The Paroli roulette strategy is a positive progression: stakes increase after wins and reset after a loss. A common version doubles twice, aiming to complete three consecutive wins before returning to the base unit. It is sometimes called reverse Martingale because the stake moves in the opposite direction.
Basic three-step rules
- Bet one unit on an even-money outcome.
- After a win, bet two units on the same outcome.
- After a second win, bet four units.
- After a third win, bank the cycle result and reset to one.
- After any loss, reset immediately to one.
A completed cycle wins 1 + 2 + 4 = 7 units. If the first bet loses, the cycle loses one. If the second loses, the first win and second loss create −1 net. If the third loses, the cycle produces 1 + 2 − 4 = −1. With this structure, every failed cycle loses one base unit.
Why the progression feels controlled
Unlike Martingale, Paroli does not respond to a loss with a larger bet. The largest stake is placed only after earlier wins have occurred. That can make it psychologically easier to stop, and a strict three-win cap prevents unlimited escalation.
Nevertheless, chips already won are still money. Calling a four-unit stake “house money” does not remove its value or change the probability of the next spin.
Probability of completing a cycle
On European roulette, a red bet wins with probability 18/37. Three specified red wins in a row occur with probability (18/37)^3, approximately 11.51%. American roulette reduces the probability to (18/38)^3, approximately 10.63%.
Most cycles therefore fail before the three-win target. The seven-unit reward compensates for many one-unit failures, but zero keeps the long-run expectation negative.
Cycle expectation
Stake sizing does not need a complicated proof. Every unit placed on standard European red carries an expected loss of 1/37. Paroli determines how many units enter action before a reset. Add the expectation across those conditional wagers and it remains negative.
Example session
Consider results W, W, L, L, W, W, W. The first cycle stakes 1, 2 and 4, then loses the four-unit decision for a cycle result of −1. The second cycle loses one. The third completes all three levels for +7. Overall session result is +5 units.
Another sequence with the same number of wins can differ because their grouping determines whether larger bets occur. Positive progressions depend on clustered wins.
Common variations
- Two-step Paroli: stakes 1 and 2, then resets after two wins.
- Four-step Paroli: continues to 8 units, increasing both target and give-back risk.
- Partial lock: removes part of the previous profit rather than doubling the entire stake.
- Fixed target: ends the session after one completed sequence.
Longer ladders create more attractive headline wins and lower completion probability. Any comparison should measure the entire cycle distribution, not only the target.
Paroli versus Martingale
| Feature | Paroli | Martingale |
|---|---|---|
| Increase after | Win | Loss |
| Typical cap | 2–4 wins | Bankroll/table limit |
| Failed cycle | Often one unit | Can equal full progression |
| Needs | Winning streak | A win before limit |
| Changes house edge | No | No |
Choosing an underlying bet
Red/black, odd/even and low/high have the same probabilities under standard rules. Switching sides after each result does not improve them. A French game with La Partage or En Prison can reduce the edge on qualifying wagers, which is a real rules improvement independent of Paroli.
How to test Paroli
Set the number of progression steps, reset rules, wheel, unit and session length. Run the same random sequences through Paroli and flat betting. Record total action, net result, maximum stake, completed cycles and drawdown.
A fair graph will show many one-unit cycle losses punctuated by completed sequences. Do not display only sessions that reach the target.
Practical safeguards
Cap the ladder before play and do not extend it because a streak “feels strong.” Treat accumulated wins as part of the bankroll. Use a time and loss boundary independent of the progression. A completed cycle is not evidence that another is beginning.
Verdict
Paroli limits the aggressive loss-chasing associated with negative progressions, but it does not create an advantage. Its most defensible use is as a finite demonstration of how outcome order and conditional stake size shape a session. Compare its exposure with other systems using the bankroll calculator.
Paroli frequently asked questions
Should the bet remain on the same colour?
The probability is equal for either colour on the same wheel. Staying consistent simplifies a test. Switching after a win or following a visual pattern adds a separate selection rule that needs its own baseline.
Why stop after three wins?
Three is a convention, not a mathematical optimum. It caps the largest common stake at four units and creates a seven-unit completed-cycle profit. Continuing to an eight-unit fourth bet greatly increases what can be returned after a rare streak.
Are previous wins really financing the next bet?
Accounting may describe them that way, but all chips in the bankroll have equal economic value. A loss on level three reduces funds by four units regardless of their history.
Does Paroli work better on dozens?
A dozen pays 2:1 but wins on only 12 of 37 European pockets. It needs a redesigned ladder and cycle calculation. The standard 1–2–4 even-money illustration cannot simply be transferred.
Cycle-based analysis
Define a cycle from the one-unit opening decision until either a loss or the chosen number of wins. Report the number of each failure type and completed ladders. This makes the attractive target visible beside the much more frequent failed attempts.
Streak-selection bias
Historical charts can select periods containing obvious winning clusters. A valid test starts every cycle according to a fixed rule and includes quiet, alternating and losing periods. Do not begin only after the first win unless that entry condition is recorded for the baseline too.
Comparing caps
Test two-, three- and four-win ladders on identical outcomes. Short caps complete more often but earn less per success. Long caps complete less often and expose more accumulated winnings. Expected value remains driven by the total negative-expectation action.
Software implementation checks
An automated Paroli tool must reset after zero, handle La Partage explicitly, prevent an unintended fourth level and retain interrupted cycles in results. Confirm whether displayed profit includes the original stake returned.
Responsible conclusion
A finite positive progression is generally easier to bound than loss doubling, but it can still encourage repeated attempts after failed cycles. Set an overall session loss and decision count; a one-unit cycle loss repeated many times is still a material loss.
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.