Grid trading
SystematicIntermediate
A grid places buy orders at steps below a price and sell orders at steps above it, so a range that bounces between the steps closes a series of small trades. It assumes the market will oscillate. Investopedia's description of the against-the-trend grid is blunt about the failure mode: if price runs one way, the grid keeps adding to a loser. Commercial bots that promise this will earn on its own are not evidence that it does.
Not investment advice. Nothing here is a recommendation to buy or sell anything. Backtests and historical studies are not guarantees, and a rule that worked in one sample can fail in the next. Trading can lose money. Links are for education, not endorsements. Do your own research.
- Rules
- Choose an anchor, a step size, often a fraction of ATR, and a hard maximum number of rungs so the entire stack can be lost without blowing up the account. Exit or shut the grid off when price leaves the planned range, when ADX shows a strong trend, or when a pre-set loss is hit. Do not increase size after a loss. That is a martingale, and it is a different and worse rule. Each rung needs a profit target at least large enough to cover the spread and the fee.
- Where AI fits
- A regime filter is the responsible use of AI here: leave the grid off when a trend model is confident, and on only in a range. An LLM can code the order ladder and the kill switch. It should not be asked to widen the grid after price has already left it. Any backtest has to include fees, because the edge per rung is small.
- Markets
- Range-bound FX pairs or other liquid markets with a real mean-reverting history. A grid on a trending crypto coin, especially with leverage, is the version that fails loudly.
- Timeframes
- Intraday to daily, matched to the step size. A tight grid on a volatile market will fill the whole book in one move.
- Risks
- A one-way trend fills every order on the losing side and none of the exits. Gaps skip rungs and then there is no average price that gets you out. Leverage and martingale sizing turn a slow loss into a wipeout. Advertised bot returns are not a track record.