Reading the win rate as an edge
A grid closes most rungs in profit by design. A 90% win rate says the grid was inside its range, not that the strategy is profitable over a full cycle.
A strategy that places orders at fixed price intervals above and below a reference price, profiting from oscillation rather than from predicting direction.
also: Grid system, Price grid, Grid EA
Updated · Reviewed
Instead of guessing which way price goes, the EA lays a ladder of orders at even distances and lets price walk into them. Every step price retraces closes a rung in profit — and every step it keeps going adds another open position.
Grid EAs are the largest single family sold to retail traders, and they are the family whose equity curve lies most convincingly. A grid can print months of small, even gains while carrying an unrealised loss that never appears in the closed-trade record until the day it does.
A grid with a 20-pip step and five levels below the reference price. Price falls through three of them, so three positions are open at once and the account is carrying the distance between each entry and the current price.
Calculation 3 filled levels × 20 pips average distance × 0.01 lots ≈ open loss of 60 pip-lots before the fourth level fills
Result 3 open positions, all in the same direction, all losing
Illustrative figures — not the record of any listed EA.
Judge a grid EA by what happens when the range breaks, not by the equity curve while it holds.
A grid can be a sound strategy with a fixed lot, a bounded rung count and a stop that closes the whole basket. It is the unbounded version that produces the equity curves that end at zero.
A grid closes most rungs in profit by design. A 90% win rate says the grid was inside its range, not that the strategy is profitable over a full cycle.
A grid is about where orders sit; a martingale is about how the size grows. They are frequently combined, and the combination is what turns a bounded loss into an account-ending one — but a fixed-lot grid is a different risk profile.
Mean reversion is a tendency, not a law. Currencies trend for months; the recovery a grid depends on may arrive after the margin call.
Choose the window and you choose the result. A grid tested only on a sideways period is being asked the one question it always answers well.
| Across the 14 EAs published here | Figure |
|---|---|
| Records where equity drawdown exceeds balance | 14 of 14 |
| Median gap between the two | 0.61 points |
| Widest gap (Cairn) | 11.10 points |
| Cairn, balance then equity drawdown | −14.52% then −25.62% |
| Longest losing streak, median | 6 |
| Longest losing streak, worst (Cairn) | 14 |
Read the Cairn row twice. It wins 64.04% of 4,725 closed trades and still lost fourteen straight, because a streak follows how many times you roll, not how often you win. A grid answers that streak by adding rungs, and a doubling ladder covering the median six needs 64 times the first lot; covering Cairn’s fourteen needs 16,384.
The gap column is the other half. These fourteen hold one position per symbol at a fixed lot, with a stop attached — the design furthest from a grid — and equity still dipped below balance in every record. Cairn’s balance drawdown of 14.52% describes an account that was really 25.62% down. On a strategy that holds losing rungs open by design, that distance is the product, not a rounding error.
MT5 prints both. The Strategy Tester’s Results tab lists Balance Drawdown Maximal and Equity Drawdown Maximal on separate lines, and a grid listing tends to quote the first. Read them as a pair: their ratio is the closest thing to a grid detector a backtest report contains, and our methodology publishes both for every listing.