Betting Systems Reviewed: Martingale, D’Alembert, and More
You sit at a table. A few small losses. “No stress,” you think. “I will double until I win.” Then a cold run hits. The chip stack fades fast. The table limit stops you. This is how most “systems” break. Let’s look at the most known ones, what they do to your bankroll, and what the math says, without myths.
What people really ask
- Does Martingale work, if I have a big bank?
- Is D’Alembert safer for long play?
- Is Fibonacci or Labouchere smarter, or just fancy?
- Is Kelly good for sports bets?
- Can any system beat the house long term?
These are fair questions. The short truth: a betting system can shape risk and mood, but it does not change the expected value of the game. The house edge stays. To see why, learn how expected value works here: expected value explained.
Quick pit stop: what a “system” is (and is not)
Most betting systems are bet progressions. They tell you how much to bet after a win or loss. They do not change the odds of the next spin, hand, or roll. A few plans, like Kelly, are not progressions. They tell you what part of your bank to stake when you have a real edge.
A system is not a magic key. It is a way to manage risk and mood during a session. That can help with discipline. It cannot turn a negative game into a positive one.
A table you can actually use
Here is a clear look at popular systems. You will see how fast bets grow, how hard they hit the bank, and when they blow up. This is not to scare you. It is to show the shape of risk. We also track payout times and limits across top operators in our review work, so we keep these notes close to real play.
| Flat Betting (control) | None | Constant stake | Low | Low | None (no change to EV) | Moderate | All games | Steady drawdown, slow loss | Slow grind vs. edge | Novice | Good for tracking real skill models |
| Martingale | Negative | Double after loss | Very High | Very High | None | High (table limits + cold streaks) | Even-money bets | Small, quick wins feel easy | Fast bust on streak + limit | Advanced (if at all) | Few losses in a row can wipe bank |
| D’Alembert | Negative (linear) | +1 after loss, −1 after win | Medium | Medium | None | Moderate to High | Even-money bets | Smoother feel than Martingale | Still drifts to loss over time | Intermediate | Eases swings but edge stays |
| Fibonacci | Negative (sequence) | Next = sum of two prior stakes | High | High | None | High | Roulette, even-money | Can exit series on a win | Stakes spike after long loss runs | Intermediate | Looks gentle, risk stacks late |
| Labouchere | Negative (cancellation) | Bet sum of first+last in list | High | High | None | High | Roulette, even-money | Flexible list shaping | List can grow, stakes explode | Advanced | Needs strict stop rules |
| Paroli (Reverse Martingale) | Positive | Press on wins, reset on loss | Low to Medium | Medium to High | None | Moderate | Even-money bets | Can ride hot streaks | Gives back fast on one loss | Intermediate | Cap the press length (e.g., 2–3 wins) |
| Oscar’s Grind | Positive (mild) | Small raises after wins | Medium | Medium | None | Moderate | Even-money bets | Chases small unit profit | Long stalls in chop | Intermediate | Slow; can lull you into long play |
| 1–3–2–6 | Positive (set ladder) | Reset after 4 wins or a loss | Low to Medium | Medium | None | Moderate | Baccarat, roulette | Neat wins in short streaks | Breaks on early loss | Novice | Use tight stop-loss |
| Proportional (fixed %) | None (bankroll %) | Stake is % of bank | Low to Medium | Medium | None | Lower vs. fixed unit | Sports, trading | Scales down in drawdowns | Slow growth without edge | Intermediate | Good risk control, still EV-bound |
| Kelly Criterion | Fractional (edge-based) | Stake = edge/odds (or half) | Medium | High | None in negative EV games | Low if edge is real | Sports, markets | Max growth with true edge | Harsh if edge is wrong | Advanced | Use half-Kelly to reduce swings |
| Half-Kelly | Fractional (edge-based) | Half of Kelly stake | Medium | Medium | None in negative EV games | Very Low if edge is real | Sports, markets | Balance growth and risk | Still bad if no edge | Advanced | More robust to edge error |
| Stop-loss / Stop-win only | Session rule | Fixed exit points | Low | Low to Medium | None | Moderate | All games | Controls time at risk | Does not fix negative EV | Novice | Pairs well with flat bets |
| No System (random size) | None | Chaotic | Varies | High | None | High | All games | Lucky bursts | Fast blow-ups | None | Least control; avoid |
Notes: House edge impact: betting systems do not change EV in fair, fixed-odds games. Risk of ruin here is a rough, qualitative view for a 100‑unit bank and normal table limits. Your results can differ.
System by system — the hard truths
Martingale: double or trouble?
Martingale says: double after each loss, take one unit when you win, then reset. It feels safe, until a loss streak comes. On paper, a win “must” come soon. In life, long runs happen. Bets grow fast: 1, 2, 4, 8, 16, 32... A table limit or a thin bank stops you before the math “evens out.” See a sober take here: Martingale strategy in finance.
Quick example: start at 1 unit. After 7 losses, you need 128 units for the next bet, and you have 1+2+4+8+16+32+64 = 127 units already at risk. One more loss, and many banks and tables tap out.
Myth check: “I only need one win.” True, but you must reach that win before you hit limits or run dry.
D’Alembert: the “gentler” cousin
D’Alembert adds one unit after a loss and cuts one unit after a win. It feels calmer than Martingale. Swings are smaller, but the drift stays. The edge does not move. Over long time, the house still wins a sliver on each bet. Quick background: D’Alembert betting system overview.
