The Fibonacci roulette strategy sizes bets with the sequence 1, 1, 2, 3, 5, 8, 13, 21 and so on. Each number after the first two is the sum of the preceding pair. Roulette versions usually advance one step after a loss and move back two steps after a win.
Standard playing rules
- Choose a base betting unit and an even-money wager.
- Start at the first 1 in the Fibonacci sequence.
- Move one position forward after a loss.
- Move two positions back after a win, never below the start.
- Define a maximum sequence position and session limit in advance.
Published versions differ. Some repeat the first unit, some reset after any net recovery and others apply the sequence to dozens. A test is meaningful only if the precise movement rules are documented.
Worked example
Assume outcomes L, L, L, W, L, W, W on red. Stakes are 1, 1, 2 and 3 for the first win. Moving back two positions returns the next stake to 1. After that loss the sequence advances to 2; a win moves back to 1, and the following win remains at the start.
Financial result: −1 −1 −2 +3 −1 +2 +1 = +1 unit. A different ordering with the same number of wins and losses can produce a different result because stake size depends on path.
Why moving back two can recover
Because each Fibonacci number is approximately the sum of the two previous stakes, a win later in the sequence can offset recent smaller losses. The relationship is not a guarantee of full recovery at every point, especially when zero, resets, maximum positions and interrupted sessions are included.
Progression cost
| Position | Stake | Cumulative stakes |
|---|---|---|
| 1 | 1 | 1 |
| 2 | 1 | 2 |
| 3 | 2 | 4 |
| 4 | 3 | 7 |
| 5 | 5 | 12 |
| 6 | 8 | 20 |
| 7 | 13 | 33 |
| 8 | 21 | 54 |
| 9 | 34 | 88 |
| 10 | 55 | 143 |
The sequence grows more slowly than Martingale but still accelerates. Ten uninterrupted losing stakes consume 143 units and require 89 next. A one-unit table minimum can therefore reach meaningful exposure without any dramatic early jump.
Fibonacci versus Martingale
Martingale doubles: 1, 2, 4, 8, 16. Fibonacci progresses 1, 1, 2, 3, 5. Fibonacci generally survives more losses for the same bankroll, but it may require several wins to retreat and recover. Martingale aims for one-unit recovery on the first subsequent win, at the cost of faster exponential growth.
Neither sequence changes the 18/37 probability of winning red on European roulette. They offer different distributions of stake and drawdown.
The effect of zero
Zero is a loss for standard even-money bets. Treating it as a “no decision” in a spreadsheet overstates performance unless the actual table uses En Prison or another documented rule. American double zero worsens the probability further.
Sequence limits are essential
Without a maximum position, Fibonacci stakes can exceed any finite bankroll. With a maximum, the system accepts a loss that may not be recovered under its usual movement rules. That is not a flaw in testing; it is the realistic consequence of finite money and table limits.
Common testing errors
- Changing from red to black after a streak.
- Skipping zero in historical data.
- Resetting only when doing so improves the chart.
- Reporting win rate without total stake.
- Comparing different session lengths.
- Ending a test at the first recovery.
A reproducible test
Fix a European wheel, one-unit base, 143-unit maximum progression budget and a ten-position ceiling. Generate at least 10,000 sessions of equal length. Record average and median ending bankroll, maximum stake, drawdown, profitable-session percentage and ceiling failures. Repeat with one-unit flat betting on red.
The comparison demonstrates what Fibonacci changes: it will not improve expected value, but its path and risk concentration will differ.
Who should avoid it?
Anyone tempted to chase losses, treat the sequence as income or breach a preset limit should not use a progression with real money. The mathematical neatness of Fibonacci can make rising stakes feel rational even when the decision is still a negative-expectation wager.
Bottom line
Fibonacci is useful as a study of path-dependent staking. It grows more slowly than Martingale, yet large stakes and terminal drawdowns remain possible. Use the roulette bankroll calculator to total the sequence and compare it with flat betting before drawing conclusions from a winning example.
Fibonacci frequently asked questions
Which of the first two ones should I start with?
Either can be labelled position one if the full rule remains consistent, but published examples may count positions differently. Write the actual sequence of stakes to prevent confusion.
What happens after a win near the beginning?
Most versions return to the first position and stay there if moving back two would pass the start. Some reset the whole cycle. This small rule affects simulations and should be fixed in advance.
Can Fibonacci be used on dozens?
It can size any wager, but a 2:1 payout requires different recovery logic from an even-money bet. Copying movement rules without accounting for payout can over- or under-recover. Zero still loses the dozen.
Is Fibonacci safer than Martingale?
Its stakes grow more slowly, so a fixed bankroll usually permits more progression steps. “Safer” must still specify risk of ruin, maximum drawdown, session length and unit. A longer ladder can encourage more total action.
Sequence math beyond the first ten steps
Fibonacci growth is not linear. Ratios between successive terms approach approximately 1.618. It is slower than doubling but eventually produces very large stakes: after 55 and 89 come 144, 233 and 377. A table maximum remains inevitable for continued losses.
Testing partial recovery
Record the bankroll after every decision rather than marking a cycle recovered merely because the sequence returned to its first position. Depending on the exact outcome path and rule variation, position and financial recovery need not be identical.
Compare at least three caps—for example six, eight and ten positions. A deeper cap reduces the frequency of a forced stop in a small sample but raises maximum exposure. There is no free extension.
Presentation traps
A chart starting just before a recovery can make the sequence look stable. Request the full test period and a fixed vertical scale. If results are quoted in “points,” confirm the monetary value and whether unit size changed.
Responsible conclusion
The familiar sequence gives the system intellectual appeal, but the origin of Fibonacci numbers does not confer predictive power. Use the progression as a well-defined object for simulation. Keep financial decisions anchored to affordable loss and the roulette payout table.
How to use this research
Treat the worked examples as explanations, not forecasts. Reproduce the assumptions with a free tool, change one variable at a time and keep losing as well as winning trials. A short favourable run cannot validate a strategy, while a short unfavourable run cannot measure its full distribution. For any real-money decision, verify local law and operator rules, set an affordable limit before play and never use a progression to recover money that has already been lost.