What it really is
Grid trading is an algorithmic strategy that capitalizes on price oscillation within a range. The trader (or bot) defines an upper and lower bound and a step size; the bot then places buy orders at fixed intervals below the reference and sell orders above it. Each time a buy fills, a sell at the next level up is queued; each time a sell fills, a buy at the next level down is queued.
Grids do not predict direction — they monetize volatility. In sideways crypto pairs they can be highly profitable; in trending ones they can quickly become a losing position.
How it actually works
Grid parameters
You set: range (e.g., $90-$110), step size (e.g., $1), and order size per level. The bot computes the number of grid levels and places limit orders accordingly. Each filled buy starts accumulating inventory; each filled sell realizes profit on the step.
Where grids profit
In a sideways market, price crosses the same levels repeatedly. Each round trip — buy at $99, sell at $100 — earns the step minus fees. Over many oscillations, the cumulative profit can be substantial. Stablecoin pairs, mature L1 pairs in low-volatility periods, and predictable sideways ranges are favored conditions.
Where grids fail
Strong directional trends. If price breaks out above the upper bound, all buys are filled and the bot is fully long — but the sells at the top are unfilled because price keeps going. If price breaks below the lower bound, all sells fire empty and the bot accumulates losing inventory. Grid bots are short volatility but long range.
Grid variants
Linear grids use equal-sized steps; geometric grids use percentage-based steps (more levels at lower prices). Some bots dynamically widen the range if price approaches the edge. None of these fix the fundamental "trend kills grid" problem.
A worked example
Range: $90-$110. Step: $2. Order size: $100. Starting at $100, the bot places buys at $98, $96, ..., $90 and sells at $102, $104, ..., $110. Over a week, price oscillates between $94 and $106 thirty times, filling about 60 round trips. At ~$2 per step, gross is ~$120 minus fees. If price breaks above $110, the bot exits with no inventory and the upside not captured. If price breaks below $90, the bot is stuck long at an average above market.
Common mistakes
- Running grids in trending regimes. The wrong regime kills profitability.
- Range too narrow. Constant breakouts.
- Range too wide. Few fills, low utilization of capital.
- Ignoring fees. Many small trades amplify fee impact.
- No risk cap. A breakout can leave the bot holding more loss than expected.
How traders use it
Grids are best deployed on assets and during periods identified as range-bound. Combine with a regime view from technical analysis in crypto and a defined kill-switch when price breaks the range. Pair with crypto risk management for position sizing.
Pick the range with context
Range selection is not just chart-based; events (unlocks, regulatory actions, listings) frequently break ranges. Zippfeed surfaces crypto headlines with sentiment and importance scoring so you can avoid running grids into known catalysts. None of this is financial advice; it is the context that distinguishes a sideways range from a trend in waiting.
