Grid trading crypto: a guarded grid bot strategy

What grid trading is

A grid divides a price range into levels. As the price falls to a level below, the strategy buys; as it rises to a level above, it sells. Each time the price swings down and back up through a pair of levels, one round trip completes with a small gain.

Grid bots are popular because crypto spends long stretches moving sideways, swinging inside a range. In those markets a grid turns noise into a steady series of small trades without anyone having to watch the chart.

Where classic grid bots go wrong

A classic grid has many levels and keeps buying at each one on the way down. That works while the price stays in the range. When it does not:

• A fall out of the bottom leaves the grid holding coins bought at every level, the deepest at the worst price. That is averaging down without a limit, and it is where most grid losses come from.
• A rise out of the top leaves the grid fully sold and in cash while the market keeps climbing.
• Fees on dozens of small round trips can take most of the gains.

The guarded grid: how the rules work

TradeCorgi keeps the useful part, trading swings inside a range, and adds the guardrails every agent has:

• A range. The agent only trades while the price is inside the range you set.
• At most 2 entries. Agents never average down beyond 2 entries, so the grid has two buy levels, not twenty.
• An exit per entry. Each buy sells at its own target above its buy price.
• A stop below the range. If the price breaks down, everything open is sold and the agent stops.
• A daily loss limit, which pauses the agent for the day if losses reach your cap.

The five rules, explained

• Range: $84,000 to $90,000. Where BTC has been swinging. Outside it, the agent does not open new trades.
• Entries: $50 at $86,000 and $50 at $85,000. Two levels, so at most $100 is ever at risk.
• Exit: +2.5% from each buy. The $86,000 buy sells at $88,150; the $85,000 buy at $87,125.
• Stop-loss: below $83,000. The range has broken; sell and stop.
• Daily loss limit: $25. A backstop for a fast, bad day.

A worked example with numbers

1. BTC falls from $87,500 to $86,000. The agent buys $50 (1 of 2 entries).
2. BTC dips to $85,000. The agent buys another $50 (2 of 2).
3. BTC recovers to $87,125. The $85,000 entry sells at +2.5%, about $1.15 after fees.
4. BTC reaches $88,150. The $86,000 entry sells, also about $1.15 after fees.
5. BTC falls back to $86,000 and the cycle starts again.

If BTC had instead fallen below $83,000 after step 2, the stop would have sold both entries at about a 3.5% and 2.4% loss, roughly $3 in total after fees, and the agent would have stopped. This is an illustrative example, not a real result.

When it tends to work

Range rules do best when a coin swings back and forth inside a band for weeks, which crypto often does between big moves. They do worst at the start of a trend, in either direction. That is why the range should come from recent price history, and why the backtest sits next to simply holding the coin.

Choosing the range and spacing

• Range: look at where the coin has traded over the past few weeks. Too narrow and the price leaves quickly; too wide and it rarely reaches your levels.
• Spacing: the gap between entry and exit must clear round-trip fees with room to spare. With a 0.1% fee each side, a 2.5% target keeps most of the move.
• Stop: just outside the range, so a real breakdown closes the trade but normal swings do not.

Grid trading versus DCA and dip buying

• A grid trades inside a range: it buys low in the range and sells high, again and again, and does best when the price goes nowhere in particular.
• DCA builds a position over time whatever the price, and does best over long periods when you want to own the coin.
• Dip buying waits for a sharp fall and sells on the rebound, and does best when falls are followed by recoveries.

Many traders run more than one: a weekly DCA buy as a base, and a range rule while the market moves sideways.

Reading the decision log

A range agent's log shows each entry, each exit and every check where the price was outside the range. Look for entries that sold at the target quickly, which suggests the spacing suits the coin, and for long stretches outside the range, which suggest the range needs updating. Every new range you set is a new version that starts on paper.

Variations to test

• ETH instead of BTC, with a wider range and spacing to suit bigger swings.
• A single entry in the middle of the range, for the simplest possible version.
• A trailing exit instead of a fixed target, to let a breakout above the range run a little.
• Alerts only: run the same rules on Copilot and place each trade by hand.

Make it your own

Describe your range, levels, targets and stop in plain words. The agent backtests the rules with your exchange's fees, runs them on paper first, and every change creates a new version that starts on paper again.