The risk of a martingale strategy is that it ties your position size to your losses, so the worst stretch of your trading is also when you risk the most. Doubling after each loss makes every small winning streak look brilliant, but it leaves you one long losing streak, one margin call or one broker limit away from losing the account. Below we run the numbers, compare martingale with anti-martingale sizing, and cover the controls that actually protect capital.
TL;DR
- Martingale sizing doubles the next position after a loss. Exposure grows exponentially while your account stays finite, so a normal losing streak can cause a catastrophic drawdown.
- Every winner only recovers prior losses plus one base unit. Every loser deepens a hole that needs a much larger percentage gain to climb out of.
- Anti-martingale sizing does the opposite: it adds after wins and cuts after losses. It caps damage better, but it has its own weakness in choppy markets.
- Leverage and margin make it worse: U.S. regulators warn that traders can lose more than they put in.
- A fixed-risk plan, a hard stop for losing streaks and honest testing beat any "recover the loss" sizing scheme.
Table of Contents
- What a martingale strategy does in trading
- The math of a losing streak
- Why leverage and margin make martingale risk worse
- Costs and slippage eat the "small win"
- The sunk cost trap behind doubling down
- Anti-martingale: the mirror image
- Is there anything better than martingale?
- Our perspective on martingale sizing
- Build discipline instead of chasing losses
- FAQ
- Sources
What a martingale strategy does in trading
The martingale comes from gambling. You bet a base amount, and if you lose, you double the next bet. When you finally win, the payout covers every previous loss plus a profit equal to your first bet. In trading, the same logic shows up as doubling position size, adding to a losing position, or averaging down with larger and larger clips, often with a fixed take-profit target.
The appeal is psychological as much as mathematical. A martingale produces a high win rate, because each cycle ends with a small gain as long as one trade wins. The equity curve looks smooth for weeks. Then one cycle fails, and it gives back everything and more.
Three assumptions have to hold for the idea to work, and none of them is guaranteed:
- You have unlimited capital to keep doubling.
- Your broker or exchange allows unlimited position size.
- A winning trade will arrive before you run out of either.
Real accounts have limits on all three. Markets can also gap through a stop, trend further than you expect, or stay irrational longer than your margin lasts.
The math of a losing streak
Here is a simple example. Suppose your account is $10,000 and your base trade risks $100. After each loss, you double. Treat each trade as a coin flip for the sake of illustration.
| Consecutive losses | Size of next trade | Cumulative loss so far | Share of $10,000 lost |
|---|---|---|---|
| 0 | $100 | $0 | 0% |
| 1 | $200 | $100 | 1% |
| 2 | $400 | $300 | 3% |
| 3 | $800 | $700 | 7% |
| 4 | $1,600 | $1,500 | 15% |
| 5 | $3,200 | $3,100 | 31% |
| 6 | $6,400 | $6,300 | 63% |
| 7 | $12,800 | $12,700 | Account is gone |
The pattern is that the next bet is always larger than everything you have lost combined. By the seventh trade, the required size is more than your entire account. That is simple arithmetic, not a prediction about any market.
Now the odds. With a 50% win rate, six losses in a row happens about once in every 64 attempts. That sounds rare until you count trades. Over a couple hundred trades, our own back-of-envelope arithmetic puts the chance of seeing at least one run of six losses at roughly 8 in 10. Real strategies rarely have exactly 50% hit rates either, and losses cluster when the market regime changes. A trend-following entry in a range, or a mean-reversion entry in a strong trend, can lose many times in a row for the same underlying reason.
Recovery gets harder as losses deepen
The table above shows only the dollars lost. The recovery math is worse. Losing 63% of an account means you need a gain of about 170% on what is left to get back to even. A 50% drawdown needs a 100% gain, which we break down in our guide to how a 50% drawdown demands a 100% gain. A martingale leaves you in exactly that territory at the moment you are most tempted to keep sizing up.
What simulations show
One 2026 simulation of 55 strategies on a trading blog compared martingale sizing with fixed larger lots. In one sample, fixed 2x sizing showed 26 units of maximum drawdown and 4.0% Monte Carlo risk of ruin, while a martingale capped at eight steps showed 35 units and 11.0%. In the same analysis, a martingale capped at 16 steps reached 52 units of maximum drawdown and an 18.9% estimated risk of ruin, against 33 units and 6.7% for fixed 2.5x sizing at roughly comparable net results. This is one author's simulation, not a universal law, but it points the same way as the arithmetic: the martingale buys similar results with a heavier tail of ruin.
An academic comparison of forex robots, from a 2018 Indonesian university repository, found a similar trade-off. In that study, martingale produced the highest profit but also high maximum drawdown, anti-martingale produced the lowest drawdown with lower profit, and fixed-lot trading was more consistent in protecting initial capital. Higher profit with high drawdown is the signature of a strategy that is collecting small premiums while selling disaster insurance.
