The Martingale Trap

Some of the most dangerous trading strategies are the ones that look like they're working. This page is about a specific class of these — strategies that survive by doubling down after every loss — and why they're a trap, not an edge.

If your only way to make a strategy profitable is to increase your position size every time you're wrong, you're not building a strategy. You're building a slow-motion account explosion.

What is a Martingale strategy?

The classic Martingale is a betting system from 18th-century gambling halls. The rules are simple:

1. Bet $1 on red
2. If you lose, double your bet to $2 and bet again
3. If you lose again, double to $4. Then $8. Then $16. Then $32...
4. When you eventually win, you recover all previous losses plus your original $1

On paper this is mathematically beautiful. As long as you keep doubling, the next win pays for everything that came before it, plus a small profit. Since red has to come up eventually, you can't lose. Right?

Wrong. Two things break it:

• You run out of money. A run of 10 losses takes a $1 starting bet to $1,024. A run of 20 takes it past $1,000,000. Real account balances are finite. Once your bet exceeds your bankroll, the system collapses, and you've lost everything.
• The casino has betting limits. Even if you had infinite money, the table won't let you bet beyond a maximum, capping the recovery mechanism at exactly the wrong moment.

In trading terms, the equivalents are: your account balance and leverage liquidations.

What this looks like in trading strategies

Trading Martingales rarely call themselves Martingales. They show up disguised:

• "After a stop loss, increase position size to recover the loss faster"
• "I'm averaging down on losers — when the price comes back, I'll be even"
• "Each losing trade just means the next trade has to be a bigger winner"
• "My win rate is low but my wins are huge, so it works out"

Any of these patterns where position size grows after losses is a Martingale. The doubling factor doesn't matter (1.5×, 2×, 3× — same trap), only that risk-per-trade scales up after a loss.

Why backtests of Martingales look incredible

A 12-month backtest of a Martingale strategy can show 80%+ returns with tight drawdowns. The numbers are real — they really did happen in the simulation. But the result depends entirely on a single variable: did the doubling cycle ever reach the limit of your account before being bailed out by a winning trade?

If yes (the strategy got "saved" before running out of money) → big positive return, looks like a great strategy. If no → the account is wiped out and the backtest ends in ruin.

The same strategy on a different time period can produce either outcome. Backtest one year and the doubling streak you were hoping for shows up; backtest a different year and a 14-loss streak wipes the account on day 200. The strategy itself didn't change; the lucky path did.

This is why running a strategy through out-of-sample testing on multiple historical periods is non-negotiable. A Martingale that's profitable in 2024 will likely be ruinous in 2022 — or vice versa — and the only way to know is to actually try it across multiple years.

The psychological trap

The financial damage from Martingales is bad. The psychological damage is worse.

Consider a trader who runs a Martingale strategy for six months and grows their account from $10,000 to $25,000. The strategy "works." They're calm, confident. The wins are infrequent but big. They're starting to think about quitting their job.

Then they hit a 14-loss streak. The doubling exceeds their account size. They liquidate at maximum leverage. The strategy that grew their money for six months wipes them out in a single afternoon.

The next day they have less than they started with. They didn't make a single mistake — they followed the rules of the strategy exactly. The mistake was choosing a strategy where rule-following inevitably leads to ruin.

The trap closes hardest on people who've already had success with the pattern. The early wins are what convinces you to size up, run multiple positions, increase leverage, and quit your day job. By the time the strategy fails, your exposure is much higher than when you started.

How to spot a Martingale you didn't know you were building

You may have built one without realising it. Check your strategy for any of these:

• A "double position after stop loss" rule of any kind
• An ever-growing collateral or amount variable that resets only after a win, never on a fixed schedule
• Logic that "averages down" on losing positions by adding more contracts at lower prices
• No hard cap on position size relative to account equity
• Strategy backtests with a high win-loss ratio (1:5, 1:10, 1:20) but a low win rate (under 30%) — the math only works if every win covers many losses, which means the wins must be very large, which usually means position sizes were inflated by a doubling cycle

Any one of these is a yellow flag. Two or more is a strategy that will eventually destroy your account.

What to do instead

The fix is not to make the Martingale "smarter." Profit-locking, milestone resets, smaller multipliers — these all just delay the inevitable rather than prevent it. The real fix is to remove the doubling mechanism entirely.

1. Use fixed position sizing

Risk a fixed percentage of your account on every trade — typically 0.5% to 2%. Whether the previous trade won or lost is irrelevant; the next trade is the same size relative to what's currently in the account. This makes blowups mathematically impossible. Your position size grows when the account grows and shrinks when the account shrinks; it never compounds adversely with bad luck.

2. Find a strategy with positive expectancy per trade

Profitability is governed by:

expectancy_per_trade = (win_rate × avg_win) − (loss_rate × avg_loss)

You can be profitable with:

Win Rate	Avg Win	Avg Loss	Style
70%	$1	$1	Mean-reversion / scalping
50%	$2	$1	Balanced swing trading
30%	$5	$1	Trend-following

All three are real. None require Martingale. A 30% win rate trend-follower with fixed 2% sizing outperforms a 60% win-rate Martingale doubler over the long run because the trend-follower can't blow up.

3. Test across multiple time periods

Even with fixed sizing and a positive-expectancy entry, your strategy needs to survive out-of-sample testing across different market regimes — bull, bear, sideways, high volatility, low volatility. If it works in 2024 but fails in 2022, you don't have a strategy; you have a one-period coincidence.

4. Set a hard account-level circuit breaker

Use the Max Drawdown block to stop the strategy if the account drops below a defined threshold (e.g., 30% from peak). This is a final backstop independent of any per-trade logic. Even strategies you're confident in occasionally do unexpected things; the circuit breaker means "unexpected" doesn't become "catastrophic."

A summary you can save

Some uncomfortable but real truths about Martingale-style position sizing:

• It feels like an edge but is a betting system, not a trading strategy.
• Backtests of Martingales reflect lucky paths, not robust performance.
• Real Martingales fail catastrophically — not gradually, not slowly. One bad day wipes out everything.
• The longer a Martingale "works" before failing, the larger your exposure when it does fail. Success is what builds the conditions for the disaster.
• Bounded risk per trade plus genuine entry edge is the only configuration that survives.

If you can't make a strategy profitable with bounded position sizing, the strategy doesn't have an edge — and adding doubling won't create one. It will just postpone learning that fact.