You find an EA with a 95% win rate. The equity curve goes up in a straight line. Starting capital $10,000, ending at $38,000 after one year. Almost no drawdown. Your first instinct is to deposit money immediately.
Stop.
Behind that curve, there's almost certainly a Martingale or grid strategy. These aren't scams. Professional traders use both. But if you don't understand how the risk works, that straight line eventually goes to zero. We've watched it happen dozens of times with traders who contacted us after the fact.
This article isn't here to scare you away from these strategies. It's here so you can evaluate them honestly and, if you choose to use them, set the safety limits that keep you in the game.
How Martingale works
Martingale originated in casino gambling. The idea is simple: after every loss, double your bet. Win once and you recover all previous losses plus profit equal to the original bet.
In forex: open 0.01 lots long on EURUSD. Price drops 30 pips, open 0.02 lots. Drops another 30, open 0.04 lots. Each new position is larger than all previous positions combined. When price rebounds even slightly, the total position turns profitable.
This is where the 95% win rate comes from. Most of the time, price does come back. The strategy closes in profit, the equity curve ticks upward, and everything looks wonderful.
Until the one time price doesn't come back.
How grid trading works
Grid trading is Martingale's cousin. Instead of doubling down in one direction, a grid EA places orders at fixed intervals (every 20 pips, for example) above and below the current price. Price goes up, buy orders trigger and profit. Price goes down, sell orders trigger and profit. Price oscillates, the grid catches money on every swing.
The simplest grid is bidirectional and can profit in either direction. More aggressive grids add a position multiplier (now it starts looking like Martingale) or only trade one direction. Many EAs combine both: grid spacing plus Martingale lot sizing. Losing positions stay open while the EA adds more, then closes everything in profit when price reverts.
Why the backtest looks too good to be true
Because it is.
Martingale and grid strategies hide their losses. A traditional strategy closes losing trades and the equity curve shows the dip. Martingale holds its losers open and keeps adding. As long as price eventually comes back, those positions close in profit and the backtest report only shows closed trades. The floating loss, which might have been 60–80% of the account at its worst, never appears in the win rate or net profit numbers.
What a typical Martin EA backtest looks like:
- Win rate: 92–98%
- Balance curve: smooth upward line
- Maximum drawdown (realized): 5–10%
What you don't see:
- Maximum floating loss: possibly 60–80% of account equity
- Some positions held for weeks or months waiting for price to revert
- If price never reverts: account zeroed
The balance curve shows closed trades. The equity curve shows reality. When evaluating any Martingale or grid EA, always look at the equity curve, not the balance curve. If the backtest tool you're using doesn't show the equity curve, the numbers are incomplete.
The math of doubling down
Starting position: 0.01 lots, doubling each layer, 30 pips between layers.
| Layer | Position | Cumulative lots | Cumulative floating loss |
|---|---|---|---|
| 1 | 0.01 | 0.01 | -$3 |
| 2 | 0.02 | 0.03 | -$12 |
| 3 | 0.04 | 0.07 | -$33 |
| 4 | 0.08 | 0.15 | -$78 |
| 5 | 0.16 | 0.31 | -$183 |
| 6 | 0.32 | 0.63 | -$408 |
| 7 | 0.64 | 1.27 | -$918 |
Starting with 0.01 lots looks harmless. By layer 7, the floating loss is $918 and total exposure is 1.27 lots. On a $1,000 account, layer 7 is game over.
And that table assumes 30 pips between layers. EURUSD regularly moves 100–200 pips per day. A trending week can push 300–500 pips. During NFP or a surprise central bank decision, 200 pips in an hour is normal.
The question isn't "what's the probability of hitting layer 7?" The question is "can your account survive layer 7?" Over a long enough timeline, extreme moves are guaranteed.
When Martingale and grid strategies blow up
Strong trending markets are the natural enemy. Price moves one direction without looking back. Every added position loses more money. In 2022, USDJPY rallied from 115 to 152, a 3,700-pip run. Any Martin EA shorting that move never recovered. The 2015 Swiss franc crisis saw EURCHF drop nearly 2,000 pips in minutes. No layer count survives that.
Liquidity dead zones amplify the damage. During holidays, spreads spike to 5–10x normal levels. The EA enters at terrible prices, the position sizing math breaks down, and the adding logic spirals out of control.
High swap costs compound over time. Martingale strategies hold positions for extended periods. If you're paying overnight swap on 10+ open positions simultaneously, the interest alone drains your account steadily. In a high interest rate environment, this cost is substantial.
How to make these strategies survivable
We're not telling you to avoid Martingale and grid strategies entirely. In ranging markets, they're genuinely efficient. The key is installing guardrails.
Cap the maximum number of layers. This is the most important rule. Unlimited position adding equals guaranteed blowup. It's only a question of when. Set the maximum to 3–5 layers. When you hit the cap, take the loss and start fresh. Three-layer Martingale totals 7x the starting position (1+2+4). Five layers: 31x. Seven layers: 127x. The difference between 3 and 7 is the difference between a manageable loss and an account wipeout.
Set a total risk exposure cap. All positions combined cannot exceed a fixed percentage of account equity. If the sum of all floating losses exceeds 20–30% of the account, force-close everything. This rule needs to be in the EA code or monitored by an external tool. Don't rely on manual intervention. When the drawdown is happening in real time, you'll convince yourself to "wait a little longer, maybe price comes back." It usually doesn't.
