Havana: the system that rebuilds a single dozen.
An unconventional strategy that occupies all 12 numbers of a chosen dozen with low units and gradually swaps them for high ones as hits arrive. It looks like a clever internal progression — but long-run tests showed a markedly negative trend. Here you can take Havana apart step by step and test it on simulated spins.
Test Havana on both a short and a long sample
Set the bankroll, the low unit and the mode. The simulator models the basic principle of Havana in simplified form: one dozen, low and high units, and a reset of the round once every low bet has been removed.
Bankroll and spins
Havana mode
The bet goes on the first dozen (numbers 1–12); a high unit is 2× a low one. Every run uses a fresh random sequence of spins.
Bankroll over time
With Havana, watch how fast the losing runs add up — the system often feels controlled, but you are paying for a whole set of 12 numbers.
What happened in the simulation: Click "Run the simulation" and you will get a short, plain-language explanation of the result here.
Low, high and uncovered numbers within one dozen
Havana does not bet on the dozen as a whole. It occupies each of its 12 numbers separately and changes the state of each during a round — that internal rebuilding is the heart of the system.
The cost of one spin
At a $10 low unit you start with a stake of $120 per spin (12 × $10). As hits arrive, some numbers move to high units and others are uncovered — so the total stake changes during the round. This is what the table might look like mid-round:
A model round (a $10 unit)
The stake per spin hovers around 12 units: a hit on a low unit leaves it unchanged, a hit on an uncovered number raises it by one, and a hit on a high unit lowers it by one.
What does a run without a hit on the dozen cost?
This calculator is not about whether Havana closes a round. It is about how likely a run of spins entirely outside your dozen is — and what such a run costs you.
Enter a low unit and the length of the dry run. The result shows the probability that your dozen never lands, and the approximate cost of such a run in money.
Is Havana a good roulette system?
Interesting as a teaching example of an unconventional strategy, poor as a route to long-term profit. Here is the whole picture, the maths and the test results, in one place.
What is the Havana system in roulette?
Havana is a lesser-known roulette strategy built on occupying the numbers inside one chosen dozen. It does not use a simple dozen bet but individual straight-up bets on all 12 of its numbers. What makes it distinctive is its handling of low and high units: a low unit carries the base value, a high one is double. As play continues, the low units are gradually eliminated and some numbers move to high units.
At first glance the system looks sophisticated: it actively “rebuilds” the table, responds to hits and has a clearly defined goal for each round. But that activity is exactly what creates a false impression of control. Under the bonnet it is still a series of straight-up bets paying 35:1 on a wheel that remembers nothing.
The rules of the Havana system
You begin by choosing one dozen and placing a low unit on each of its numbers. So you start with 12 units on the table. If a number outside your dozen lands, the layout does not change and you carry on — you have simply lost the whole current stake.
If a number carrying a low unit lands, that low unit becomes a high one and at the same time you remove a low unit from another number. If a number with a high unit lands, the high unit stays and you again remove a low unit elsewhere. If an uncovered number from the dozen lands, you place a high unit on it and once more remove a low unit from another number.
The round ends the moment no low unit is left on the table. The system then resets and you start again with a low unit on every number of the chosen dozen.
What did the long-run tests of Havana show?
The original test first tried Havana without a progression. The result was negative, with an average loss of roughly 10,000 units per 25,000 spins. A linear progression of runs was then tested, raising the value of the low unit after losing runs. That adjustment did not help either — the test showed a still sharper decline, and in some samples the figures ran to very large losses.
The original analysis also examined 986 individual runs. The losing ones outweighed the winning ones both in number and in money, confirming the problem was not one unfavourable chart but the structure of the system itself. You can switch between both modes — basic and linear progression — in the simulator above and compare them over 25,000 spins.
The biggest risks of the Havana system
The first risk is the number of bets on the table. Although the system talks about low units, twelve bets are on the table from the start. That means every number outside your dozen costs the player the entire current stake — at a $10 unit, $120 on every unsuccessful spin.
The second risk is complexity. The player has to track which numbers are low, high and uncovered. At a real table that would be impractical and error-prone — one misplaced unit changes the whole course of a round.
The third risk is a false impression of control. Havana looks like a system that actively works with the internal structure of a dozen. In reality it still runs into roulette’s payout ratio over time: a straight-up bet pays 35:1, while the fair ratio on 37 numbers would be 36:1.
How to read the simulation chart
A Havana bankroll chart typically has a saw-toothed shape: long declining stretches while the dozen fails to land, broken by jumps upward after hits. Do not be misled by individual wins — the slope of the whole curve and the deepest drawdown are what matter. The toggle above the chart offers a second view: the stake per spin, which changes with how many low and high units remain on the table.
The best way to understand Havana is to run several simulations with the same settings. A short close-up run over 1,000 spins shows the mechanics of the rounds; a long test over 25,000 spins shows the trend. To compare it with other systems on identical spins, use the Roulette system tester.
