Fibonacci — Climb the sequence after losses; step back two places after a win.

IntermediateRisk: HighNegative progressions

Strategy guide

Fibonacci Strategy

Climb the sequence after losses; step back two places after a win.

Best for: Even-money players who want slower stake growth than Martingale but more structure than flat unit steps

Overview

Fibonacci is a negative-progression system for even-money bets (red/black, odd/even, high/low). Stake sizes follow the classic sequence 1, 1, 2, 3, 5, 8, 13, 21… measured in your chosen unit.

After a loss you move one term to the right. After a win you move two terms toward the start (standard educational rule). When you are already on the first term, a win simply keeps you there.

The climb is slower than Martingale’s pure doubling, so unfinished losing runs usually hurt less per spin—but deep sequences still produce large bets, and table maximums can freeze recovery mid-way.

No progression changes the math of the wheel. On European single-zero even-money bets the house edge is about 2.70% (roughly 1.35% only when French La Partage or En Prison applies). American double-zero even-money is about 5.26%.

Treat Fibonacci as a way to organize stake changes for short sessions, not as a method that beats the edge over time.

Visual progression

Fibonacci stakes (units along the sequence)

How it works

Choose a base unit small enough that many sequence steps still fit under the table maximum and your bankroll stop (for example $2–$5 if the max is a few hundred).
Write the unit sequence you will actually use: 1-1-2-3-5-8-13-21-34… and stop the list where your money or the table would break.
Start on the first 1. Place that many units on one even-money outcome.
If the spin loses, advance one position and bet the new term on the next spin.
If the spin wins, move two positions toward the start (or stay on the first term if you cannot move back two full steps).
When you return to the opening term after a completed recovery path, treat the cycle as reset and decide whether to continue or stop.
Optional: cap the highest allowed term (for example never past 13 units) so a long cold streak ends the system on purpose instead of improvising.
Always prefer European or French single-zero layouts when you have a choice; the edge is lower, but the system still cannot invert it.

Worked path (unit $5, European red)

Sequence in units: 1, 1, 2, 3, 5, 8… Start at the first 1.
Spin 1: bet $5 — lose. Running net −$5. Advance one step (still 1 unit).
Spin 2: bet $5 — lose. Running net −$10. Advance to 2 units.
Spin 3: bet $10 — lose. Running net −$20. Advance to 3 units.
Spin 4: bet $15 — lose. Running net −$35. Advance to 5 units.
Spin 5: bet $25 — win $25. Running net −$10. Move two steps back to the 2-unit term.
Spin 6: bet $10 — win $10. Running net $0. Move two steps back to the first 1. Cycle reset; the next bet would be $5 again.

Pros

Stake growth is more gradual than classic Martingale doubling
A fixed written sequence reduces mid-session improvisation
Even-money focus keeps payouts simple at 1:1
Easier to pre-map “how many losses until I stop” than pure ad-hoc chasing
Common educational baseline for comparing other negative progressions

Cons

Long losing runs still push stakes into uncomfortable territory
A win does not always erase the full dollar loss of the prior chain
Table maximums or a depleted bankroll can stop the sequence before recovery
Win-rule variants (one step back vs two) produce different results if mixed carelessly
Does not overcome house edge or zero/double-zero outcomes

Bankroll & table limits

Size the unit so a realistic losing streak (often planned to 6–8 losses) still leaves cash and headroom under the table max—many players keep the deepest planned bet under about 2–5% of total session funds only if the early terms are tiny.
Pre-compute dollar stakes for each term (5, 5, 10, 15, 25, 40…) and stop when the next term would break your written loss limit.
Fibonacci needs a larger cushion than flat betting for the same base unit because you are still raising after losses.
If the table max sits inside your sequence, treat that as a hard system failure point, not a reason to invent a new rule on the fly.

Full bankroll guide →

Common mistakes

Using a unit that is already large relative to bankroll so three or four losses force a stop far from any recovery
Moving only one step back after a win (or skipping steps) and then blaming the “system” for inconsistent results
Switching bet types mid-sequence (for example red to dozens) and breaking the 1:1 recovery assumptions
Continuing past a pre-set loss limit because “the next Fibonacci win will fix it”

FAQ

Does Fibonacci beat roulette in the long run?

No. It only rearranges how much you bet after wins and losses. Over enough spins the house edge on even-money bets still dominates—about 2.70% on standard European wheels and about 5.26% on American double-zero wheels.

Why step back two places after a win?

That is the usual teaching rule: one win does not fully cancel the faster growth of earlier losses, so retreating two terms aims to ease back toward the base. Some players use one-step retreats; pick one rule and stick to it.

Is Fibonacci safer than Martingale?

It usually climbs more slowly, so average stake after a short losing run is lower than pure doubling. Safer path is not the same as positive expectation—deep sequences still risk large drawdowns.

Can I use Fibonacci on dozens or columns?

People experiment, but classic Fibonacci assumes 1:1 even-money payouts. On 2:1 bets the recovery geometry changes and the textbook sequence no longer maps cleanly.

How long should the sequence be?

Only as deep as your bankroll and table max allow. Pre-compute the dollar amount at each step and stop before the next term would violate either limit.