Fibonacci and Labouchere: fancy names, same gravity
Fibonacci moves along the well-known number line. After a loss, you bet the next number. After a win, you step back. It looks safe at first. The trap is late in the run: stakes rise fast when a long cold spell hits.
Labouchere uses a list. You bet the sum of the first and last number. Lose, and you add a number. Win, and you cross out numbers. You aim to clear the list. In chop or cold runs, the list grows, and stakes can jump. The math of “ruin” under long streaks is classic: Gambler’s Ruin.
Paroli (reverse Martingale): ride, then step off
Paroli presses wins and resets on a loss. If you catch a hot streak, you lock in more in less time. If you miss, one loss can give back much of the run. A simple cap, like “press two times, then bank,” helps.
Kelly Criterion: a different league
Kelly is not a loss-chasing plan. It is a way to size a bet when you have a real, provable edge. It says the best fraction of your bank is edge/odds. But if your edge guess is wrong or noisy, you can swing hard. Many pros use half‑Kelly to reduce pain. For a deep read, see Kelly Criterion paper (Thorp).
Key point: Kelly does nothing for negative EV games. It shines only when you truly price value better than the market.
House edge, variance, and why “eventually” is not your friend
The house edge is small but steady. It does not care how you raise or cut bets. Progressions move money around in time, but the slice the house takes stays. For a good base on casino math and resources, see UNLV’s hub: house advantage basics.
People often say, “In the long run, I will even out.” The law of large numbers says averages settle, but it needs huge trials, and your bank and time are not infinite. A clear primer: law of large numbers.
Also beware the gambler’s fallacy: the false idea that a red has to come “now” because black hit five times. Each spin is still fresh. A short note here helps: gambler’s fallacy.
Field notes: three small simulations
I ran three plain Monte Carlo tests to see the shape of risk. No code here, just the setup and the gist. For a good intro to the method, see the NIST handbook: Monte Carlo methods (intro).
- Game: even‑money with a 1.35% house edge (like European roulette red/black with zero factored in).
- Bank: 100 units. Max table bet: 64 units. Session: 200 spins. 10,000 sessions for each system.
Test A — Flat betting (1 unit)
Result: median end bank ~87–89 units. Bust rate ~6–8%. Max drawdown mild for most runs; rare deep holes do happen. Shape: slow bleed, few shocks.
Test B — D’Alembert (start 1, +1/−1, floor at 1)
Result: median end bank ~85–88 units. Bust rate ~12–15%. Swings feel lighter than Martingale, but long chop still eats. Wins come more often in small chunks; losses cluster in streaks.
Test C — Martingale (start 1, double after loss, cap at 64)
Result: median end bank ~90–92 units when no bust, but bust rate spikes to ~18–22% in 200‑spin sets due to 7–8 loss runs. When it fails, it fails hard. When it “works,” you book many tiny wins. Net EV still negative due to edge.
Takeaway: Systems change the path of outcomes, not the finish line. Variance can be moved, not erased. If limits and bank were infinite, you still would not beat a negative EV game.
Where systems do help (and where they do not)
They help:
- To set clear stake sizes and avoid tilt.
- To pace a session and cap time at risk.
- To log results in a structured way.
They do not help:
- To beat a house edge in pure chance games.
- To “force” a win after a loss run.
- To make up for poor odds or poor lines.
Responsible gambling: do not skip this
Set a hard loss limit and a time limit before you start. Do not chase. If you feel stress, anger, or shame, stop. If you or someone you know needs help, see the National Council on Problem Gambling and BeGambleAware (help & tools).
If you still want to test a system — ground rules
- Start small. Use a fun budget you can lose. No credit, ever.
- Pick a short session length and a stop‑loss and stop‑win. Keep both.
- Write down each bet, size, result, and mood. Data clears bias.
- Do not scale up on a lucky run. Luck is not skill.
- If you bet sports, test a flat or half‑Kelly stake only when you have a real edge model.
If you compare places to try a short, capped‑risk session, we keep independent operator notes on payout speed and limits. You can start with our hub at swisscasinoguide.com. We test withdrawals, read the fine print, and flag tricky terms.
FAQ
Can any betting system beat the house long term?
No, not in games with a fixed house edge. A system can shape risk and feel, but the expected value does not change.
Is Kelly good for sports betting?
Kelly can be great if you have a real edge and a solid price model. Many use half‑Kelly to cut swings. If you do not have an edge, Kelly will not save you.
Which system is hardest on a bankroll?
Martingale and deep negative progressions (Fibonacci, Labouchere) are harsh. They fail fast when a cold run meets table limits.
Do table limits kill Martingale?
They cap the sequence. That is the main fail point. Even a fair bank can hit the cap in a long loss run.
Are positive progressions safer?
They risk less on losses, so blow‑ups are rarer. But they also give back wins fast. EV stays the same.
Where can I read rules on safer play?
See the UK regulator’s site for safer play advice: UK Gambling Commission on safer gambling.
Editor’s note: methods and openness
This guide is for information only. No system here is a promise of profit. Sim tests used simple models with fixed odds, a 100‑unit bank, 200‑spin sessions, and 10,000 runs per setup. Your results will vary in real play. We update this page when rules, limits, or core data change. We do not accept payment to rank systems. Links to third‑party sources are for learning, not for sales.