If you want to measure this for your own system, our guides to Monte Carlo trading simulation and cutting risk of ruin under 1% show how to stress-test position sizing before real money is on the line.
Why leverage and margin make martingale risk worse
Most traders who double down are not using only cash. Leverage turns a bad streak into a margin event.
The SEC's margin-account bulletin gives a clean example: a stock falling from $50 to $25 loses 50% in a cash account but 100% of the investor's equity when bought with 50% margin, excluding interest. The same bulletin explains that margin investors can lose more than they invested, can face short-notice demands for more cash or securities, and can have positions sold by the broker without consultation. Investor.gov repeats those warnings for U.S. investors, including that a firm may sell securities, potentially without advance notice, to cover a shortfall.
That matters for martingale traders in a specific way. The moment you most want to add size is when your collateral has already declined. FINRA's margin guidance says declining collateral can trigger a margin call and lets the firm liquidate securities, and that in volatile markets the time you get may be very limited. Your doubling plan does not get a vote.
Forex is no kinder. The SEC's forex bulletin says leverage magnifies small currency moves into larger gains and losses, and that individual traders can lose all of their initial capital or more. The SEC's day-trading bulletin likewise warns that borrowing to trade is risky because traders can lose all their money and potentially end up in debt. Crypto venues add their own liquidation rules, which differ by exchange, so read the liquidation terms of the specific venue you use.
To see how leverage changes the size of the move that wipes you out, work through our practical guide to leverage calculation and the leverage risk management math for retail traders.
Costs and slippage eat the "small win"
A martingale cycle usually aims for a small profit, equal to one base unit. That is exactly the kind of edge that spreads, commissions and slippage can erase. Each added leg of a cycle pays costs again, on a larger size, often in fast markets where fills are worse.
FINRA has shown how serious cost drag can get in frequent trading. A 2017 disciplinary decision describes a pattern where commissions averaged more than 4% per transaction and 8% to 9% of a round-trip return, and costs consumed returns. Retail commissions today are very different from that case, but the principle holds: when your target profit per cycle is thin, costs decide whether the strategy has any edge. If you backtest a martingale, model transaction costs in backtesting across different market regimes rather than assuming ideal fills. Our own slippage analysis playbook shows how to measure what you actually pay.
The sunk cost trap behind doubling down
Martingale thinking is the sunk cost fallacy with a formula attached. The fallacy is letting money you have already lost drive your next decision, rather than asking whether the trade is good from here. "I'm down $1,500, so I have to size up to get it back" is a sentence about the past. The market does not know your entry price and does not owe you a bounce.
You can spot sunk-cost behavior in a few habits:
- Holding a trade past your stop because closing it would make the loss "real".
- Adding to a loser with a bigger size so that a small bounce gets you out even.
- Refusing to take a new, better setup because capital is tied up in an old one.
- Moving a stop further away after price approaches it.
A useful test is to ask: if I had no position right now, would I open this exact trade at this size at this price? If not, the reason you are still in it is sunk cost.
One note on taxes, since people often cite them when they refuse to sell. FINRA explains that selling below your cost basis creates a capital loss that may generally offset capital gains, and that the wash-sale rule generally disallows a loss if you buy the same or a substantially similar security within 30 days before or after the loss sale. Tax treatment depends on the asset and your situation, so check with a tax professional about how it applies to you. The key point: a realized loss is not a reason to hold a bad position.
Anti-martingale: the mirror image
Anti-martingale sizing, sometimes called reverse martingale, flips the rule. You increase size after wins, when the market is confirming your read and you are trading with house money, and you cut size back to base after a loss. Traders use it with trend following and pyramiding: add to a position as it moves in your favor, with stops moved up to protect earlier gains.
Martingale vs. anti-martingale
| Feature | Martingale | Anti-martingale |
|---|---|---|
| Size after a loss | Increases | Decreases or resets to base |
| Size after a win | Resets to base | Increases |
| Typical win rate | High, with small wins | Often lower, with larger wins |
| Worst case | One long streak can wipe out the account | A string of small losses after sizing up, and giveback of recent gains |
| Exposure when wrong | Largest | Smallest |
| Market fit | Rarely fits any regime reliably | Trending markets |
| Psychological pull | "I must win it back" | "I'm hot, I can't lose" |
The academic comparison above is consistent with this table: anti-martingale had the lowest drawdown among the methods tested, at the cost of lower profit.
Anti-martingale is not safe by default. It tends to get hurt in choppy, range-bound markets, where you add size on a breakout that fails and then take the loss at the larger size. Overconfidence after a winning streak is the behavioral version of the same risk. The difference is in the structure: your largest positions are funded by profits, and the damage from a losing streak is bounded by a small base size rather than growing with it.
Is there anything better than martingale?
For most traders, yes, and it is less exciting. These are the controls we would put in place before touching any size-changing scheme.
- Fix your risk per trade. Many traders risk 1% to 2% of equity per trade, sized from the stop distance rather than from the last result. Our guide to sizing stop losses to 1 to 2% risk walks through it.
