.

After four losses, the first win only brings you back to −2. You need three wins in total to get back to +1. With the Martingale, one win would have been enough. This is why the Fibonacci system needs so many spins: most of the time it is slowly climbing back to where it started.

3. Where the Fibonacci system breaks

With a bankroll of 200 units, the Fibonacci system can place ten bets in a row: 1, 1, 2, 3, 5, 8, 13, 21, 34 and 55 units, 143 units in total. The eleventh bet would be 89 units, but only 57 are left. So ten losses in a row cost 143 units. That happens in 0.13% of series, about one in 784, much more rarely than the seven losses that break a Martingale with the same money.

But the Fibonacci system does not need ten losses in a row to get into trouble. Because a win only moves two steps back, the ladder keeps climbing in any bad stretch where you lose more than two bets for every win, for example loss, loss, loss, win, loss, loss, loss, win. In our test, 33% of all sessions reached a point where the next stake was more than the money left or above the table maximum. Of the sessions that lost the whole bankroll, only 45% ever had ten losses in a row. The rest got there through such stretches.

Many players do not know the limit of their table before they start. Check it: with a low maximum the ladder breaks earlier. At a table maximum of 50 units, for example, the tenth stake (55 units) is not allowed, so the ladder ends after nine bets.

4. What our simulation shows

Our standard test is the same as on our Martingale page: bankroll 200 units (for example $1,000 with $5 bets), table maximum 100 units, bets on red, zero loses, and the session ends at +50 units, when the money runs out, or after 300 spins. When the next stake was not possible, the player took the loss and started again at 1 unit. We played this one million times.

40 sessions, spin by spin

Result after each spin in 40 Fibonacci sessions
Reached +50 (18 of 40)Stopped after 300 spins (17 of 40)Lost the bankroll (5 of 40)
40 Fibonacci sessions: result after each spin −200 −150 −100 −50 0 +50 0 50 100 150 200 250 300 Session 1: ended −91 units after 300 spins Session 2: ended +50 units after 216 spins Session 3: ended +45 units after 300 spins Session 4: ended −93 units after 300 spins Session 5: ended +50 units after 297 spins Session 6: ended +50 units after 248 spins Session 7: ended −96 units after 300 spins Session 8: ended −99 units after 300 spins Session 9: ended −100 units after 300 spins Session 12: ended +50 units after 189 spins Session 13: ended −93 units after 300 spins Session 14: ended +50 units after 229 spins Session 15: ended +50 units after 252 spins Session 16: ended +30 units after 300 spins Session 17: ended −193 units after 300 spins Session 18: ended +50 units after 247 spins Session 20: ended −103 units after 300 spins Session 21: ended +47 units after 300 spins Session 22: ended −98 units after 300 spins Session 23: ended +50 units after 239 spins Session 24: ended +50 units after 284 spins Session 25: ended +50 units after 281 spins Session 26: ended +50 units after 269 spins Session 27: ended +50 units after 286 spins Session 28: ended −179 units after 300 spins Session 30: ended +50 units after 286 spins Session 32: ended +50 units after 226 spins Session 33: ended +50 units after 185 spins Session 34: ended −89 units after 300 spins Session 35: ended +50 units after 232 spins Session 36: ended +26 units after 300 spins Session 37: ended −105 units after 300 spins Session 38: ended +50 units after 243 spins Session 39: ended +47 units after 300 spins Session 40: ended +50 units after 183 spins Session 10: ended −200 units after 241 spins Session 11: ended −200 units after 264 spins Session 19: ended −200 units after 65 spins Session 29: ended −200 units after 220 spins Session 31: ended −200 units after 209 spins Spin Result (units)
The first 40 of one million simulated sessions in the standard test, not selected. Blue: target reached. Grey: stopped after 300 spins. Orange: bankroll lost. Hover over a line for its result. Source: GamblingSociety simulation.

The lines climb slowly. Bad stretches pull them down in deep dips that the system often wins back over the next few dozen spins, but sometimes a dip ends in a broken ladder and a loss of more than 100 units within about ten spins. Of these 40 sessions, 18 reached the target, 5 lost the bankroll and 17 were still going after 300 spins. Of those, 12 were around 90 units or more behind and 5 were ahead but short of the target.

