Sunk Cost Fallacy

Sunk Cost Fallacy is the pull to keep investing in a plan, project, or purchase because of resources already spent, even when the forward-looking case no longer justifies it. The money, time, or code already sunk is gone regardless of what you do next — the only question that should matter is what happens if you continue versus if you stop, and past spend is not an input to that comparison.

The clean version is a movie you're not enjoying: staying because you paid for the ticket doesn't get the ticket price back, and it costs you two more hours you'll never get back either. The costly version is a six-month rewrite that turns out to be worse than the system it replaced — team keeps going because "we've come this far," when the honest comparison is between finishing the rewrite from here and reverting, evaluated purely on the effort and risk remaining in each. The fallacy is emotionally powerful because abandoning something makes the original investment feel wasted in a way that's psychologically vivid even though it's economically irrelevant — the waste already happened either way.

It compounds with identity: the person who chose the failing approach has more to lose personally by killing it than a neutral party would, which is why the decision to cut a project is best made with input from someone who wasn't the one who started it. It also masquerades as diligence — "let's not waste what we've built" sounds like prudence, but it's evaluating the wrong ledger.

The engineering-specific form is the branch nobody wants to close: hundreds of commits into an approach that a fresh read of the problem would no longer choose. The tell is language that cites the past ("we already did X") rather than the future ("X gets us Y"). If a decision can't be justified without a reference to how much has already been spent, that's the fallacy talking.