When to sell your crypto: writing your exit plan

Once you are in a position, your judgement changes. A position in profit makes you want to wait a little longer; a position at a loss, to wait for it to come back. In both cases, the decision is put off and ends up being made under pressure.

Writing your exit rules with a cool head solves this problem. Selling at the top is a matter of luck. An exit plan sets, while nothing is urgent, what you will do if the price reaches a given level or if a given fact occurs.

A complete plan covers four cases: the rise, the decline, the thesis that no longer holds, and the position that has grown too large.

Selling in stages means selling a fraction of the position at each level reached. It avoids the choice between keeping everything and selling everything.

Example: bitcoin bought at $50,000, with 25% sold at $75,000, 25% at $100,000, and half kept as long as the thesis holds. At $100,000, the two sales have recovered 87.5% of the initial stake and half of the position is still running.

The levels can be multiples of the entry price or market levels. In both cases, write the share sold at each one: a level without a quantity leaves the decision entirely open.

The stop sets the maximum loss accepted on a position. It is placed where the scenario behind the purchase stops being plausible, at a distance compatible with the asset’s volatility. On bitcoin, moves of 5% within a day are common: a stop at 3% will trigger on an ordinary day.

The stop also determines the size of the position. If you accept losing 1% of your capital on an idea and the stop is 15% from the entry price, the position must not exceed 6.7% of capital.

A stop order does not guarantee the execution price. In a fast fall, the sale can go through below the threshold. That is one more reason to size the position with a margin.

The stop reacts to the price. The invalidation condition reacts to the facts. It describes what, if it happened, would prove wrong the reason you hold the asset: a change in regulation, a lasting loss of users, a stablecoin that loses its peg.

A good invalidation condition is observable and dated. “If the fundamentals deteriorate” allows no decision. “If network activity falls for two quarters in a row” allows one.

When the condition is met, the position is sold whatever the price, including at a profit. This is the hardest rule to follow, because the market can keep rising after the reason to buy has disappeared.

An asset that rises a lot ends up weighing more than planned. A portfolio built with 50% bitcoin can find itself at 65% after a rise, with a risk you never chose.

Rebalancing sells part of what has risen to return to the target. There are two triggers: a fixed date, every quarter for example, or a gap, when an asset moves away from its target by more than ten points. The gap rule acts less often, and only when the portfolio has actually drifted.

Each sale has a cost: fees and, depending on your country, tax on the capital gain. These costs are part of the choice of frequency.

An exit plan fits in a few lines per position: the selling levels and the share sold at each, the stop, the invalidation condition, the rebalancing rule. If one is missing, the corresponding decision will be made in a hurry.

A backtest shows how these rules would have behaved over the price history: how many times the stop would have triggered, what the selling levels would have left on the table, the maximum loss endured. It does not predict what comes next, but it reveals the rules you could not sustain.

What remains is sticking to it. Keeping a journal of your sales, with the rule applied each time, is the simplest way to see whether the plan is followed or bypassed.