The Concorde Fallacy is the philosophy that if you realise a project is doomed to failure, it's better to cut your losses than insist on finishing.
This is a horribly defeatist strategy, and worse, it doesn't always apply.
1. You can be wrong
On one occasion, we were travelling to an airport when the radio warned that all flights by our airline that day had been cancelled. Rather than waste the effort, we continued on. It emerged that the news report was slightly exaggerated, and the company was able to secure us alternative seats.
2. A half-success is better than none
A party of adventurers has battled through two levels of an ancient dungeon in the belief that it contains a powerful book of spells worth forty thousand gold pieces. Suddenly, the wizard realises they're in the wrong dungeon and the treasure in this one is only worth ten thousand. They continue on regardless, because ten thousand is still worth it.
3. Your luck may change
A scientist discovers a way to create limitless energy, but is pressed to abandon years of research when it's discovered only to work in zero-gravity. Five years later, another scientist discovers anti-gravity. The first scientist laments his decision, and cold fusion is delayed by another five years.
4. Other factors are at stake
There are other important factors than achieving a goal. Your reputation may be at stake, or customer satisfaction may ride on the fallacy. Concorde itself may have inspired a generation of aerospace engineers, and the lessons learned from the project may prove invaluable in the long run.
5. Bravery is commendable
There's a fine line between stupidity and bravery. Sometimes it's important to set an example. Sometimes, if the stakes are high enough, you just have to kick reason to the kerb, pierce the heavens, and so on.
When the Concord Fallacy applies
Sometimes, you genuinely can't win. In such a case, any further expenditure is further loss, at least as far as pertains your impossible goal. Some projects realise this before they even begin, others three years and twelve-billion of government money later.
Take an example: Say an ancient general marches his troops for ten weeks to capture a walled city. When he gets there, it turns out that the city is well-defended. If he retreats now his march is wasted; on the other hand, if he lays siege, he may waste ten months and half of his men, in addition to wasting his ten week march.