Tiers du Cylindre (Serie 5/8) with Doubling: A Roulette System Played for Months
The Serie 5/8, called tiers du cylindre in French casinos, covers the twelve numbers on the side of the wheel opposite the zero. Played with doubling after every miss, it wins most series and wins more the longer a series runs. I played exactly this system for months at an online live casino in the early 2000s and finished more than $10,000 ahead. This page explains the system, tells that story and shows with our own simulation how likely such a run really is.
The short version: with $1 per split, 93% of all series end with a win, on average about $48. About one series in 13 loses every bet, which costs $634. Starting with $600 and aiming for +$10,000, about 2% of players get there. The other players lose the $600, most of them quickly. You can try it yourself in the simulator right below.
32.4%
Chance per spin: 12 of the 37 numbers
1 in 13
Series lose every bet (−$634 with $1 per split)
2.1%
Of players starting with $600 reach +$10,000
2.7%
Lost on average per dollar staked
1. Try the Serie 5/8 in the simulator
The simulator starts with the Serie 5/8 and doubling. The base bet is the total on all six splits together. Change the bankroll, the table limit or the target and compare the system with the Martingale on red or with any other system.
2. What the Serie 5/8 is
On a European wheel, the twelve numbers 27, 13, 36, 11, 30, 8, 23, 10, 5, 24, 16 and 33 sit next to each other, exactly opposite the zero. They fill about a third of the wheel, hence the French name tiers du cylindre, “third of the wheel”. In German casinos the same bet is called Serie 5/8, after its first split.
On the table these twelve numbers form six pairs that lie next to each other on the layout, so they can be covered with six split bets:
Split
Numbers
Pays
1
5 and 8
17:1
2
10 and 11
17:1
3
13 and 16
17:1
4
23 and 24
17:1
5
27 and 30
17:1
6
33 and 36
17:1
In land-based casinos you can usually call the bet (“tiers” or “Serie 5/8”) and the dealer places the six chips. Online you place them yourself or use the racetrack.
With one chip on each split you stake six chips. When one of the twelve numbers comes up, that split pays 17 to 1, and you get 18 chips back for the winning split while the other five are lost. That is 12 chips profit on 6 staked: the whole bet pays 2 to 1. The chance of a hit is 12 in 37, 32.4%. Like every bet on a single-zero wheel, it costs 2.7% of your stake on average.
3. Doubling on the Serie 5/8
The idea is the same as in the Martingale: after every miss you double the stake on all six splits, and after a hit you start again at the base bet. The difference is the payout. Because a hit pays 2 to 1, a late hit does not only win back the losses, it wins more than an early one:
Try
Per split
Total stake
Result of the series if this try hits
Chance
1
$1
$6
+$12
32.4%
2
$2
$12
+$18
21.9%
3
$4
$24
+$30
14.8%
4
$8
$48
+$54
10.0%
5
$16
$96
+$102
6.8%
6
$32
$192
+$198
4.6%
7
$64
$256 (capped)
+$518
2.1%
All miss
−$634
7.5%
Serie 5/8 with $1 per split and doubling after every miss. At the seventh try a table maximum of $256 per spin allows only four splits to be filled with $64; the other two stay empty, so only 8 of the 12 numbers win. Chance = probability that the series ends exactly at this try.
In total, 93% of all series end with a win, on average about $48. The other 7.5%, about one series in 13, lose all seven bets and $634. Compared with the Martingale on red, the misses come more often (a miss has a chance of 25 in 37, a loss on red 19 in 37), but every series that ends in a hit brings more.
4. How I played it
In the early 2000s I played a lot of roulette, in real casinos and at online live tables. At one live casino my system was the Serie 5/8 with doubling: $1 on each of the six splits, and after every miss I doubled. For a while I also tried a version that bet the base twice and only then started doubling, but most of the time I doubled straight away: 1, 2, 4, 8, 16, 32 and 64 dollars per split.
The table limit allowed six doublings. At the last step two or three of the splits could no longer be filled completely, because the total would have been above the maximum.
I had one rule: the longer the Serie 5/8 had not come up, the better the moment to start a series. To be honest, I often did not stick to it.
My goal was never the big win. I wanted small profits over a long time. There were negative phases, and at times I lost a few hundred dollars, but by then I was already well ahead. Before this run I had deposited, won a bit, lost a bit and experimented. After I had lost about $1,000 of deposits and winnings, I deposited another $500 and started again with about $600.
For two to three months I played almost every day. Then the casino lowered its table limits, and I could only double four times. With that limit one bad series would have finished me, so I cashed out. In the end I was more than $10,000 ahead at that casino.
I know that roulette loses in the long run. In my case the short-term streaks and luck were on my side, and I stopped at the right time.
Andreas, GamblingSociety
5. How likely is a run like this?
We simulated exactly this situation 50,000 times: $600 to start, $1 per split, doubling up to $64 per split with the last step capped at $256, and play until the bankroll reaches $10,600 or the money runs out. When the next step was not affordable any more, the player took the loss and started again at the base bet.
300 runs, spin by spin
Starting with $600 and aiming for +$10,000
Reached $10,600 (5 of 300)Went broke (295 of 300)
300 simulated runs, not selected. Blue: reached $10,600. Orange: went broke. Hover over a line for its result. Source: GamblingSociety simulation.