The verdict: is the Havana system worth it?
Havana is interesting as a teaching example of an unconventional roulette strategy. It shows that even sophisticated shuffling of bets within one group of numbers need not create an edge — neither the probabilities of individual spins nor the payout ratio change.
As a long-term strategy, Havana in this form cannot be recommended. The original test called it unplayable without further reworking, because the losing runs clearly outweighed the winning ones. If you are tempted to experiment, do it in the simulator — there a long dry run costs you nothing but a click.
The biggest misconceptions about Havana
More complicated systems invite the impression that “more rules means more control”. With Havana that goes double.
Shifting low and high units only changes how the stake is distributed. The expected value of every straight-up bet is still −2.70% (EU).
A dozen hits on only ~32% of spins. The other ~68% cost the whole stake on the table — and they can line up into long runs.
Every spin is a separate event. After ten spins outside, the chance of a hit is still the same 12/37.
Raising the unit after losing runs only increased the volume of bets on identically negative mathematics. The results were worse than the basic version.
Havana vs. other roulette systems
Havana belongs to the less usual systems that work internally with a group of numbers. Here is how it compares with other unconventional strategies.
| System | Risk | Principle | Best for |
|---|---|---|---|
| Martingale | Very high | Doubling the bet after a loss | Understanding the risk of a progression |
| Fibonacci | High | A progression following the Fibonacci sequence | A slower progression after a loss |
| Labouchere | High | A cancellation line; the bet is first + last | More advanced players and experiments |
| Reverse Labouchere | High | The line grows on a win and is crossed off on a loss | Hunting streaks with a capped loss |
| Parlay | Medium | The win stays in play (let it ride) | Short runs after a win |
| Bally | Very high | 24 numbers with a 12–36–108–324 progression | Blanket bets with a harsh progression |
| Charlotte | High | A single-number progression over 88 spins | Hunting the 35:1 payout to a plan |
| Dominance | High | A double street following the dominant side of the table | Players who follow trends |
| Double 10 | High | One number, stepping up after 10 misses | Patient single-number hunters |
| FAST | High | 33 numbers, low and high units | Maximum table coverage |
| Havana | High | One dozen, two levels of units | Dozen players |
| Holy Grail | High | 18 distinct numbers with a progression | Covering half the wheel |
| John Wayne | High | A spread of 18 numbers for 10 units | Combined bets on numbers |
| Numeric Scanner | High | Repeating numbers with a progression | Players who love stats and tables |
| Seven Corners | Medium | Seven corner bets, 27–28 numbers flat | Blanket bets with no progression |
| System 300 | Very high | 34 of 37 numbers, a rare 300-unit loss | Understanding asymmetric risk |
| Five-number system | High | Five numbers, progression blocks of 1-6-11-16 | Hunting the 35:1 payout on a budget |
The name of a system takes you to its page with an interactive test — the current system is highlighted. Want to compare several systems on the exact same sequence of spins? Use the Roulette system tester.
Common questions about the Havana system
Is the Havana system profitable?
Not in the long run. The original test showed a markedly negative trend both without a progression and with a linear one — an average loss of roughly 10,000 units per 25,000 spins.
What is bet on in the Havana system?
You choose one dozen and bet straight up on its individual numbers — 12 separate bets on the table at once.
What do the low and high units mean?
A low unit is the base bet on a number. A high unit is double that in the original description, and arises after a number hits — either from a low unit or on an uncovered number.
What is the chance my dozen lands?
On European roulette 12 of 37, so roughly 32.43%. On American roulette 12 of 38, roughly 31.58%. The rest of the spins mean losing the entire stake on the table.
How much money do I need to start a round?
A round starts with 12 low units, so the stake per spin is 12× the low unit. At a $10 unit that is $120 every spin — and a dry run of ten spins costs over $1,000.
Does a linear progression of the low unit help?
According to the original tests, no. Raising the unit after losing runs led to a still sharper decline and, in some samples, to very large losses.
Does Havana change roulette’s house edge?
No. On European roulette every straight-up bet has an expected value of −2.70%, and on American −5.26%. Moving units between numbers changes nothing about that.
Why does a round end when all the low units are gone?
That is how the system is defined: the goal of a round is to eliminate every low unit through hits. Once none remain, the round closes, the result of the run is settled and the layout resets.
Can I play Havana on a different dozen or column?
Yes — mathematically it is exactly the same. Every dozen and every column covers 12 numbers, so neither the hit probability nor the expected value changes.
Is Havana suitable for beginners?
Not really. The system is harder to keep track of — you have to watch the state of all 12 numbers — and its long-term mathematics is no better than that of simpler roulette bets.
Related tools and systems
First understand the principle, then work out the risk — and test longer scenarios over thousands of spins.