What is a stop-loss in crypto trading, and how do you set one?

What a stop-loss is

A stop-loss is a rule that sells a position when its price falls a set amount below what you paid. If you buy BTC at $90,000 with a 6% stop-loss, the position is sold if BTC falls to about $84,600. The point is simple: one bad trade should never be allowed to grow into a large loss.

On an exchange, a stop-loss is usually an order you place in advance. In an automated strategy, it is one of the rules: the agent watches the price and sells when the stop level is reached.

Either way, a stop-loss turns an open-ended risk ("I will sell when I feel it has fallen too far") into a known one ("I will lose at most about 6% of this position").

Why stop-losses matter more in crypto

Crypto moves fast and trades around the clock. A coin can fall 10% while you sleep, and there is no closing bell to pause the slide. Without a stop-loss, a position you meant to hold for a few days can become a position you hold for months, hoping it comes back.

A stop-loss also protects the strategy, not just the trade. A dip-buying rule that keeps buying into a long fall can lose far more than any single trade. A stop on each position, plus a daily loss limit across the strategy, keeps the damage bounded.

How to choose a stop-loss level

There is no single right number, but there is a sensible process:

1. Look at the coin's normal swings. If BTC often moves 3% to 4% in a day, a 2% stop will trigger on ordinary noise. A stop wider than the usual daily range triggers less often by accident.
2. Decide what one loss may cost. With $50 per trade and a 6% stop, one stopped-out trade loses about $3 plus fees. Make sure that number is comfortable.
3. Match the stop to the target. If you aim to sell at 4% above the buy, a stop at 6% below means each losing trade costs more than each winning trade brings in, so the rule needs more than about 60% of its trades to end above the buy price just to break even before fees. That is fine if the backtest supports it, but know it.
4. Test a few levels. Backtest the same rules with stops at 4%, 6% and 8%. Prefer a level where the result is reasonable across all three rather than one that looks best by a small margin.

Why a stop-loss can fill worse than planned

A stop-loss order turns into a market order when triggered. A market order fills at the best price available at that moment. In a calm market that is close to your stop. In a sudden fall, buyers can disappear for a few seconds, and the order fills lower. This gap is slippage.

Slippage is larger on coins with thin order books, during big news, and when the price gaps through your level. It is the reason a 6% stop can occasionally become a 7% or 8% loss. A backtest that includes slippage, and a decision log that records the actual fill price, show you how large it tends to be.

Stop-loss versus stop-limit

A stop-limit order has two prices: the stop, which triggers it, and the limit, the lowest price you accept. It never sells below the limit, but in a fast fall it may not sell at all, leaving you holding a position that keeps falling. For protecting a position, a plain stop-loss is usually the safer choice because it gets you out. Read the full comparison in stop loss vs stop limit.

A worked example

You run the BTC Dip Buyer with $50 per buy, a 4% exit and a 6% stop-loss:

• BTC falls 5.3% in a day and the agent buys $50 of BTC at $86,930.
• BTC keeps falling. At $81,714, 6% below the buy, the stop-loss triggers.
• The market is moving fast, and the sale fills at $81,650.
• The trade loses about $3.05 before fees and about $3.15 after a 0.1% fee on each side.

The decision log shows both the stop level and the actual fill price, so you can see that slippage added a few cents. This is an illustrative example, not a real result.

How TradeCorgi agents use stop-losses

• Mandatory. An agent cannot run without a stop-loss, a budget, a daily loss limit and a maximum number of open positions.
• Checked by a rules engine, not the AI. The language model drafts your rules; a rules engine with no AI in it applies the stop exactly as confirmed.
• Sells at market, so the agent exits rather than waiting for a price that may not come.
• Backed by a daily loss limit. If several stops hit on the same bad day, the agent pauses once the day's losses reach your cap.
• Recorded with the fill price, so slippage is visible.

A stop-loss checklist

• The stop is wider than the coin's normal daily swing.
• One stopped-out trade costs an amount you are comfortable losing.
• The stop and the exit target make sense together in a backtest with fees.
• A daily loss limit covers the case where several stops hit at once.
• You will not move the stop further away once the trade is open.