Use a trend filter. Add a simple directional check: only go long above the 200 MA, only short below it. This prevents the EA from adding positions against an obvious trend. Multi-timeframe filters are even better. Daily chart confirms direction, hourly chart executes the grid. When the big picture is wrong, the EA sits out. We use this approach in several of our EAs that include grid elements.
Pick the right instruments. Cross pairs like EURGBP and AUDNZD tend to range more than major pairs. They're naturally better suited for grid strategies. EURUSD and GBPUSD are more trend-prone and more dangerous for Martingale. Also check the ADX indicator: below 25 usually means ranging conditions (good for grids), above 30–40 means a strong trend (stay out).
Set a hard stop for the entire grid. Not per-order stop losses. A total account stop loss. When equity drops below a specific level, close everything and walk away. Define this number before you start. For a $1,000 account: "if I'm down to $750, everything closes." Write it down. No exceptions.
Softer alternatives to pure Martingale
If you like the concept but not the risk curve, there are less aggressive versions.
Fibonacci scaling uses the sequence 1, 1, 2, 3, 5 instead of doubling. After 5 layers, total position is 12x the starting lot (vs 31x for pure Martingale). Same concept, much slower growth, more room for error.
Fixed lot grid uses the same lot size at every level. No multiplier. Five layers of 0.01 lots equals 0.05 lots total. Risk grows linearly instead of exponentially. Recovery requires a larger price retracement, but your account isn't on the edge of a cliff after layer 4.
Small grid with stop losses uses tight spacing (10–15 pips) with individual stop losses on each order (20–30 pips). Each loss is tiny. You win by volume. This is closer to high frequency scalping than traditional Martingale, and the risk profile is completely different.
How to evaluate a Martingale EA properly
When you're looking at a Martin or grid EA's backtest report, these metrics matter far more than win rate or net profit:
Maximum floating drawdown. Not "maximum drawdown" from the report summary, which usually only counts closed positions. You need the equity curve, not the balance curve. If equity dipped to 40% of its peak while the balance curve stayed flat, the balance curve is lying to you.
Maximum simultaneous positions. If the EA held 15–20 orders at once during backtesting, that tells you how aggressive the adding logic is. Cross-reference this with the account size to determine if the margin requirements are reasonable.
Maximum holding time. If the longest position was held for 3 months before closing in profit, ask yourself: can you handle watching an unrealized loss for 90 days without turning off the EA?
Backtest coverage. The data must include at least one strong trending period. The COVID crash of March 2020, the 2022 dollar rally, the Swiss franc crisis. If the backtest conveniently avoids these events, the results mean nothing. Use 3–5 years minimum covering different market regimes.
Account-to-lot ratio. If the backtest uses $10,000 with 0.01 starting lots and 5 max layers, the risk might be reasonable. If you run the same parameters on $1,000, risk is 10x higher. Make sure your account matches the backtest assumptions. Our position size calculator can help with the math.
Rules for live trading
After years of testing Martingale and grid strategies at FXTool, these are non-negotiable:
Demo test for at least 3 months. Not 3 days. You need to see how the EA handles different market conditions, including at least one trending week. If it can't survive demo, it won't survive live.
Use money you can genuinely lose. Martingale has inherent blowup risk no matter how many safeguards you add. Allocate 20–30% of your total trading capital to Martin strategies. Manage the rest conservatively with trend following or single-entry approaches.
Withdraw profits regularly. Account grows from $1,000 to $1,500? Pull out $300. Even if the account eventually blows up, you've banked real money along the way. We recommend withdrawing at least 30% of profits monthly.
Never run multiple Martingale EAs in the same account. Two Martin EAs can add positions in the same direction simultaneously. The risk multiplies instead of diversifying. Use separate sub-accounts or make sure the EAs trade uncorrelated pairs.
Have a written exit plan. At what equity level do you turn off the EA? After how many consecutive losing months do you abandon the strategy? Decide this before you start, write it down, and follow it. Deciding during a drawdown is too late because emotions are already in control.
FAQ
Will Martingale EA definitely blow up eventually?
Unlimited Martingale with no layer cap? Yes, given enough time. But with layer limits, total risk caps, and trend filters, the behavior becomes closer to "planned counter-trend trading" than pure gambling. Blowup is no longer inevitable, but occasional significant losses are. The tradeoff is accepting controlled losses instead of catastrophic ones.
What market conditions suit grid EAs best?
Low volatility, narrow range markets. Specifically: a currency pair moving back and forth within 100–300 pips without a directional breakout. The Asian trading session, pre-holiday markets, and naturally range-bound cross pairs (EURGBP, AUDNZD) tend to be the best environments.
How much capital do I need for a Martingale EA?
Calculate the total floating loss at maximum layers and make sure it's under 30% of your account. If max layers produce $500 of floating loss at worst, you need at least $1,700. Give yourself margin beyond the minimum because markets are more creative than your worst case assumptions. See our realistic income expectations guide for sizing context.
Will the live win rate match the 98% backtest win rate?
No. Live trading adds slippage, spread variation, and execution delays that the backtest doesn't simulate. Live win rates typically run 3–8% lower than backtest. More importantly, the 2% of trades that lose in live trading tend to lose more than the backtest predicted, because real market conditions during extreme events are harsher than historical simulation.
About the author: The FXTool team builds and tests MetaTrader trading tools daily. We run every EA we sell on live accounts and publish the results. This guide reflects what we've learned from building 50+ EAs and working with thousands of retail traders.
Forex trading involves significant risk and may result in total loss of capital. This article is for educational purposes only and is not investment advice. Understand the risks and consider your financial situation before trading.