- Set a losing-streak rule. Decide in advance that after a set number of consecutive losses or a daily loss limit, you stop and review. Our piece on trading streak management covers how to define that for your own trade count.
- Cap drawdown with hard bands. Prop firms use drawdown limits for a reason. See how drawdown bands protect funded accounts.
- Do not deploy all your funds. FINRA advises margin users not to use all available funds, to keep accessible cash reserves, and to expect that firms can change margin policies at their discretion. The SEC bulletin adds that Regulation T generally permits borrowing up to 50% of the purchase price of marginable equity securities, FINRA's minimum maintenance requirement is at least 25%, and firms often set higher levels, commonly 30% to 40% or more, per the SEC margin bulletin.
- Test before you trade. Run the sizing rule through paper trading and a Monte Carlo simulation. If the risk of ruin is not close to zero, the rule is not ready.
- Stand aside. Not every market condition suits your setup. Skipping a trade is a position size of zero, and it is the cheapest risk control there is. Our overview of the best risk controls for new traders covers the basics.
FINRA's guidance on frequent trading adds one more reason for caution: frequent intraday trading, especially on margin, can lose some or all of an investment and potentially more than the funds deposited, and is generally inappropriate for investors with limited resources, limited experience or low risk tolerance. If that describes you, a sizing scheme that escalates risk after losses is the wrong tool.
Our perspective on martingale sizing
We built Discipline AI around a simple belief: a trader's biggest enemy is rarely the market, it is the decision made right after a loss. Martingale sizing is that decision turned into a rule. It feels like discipline because it is systematic, but it is the opposite. It links your risk to your emotions rather than to the evidence in front of you.
So we look at trading through measurable evidence: win rate, risk-adjusted expectancy, reward-to-risk, sample size, and whether current market conditions actually suit a setup. Our Sentinel research system evaluates market structure, volatility, liquidity events and regimes, and helps a trader see when the right call is caution or standing aside. A martingale cannot say "not now". It can only say "more".
We also tell traders plainly that no tool, ours included, guarantees profitable trades. What a good platform can do is make your process visible: what you planned, how you executed, how a losing streak compares with your history, and whether you are drifting toward revenge sizing. If your plan needs you to double after a loss to work, the plan has no edge. Fix the edge, then keep size constant.
— Tony
Build discipline instead of chasing losses
If you have been tempted to size up after a bad day, the most useful step is to see your own behavior in data. Discipline AI is a mobile app for iOS and Android that combines AI-generated trade setups with confidence assessments, position sizing and leverage tools, paper trading, automated journaling, and AI Trade Autopsies that show where execution went wrong. Practice a sizing rule in paper trading before it touches real capital.
The Pro plan is listed at $8.99 per month, $79.99 per year, or $199.99 one-off on our pricing page. If you prefer structured learning, The Disciplined Trader is a $79 one-off. You can also see how Sentinel approaches market conditions.
FAQ
Can you make a living off the martingale strategy?
Not reliably. Martingale sizing does not create an edge; it reshapes risk into many small wins and rare, very large losses. Because the next position must exceed all prior losses combined, one long streak, a margin call or a broker size limit can erase months of gains.
What is the difference between martingale and anti-martingale?
Martingale increases size after losses; anti-martingale increases size after wins and cuts it after losses. Martingale risks the most when you are losing, while anti-martingale risks the most when you are winning. In a 2018 forex-robot study, anti-martingale had the lowest drawdown but lower profit than martingale.
Is averaging down the same as a martingale?
Often it is a softer version of one. If you add to a losing position in larger and larger amounts to lower your average entry, your exposure grows as the trade moves against you. Adding a fixed, pre-planned amount at a level backed by your analysis and a defined stop is different from adding because you cannot accept the loss.
How do I stop sunk cost thinking when I'm in a losing trade?
Ask whether you would open the same trade, at the same size, right now with no position. If the answer is no, close or reduce it. Write your stop and invalidation level before entry, and treat moving it as a rule violation you log in your journal.
What risk controls should I use instead?
Start with a fixed percentage of equity at risk per trade, a daily loss limit, a losing-streak pause and sensible leverage. Test the plan with paper trading and Monte Carlo simulation before going live, and keep cash reserves so a bad week never forces you out of positions.
Sources
- Understanding Margin Accounts (SEC)
- Investor Bulletin: Understanding Margin Accounts (Investor.gov)
- Frequent Intraday Trading: Understanding the Basics (FINRA)
- Capital Gains Explained (FINRA)
- Guidance on Margin Flyer (FINRA)
- Investor Bulletin: Foreign Currency Exchange (Forex) (SEC)
- Day Trading: Your Dollars at Risk (SEC)
- FINRA Office of Hearing Officers decision, 2017
- Martingale versus a bigger fixed lot: 55 strategies (MQL5)
- Pembangunan robot forex dengan manajemen keuangan (Universitas Katolik Parahyangan repository)
- Transaction costs in backtesting (BacktestMarket)