Why Never Sell Down Is Dangerous

Your idea has an intuitive appeal:

Buy when movement is expected.
Hold while down.
Sell as soon as profitable.
Never sell at a loss.

The problem is that this becomes a known behavioral trap: sell winners too early and hold losers too long.

The Trap

Suppose the bot buys three positions:

TradeEntryCurrentRule saysProblem
A100101SellTiny win is locked early
B10095HoldLoss can keep growing
C100120SellBig winner may be cut too soon

If this repeats, the account can fill with stale losers while winners disappear quickly.

The Better Rule

Do this instead:

Never sell only because it is down.
Never hold only because it is down.
Sell or reduce when the thesis is invalidated.
Let winners continue while the thesis remains strong.
Protect winners with trailing stops or partial profit-taking.

Invalidation Beats Hope

An invalidation is the condition that proves the trade idea was wrong.

Examples:

Mean reversion trade:
Invalid if price breaks below support and stays there for two cycles.

Momentum trade:
Invalid if volume fades and price falls back below breakout level.

Portfolio trade:
Invalid if the asset becomes too concentrated or drawdown worsens.

Profit Is Not The Only Sell Signal

The bot can sell when:

  • The thesis is wrong.
  • Risk is too high.
  • Concentration is too high.
  • A better opportunity needs cash.
  • A planned timeout expired.
  • A trailing stop protects a winner.
  • The target was reached and reward/risk is no longer attractive.

Beginner Example

Bad policy:

If position is profitable, sell immediately.
If position is losing, hold until profitable.

Better policy:

If position is profitable and trend remains strong, hold part and trail the stop.
If position is profitable but near resistance, take partial profit.
If position is losing but thesis remains valid, hold or reduce size.
If position is losing and thesis is invalidated, close or reduce.

Gotcha: Refusing to realize losses does not remove losses. It only hides them as unrealized losses.