Buy the dip strategy for crypto: rules, risks and a backtest

What buying the dip means

Buying the dip means buying a coin after its price has fallen sharply, on the idea that part of the fall will reverse. Traders have done it by hand for as long as markets have existed. The trouble by hand is twofold: crypto falls at any hour, and in the moment it is hard to tell a dip from the start of a crash.

An automated dip-buying strategy fixes the first problem by watching around the clock, and handles the second with rules decided in advance: how big a fall counts, how much to buy, when to take the gain, and when to admit it was not a dip.

How the rules work

The agent watches the price around the clock. When the coin has fallen by your trigger within your window, for example 5% within 24 hours, it buys a fixed amount. It then waits for one of two exits: the price rises to your target above the buy price, or it falls to your stop-loss below it.

Two more limits keep it honest. A maximum number of open buys stops it from buying every step of a long fall, and a daily loss limit pauses the agent for the rest of the day once it is hit.

The five rules, explained

• Entry: a 5% fall within 24 hours. Large enough to skip ordinary daily noise on BTC, small enough to trigger several times in a volatile quarter.
• Size: $50 per buy, at most 2 open. The total at risk is never more than $100, however far the price falls.
• Exit: sell at 4% above the buy price. A modest target that a bounce often reaches.
• Stop-loss: sell at 6% below the buy price. The rule's admission that a fall is not a dip.
• Daily loss limit: $25. If two stops hit on the same bad day, the agent stops for the day.

Each number is a starting point, not a recommendation. Change them in plain words and compare backtests.

A worked example with numbers

Over one stretch, the agent might see this:

1. BTC falls 5.3% in 24 hours. The agent buys $50 of BTC at $86,930 (1 of 2 buys open).
2. Two days later BTC trades at $90,410, 4% above the buy. The agent sells for about $52, a gain of about $1.79 after fees.
3. A week later BTC falls 4.6% in 24 hours. The agent stays out and logs why: short of the 5% entry rule.
4. Later BTC falls 5.1% from a higher level. The agent buys at $91,120, then BTC keeps sliding. At $85,650, 6% below the buy, the stop-loss sells, a loss of about $3.11 after fees.

Two trades, one gain and one loss, and one recorded decision to stay out. The loss was larger than the gain, which is why the share of trades that gain matters as much as the target. This is an illustrative example, not a real result.

When it tends to work

Dip buying does best when a coin swings around a level and recovers after sharp falls. It does worst in a steady decline, where each buy is followed by a lower price and the stop-loss is hit again and again. That is why the backtest always sits next to simply holding the coin over the same dates.

It also does little in a steady rise. If a coin climbs without sharp falls, the rule never buys and you sit in cash while the market moves without you.

When it fails, and what limits the damage

• A crash that keeps going. The first buy hits its stop; the next dip triggers another buy, and it hits its stop too. The cap of 2 open buys and the $25 daily loss limit bound how much this costs in a day.
• A fast fall through the stop. In a sudden move, the stop-loss sells at market and can fill below the stop level. The decision log records the actual fill price.
• Fees on many small trades. A smaller dip trigger trades more and pays more. Every TradeCorgi backtest includes fees, so you see this before going live.

Variations to test

• Different coins. ETH and SOL usually swing more than BTC, so the same 5% trigger fires more often. Consider a larger trigger for more volatile coins.
• A larger dip, a larger target. A 10% fall with a 7% target trades rarely but catches bigger swings.
• An RSI filter. Only buy the dip when RSI is below a level, so the rule waits for an oversold market.
• Alerts only. Run the same rules on Copilot to get an alert and trade by hand.

Reading the decision log

The log is where you learn whether the strategy behaves as you meant. Look for three things: trades that happened when you did not expect them, dips the agent skipped and why, and the gap between each stop level and its actual fill price. If the agent stays out far more often than you expected, the dip trigger may be too large for that coin.

Make it your own

Change the coin, the size of the dip, the window, the target and the stop-loss in plain words, then compare backtests on the same period. Every change creates a new version of the agent and starts on paper before any real order.