In all 50,000 runs, 2.1% reached +$10,000, about one player in 47. Half of the others were broke within the first 91 spins. $600 does not even cover one complete losing series ($634), so a run like mine survives only if the start goes well. The successful runs needed about 2,400 spins (median), which fits two to three months of almost daily play.
Smaller goals, better chances
Chance to reach a profit goal before the $600 are gone
Same simulation, 50,000 runs. The player stops as soon as the goal is reached. Hover over a bar for the numbers. Source: GamblingSociety simulation.
Even the smallest goal is far from certain: only 46% of players double the $600 before losing it, because one bad series at the start is enough. The most telling number is a different one. About 6.5% of all players were +$5,000 ahead at some point, but 67% of them kept playing and lost everything. The profit came from a good run; it stayed with the players who stopped.
What the lower table limit changed
With only four doublings ($1 to $16 per split), a series loses every bet much more often, in 14% of cases instead of 7.5%. Each of these losses costs $186 instead of $634, and the wins are smaller too. The price per dollar staked stays at 2.7% either way. What changes is the shape: smaller, more frequent full losses. For a player with a big profit on the table, cashing out was the decision that made sense.
6. Does it help to wait until the series is “due”?
My own rule was to start after the Serie 5/8 had not come up for a while. We tested it on almost 400,000 real spins from one roulette table in a German casino (2023 to 2026):
Serie 5/8 missed in a row before
Next spin hit the Serie 5/8
none (it had just come up)
32.4%
3
32.7%
6
32.1%
9
32.8%
12
33.5%
Real spins, each casino day counted separately. A fair wheel gives 32.4% every time. Source: GamblingSociety analysis.
The chance stays at about 32.4%, no matter how long the series has been missing. The longest gap we found in one day was 29 spins in a row without any of the twelve numbers. Waiting does not change the odds, so not sticking to my rule did not cost me anything either.
7. Serie 5/8 compared with the Martingale on red
In our standard test for all systems (bankroll 200 units, table maximum 100 units, target +50 units, at most 300 spins) the base bet is 1 unit on all six splits together. The Serie 5/8 reached the target in 73% of sessions and lost the bankroll in 27%. The Martingale on red reached the target in 62% and lost the bankroll in 16%.
The Serie 5/8 is faster and more volatile. Because every hit brings at least 2 units, its sessions lasted 47 spins on average, those of the Martingale 155. It also staked only about half as much per session (272 units instead of 538), and so it lost less per session on average (7.3 units instead of 14.6). In exchange, more sessions ended with the whole bankroll gone. Our overview of all ten systems shows how this compares with the others.
8. How to play the Serie 5/8 sensibly
Bring enough for a full series. With $1 per split, seven bets cost $634. With less money, one bad series at the start ends the evening.
Check both limits. Online tables often have a maximum per split and one for the whole table. The lower one decides how many doublings you really have.
Set a goal and a cash-out rule. For example: after every +$500, withdraw half of it. What is withdrawn cannot be lost in the next bad series.
Stop when the conditions change. Lower limits, a smaller bankroll or a long session all raise the risk of a full loss.
Do not chase. After a lost series, the next one starts at the base bet, never with a bigger stake to win the loss back.
See it as entertainment with a price. Over time every series costs 2.7% of what you stake. The good runs are real, and so is the price.
Frequently asked questions
Which numbers are in the Serie 5/8?
27, 13, 36, 11, 30, 8, 23, 10, 5, 24, 16 and 33. On the layout they form the splits 5/8, 10/11, 13/16, 23/24, 27/30 and 33/36.
What does the tiers du cylindre pay?
Each split pays 17 to 1. With equal chips on all six splits, a hit returns three times the total stake, a net profit of 2 to 1 on the whole bet.
Is the tiers du cylindre a good bet?
It has the same house edge as every other bet on a single-zero wheel, 2.7%. La Partage does not apply to it, because it only covers even-money bets.
How much bankroll do I need for the Serie 5/8 with doubling?
With $1 per split, a full series of seven bets costs $634 if the last step is capped at $256 by the table limit, or up to $762 without a cap. With less money you cannot finish a bad series.
Does waiting for a missing Serie 5/8 help?
No. In almost 400,000 real spins, the twelve numbers came up 32.4% of the time after no miss and about as often after three, six, nine or twelve misses in a row.
What is the difference between tiers, voisins and orphelins?
They are the three classic call bets of French roulette and together cover the whole wheel. Voisins du zéro covers the 17 numbers around the zero, tiers du cylindre the 12 numbers opposite, and orphelins the remaining 8 numbers.
How we tested it
Simulator: our own program. 50,000 runs from $600 for the run in section 5, one million sessions per scenario for the comparison in section 7. The results for one series match the exact calculation.
Rules: single-zero wheel, six equal split bets on the Serie 5/8, doubling after every miss up to $64 per split, a table maximum of $256 per spin for the last step. When the next step is not affordable, the player takes the loss and starts again at $1 per split.
Real data: almost 400,000 spins from one roulette table in a German casino, January 2023 to September 2026, from the casino’s published records.
The story: told from memory, more than twenty years later. The exact limits and the number of spins are not documented any more; the simulation uses the setup as it is remembered.
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.