The sunk-cost fallacy is continuing something because of what you've already invested — money, time, or ego — rather than its future prospects. In trading, it's holding (or adding to) a loser because you're "already in this deep."

The past is gone

What you've already lost on a position is a sunk cost — it's gone whether you hold or sell, and it should have zero weight in the decision (how to take a loss). The only rational question is forward-looking: given the current setup, would I open this position right now? If not, the fact that you're already in doesn't change the answer (process over outcome).

Averaging down: the trap in action

Sunk cost powers the worst version of this — averaging down, adding to a loser to "lower your average" and justify the original bet (the danger of averaging down). Now you've thrown good money after bad and grown the position exactly as it's proving you wrong (the martingale trap).

The money you've already lost isn't a reason to stay — it's the exact thing clouding your judgment about whether to leave. Decide as if you're flat.

Letting go

Anchor every decision to a pre-set stop and the question "would I enter here now?" — not to your accumulated loss or ego (stop-loss orders, anchoring bias). A mechanical system carries no sunk cost — it never holds a loser out of pride or to justify a prior decision; it exits on rule (emotional discipline).