Fibonacci against Martingale

Fibonacci and Martingale in the same test
FibonacciMartingale
Fibonacci and Martingale in the same test 0% 25% 50% 75% 100% Target +50 reached: Fibonacci 48.2%, Martingale 62.3% 48% 62% Target +50 reached Ended ahead: Fibonacci 65.9%, Martingale 71.1% 66% 71% Ended ahead Whole bankroll lost: Fibonacci 11.6%, Martingale 16.2% 12% 16% Whole bankroll lost Bankroll 200 units, table maximum 100, target +50, at most 300 spins
Both systems with a bankroll of 200 units, table maximum 100, target +50, at most 300 spins, one million sessions each. Hover over a group for the numbers. Source: GamblingSociety simulation.

The Fibonacci system lost the bankroll less often: 12% against 16%. That is the safer part. Everything else went the other way. It reached the target in 48% of sessions instead of 62%, and it needed a median of 247 spins to get there, compared with 99 for the Martingale. Many sessions ran out of time while still behind: 22% ended with a loss but with money left, against 13% for the Martingale.

Because the Fibonacci system plays longer and with more money on the table, it also staked more: 718 units per session on average, against 538. At 2.7% per unit staked, that made it more expensive. The average session lost 19.4 units with the Fibonacci system and 14.6 with the Martingale.

A smaller target helps most

Target reached and bankroll lost, by target size
Target reachedWhole bankroll lost
Fibonacci: target reached and bankroll lost, by target size 0% 25% 50% 75% 100% Target +10 units: reached in 90.6%, whole bankroll lost in 3.6%, average result −5.2 units 91% 4% +10 units Target +20 units: reached in 82.1%, whole bankroll lost in 5.8%, average result −9.6 units 82% 6% +20 units Target +30 units: reached in 74.2%, whole bankroll lost in 7.2%, average result −13.5 units 74% 7% +30 units Target +50 units: reached in 48.2%, whole bankroll lost in 11.6%, average result −19.4 units 48% 12% +50 units Target +75 units: reached in 2.6%, whole bankroll lost in 11.9%, average result −21.4 units 3% 12% +75 units Win target (bankroll 200 units, at most 300 spins)
Fibonacci with a bankroll of 200 units and a table maximum of 100 units, at most 300 spins, one million sessions per target. Hover over a group for the numbers. Source: GamblingSociety simulation.

With +10 units, 91% of sessions reached the target and 4% lost the bankroll. With +20 it was 82% and 6%. A target of +75 units was almost out of reach: only 2.6% of sessions got there within 300 spins. The Fibonacci system wins slowly, so a big target mainly means more time at the table and more chances for a bad stretch.

All scenarios

ScenarioTarget reachedSession ended aheadWhole bankroll lostAverage result (units)Lost per 100 staked
Standard: bankroll 200, table max 100, target +5048%66%12%−19.42.7
Table max 5037%48%4%−18.12.7
La Partage (half back on zero)65%70%10%−9.01.3
Stop at a loss of 100 units47%63%0%−16.02.7
Target +20, stop at a loss of 100 units82%82%0%−7.52.7
Bankroll 50, target +1273%74%25%−4.02.7
Bankroll 100, target +2568%68%21%−9.52.7
Bankroll 500, target +1250%71%1%−24.72.7
Target +1091%91%4%−5.22.7
Target +2082%82%6%−9.62.7
Target +3074%74%7%−13.52.7
Target +753%63%12%−21.42.7
One million sessions per scenario, at most 300 spins. Everything not mentioned is as in the standard test. A stop at a loss of 100 units can end a little lower, because the last bet may take the loss past 100. Source: GamblingSociety simulation.

A few lines stand out. A stop-loss at half the bankroll meant nobody lost everything, while the target was reached almost as often (47%) and the average loss fell from 19.4 to 16.0 units. La Partage halves the price: 9.0 units lost per session on average. A table maximum of 50 cuts the lost bankrolls to 4%, but the target is reached less often (37%). With a bankroll of 500 units almost nobody goes broke (0.5%), but the target of +125 is out of reach in 300 spins.

