Grid trading

In a sideways range a grid prints steady small wins and looks unstoppable. Every filled order spawns a fresh one at the next level, so the ladder never stops trading and the equity curve climbs in a neat, hypnotic staircase.

Then the market trends. The losing side of the ladder stacks up, and open drawdown swallows every profit the grid ever booked. There's no exit built into the logic that says stop, because the whole system is a bet, never quite stated out loud, that price eventually returns to where it started. A martingale adds size to a losing position on the assumption of eventual reversion; a grid does the same thing spatially, laddering size across price instead of across time. Different mechanics, same wager. See martingale.

We ran 76 grid and DCA systems through real bid/ask data. All of them failed, 100 percent, no survivors, the sharpest single result in the audit. DCA variants that scale into a position as it moves against them share the same fate for the same reason: size grows exactly when the trade is going wrong.

The failure isn't bad settings. It's structural: a grid is short volatility and long hope, and eventually volatility sends the bill. Short volatility means the strategy profits while price stays inside a range and loses, without limit, the moment it doesn't. No limit, no exit. No amount of grid spacing, take-profit tuning, or position sizing changes that exposure. It only changes how long the tidy equity curve lasts before the trend arrives.

Reported returns on a grid almost never show this honestly. A demo run over a calm three-month window can look flawless while the tail risk sits off-screen, invisible until the one month it isn't, and that gap between the demo and the real distribution is exactly where grids get sold. Drawdown, not the smoothness of the equity curve, is the number that tells the real story.

The tidy equity curve is real right up to the day it isn't.