5. Variants of the Fibonacci system

  • Fibonacci with a stop-loss. You stop the session when you are 100 units down. In our test that removed the lost bankrolls completely and made the average session cheaper (−16.0 instead of −19.4 units). Together with a target of +20, 82% of sessions reached the target. This is the most useful change you can make.
  • Reverse Fibonacci. You move one step up after a win and two steps back after a loss, so you bet more while you are winning. The idea is the same as with the Paroli: let wins run instead of chasing losses. We have not tested this variant separately. It changes the size of wins and losses, but the cost stays at 2.7% of every stake on red or black.
  • D’Alembert. An even slower progression: one unit more after a loss, one unit less after a win. In our standard test it reached the target in 67% of sessions but lost the bankroll in 28%. It staked the most of all ten systems we tested (1,050 units) and lost the most per session (−28.4 units). Our D’Alembert page has the full test.
  • Martingale. Doubling after every loss ends each series with a win of 1 unit, so it reaches small targets fast, but long losing runs cost more. Our Martingale page has the full test.

6. How to play the Fibonacci system sensibly

The Fibonacci system cannot beat the house edge either. If you treat it as entertainment with a fixed price, these rules from the numbers above help:

  • Know how far your ladder goes. Add up the sequence until you reach your bankroll. With 143 units you can make ten bets in a row, with 232 units eleven. If the table maximum stops you earlier, that is your real limit.
  • Keep the target small. With +20 units on a bankroll of 200, 82% of our sessions reached the target. With +50 it was less than half.
  • Plan for the time it takes. The Fibonacci system needed a median of 247 spins to win 50 units. At 30 to 60 spins per hour, that is four to eight hours at a real table.
  • Set a stop-loss. Stopping at half the bankroll removed the lost bankrolls in our test and cut the average loss.
  • Do not raise the base bet to win back a loss. After a bad stretch the next spin is as random as any other. Starting a new ladder with bigger stakes only makes the next bad stretch more expensive.
  • Play at a table with La Partage if you can. It halves the cost of every bet on red or black.

Frequently asked questions

Does the Fibonacci system work in roulette?

It wins most sessions with a small target, but not in the long run. Over many sessions it loses 2.7% of everything you stake, like every other roulette system. In our standard test it ended 66% of sessions ahead and lost 19.4 units per session on average.

Is the Fibonacci system safer than the Martingale?

It loses the whole bankroll less often (12% against 16% in our test), because the stakes grow more slowly. But it reaches the target less often, takes about two and a half times as long and loses more per session on average, because it puts more money on the table.

How much bankroll do I need for the Fibonacci?

For ten bets in a row you need 143 base bets, with $5 bets that is $715. Eleven bets need 232 base bets. More bankroll means fewer busted sessions but bigger losses when they happen.

What happens after a win in the Fibonacci?

You move two steps back in the sequence, for example from 8 to 3 units. At the first two steps you stay at 1 unit. A win only cancels the last two losses, so after a long losing run you need several wins to get back to a profit.

Which bet is best for the Fibonacci?

Red or black, odd or even, and 1 to 18 or 19 to 36 all have the same odds. The table rules matter more: with La Partage you get half your stake back on zero, which halves the house edge on these bets.

Does the Fibonacci system work online?

The odds are the same online and in a casino. Online tables play faster, which helps with the time the Fibonacci system needs, but it also means more bets per hour and therefore more of the 2.7% edge per hour. Check the table maximum before you start.

How we tested it

  • Simulator: our own program, one million sessions per scenario. We checked the standard test with a second, independent program (400,000 sessions): it found 48.2% of targets reached and 11.5% lost bankrolls, the same as the main simulation within its precision.
  • Rules: single-zero wheel, bets on red, all even-money bets lose on zero (variant: La Partage). One step up after a loss, two steps back after a win. A session ends at the target, when the money runs out, or after 300 spins. When the next stake is above the table maximum or above the money left, the player takes the loss and starts again at 1 unit.
  • Limits: one player, no mistakes, no side bets. Real tables play roughly 30 to 60 spins per hour, so 300 spins are a long evening.

Related guides

About the author: GamblingSociety is an independent site run by Andreas, who has operated casino and betting guides since the mid-2000s. No bookmaker or casino owns or pays for our content. More about us.