Resign or play on: what chess gets right about quitting

Strong chess players resign. Once the evaluation stops being in dispute, playing on only burns time and energy on a position whose outcome they already know, and they would rather spend that energy on the next game. There's no shame attached to it. Resigning is honest accounting about where things actually stand.

Most other fields have no such norm. Projects keep running long after the evaluation is settled. A trading strategy keeps taking positions after its edge is gone, and papers get written to defend the result they were supposed to be testing.

There are two ways to be lost, and they ask different things of you. The first is lost with counterplay. You're objectively worse, but the position still puts hard questions to your opponent and your practical chances are real. The second is flat lost, where the only variance left is whether the other side blunders.

The skill is telling those apart, and the discipline is acting on what you find. My microstructure project went flat lost as a trading strategy the moment the fee math came back, so I resigned the trade and kept everything it had built, the capture infrastructure and the data, along with a negative result worth publishing.

When you go back over your own games honestly, the alibis don't survive. The engine doesn't care that your opponent's move was ugly or that you were down to your last thirty seconds. There was a number on the position, and the number was knowable. The only useful thing to ask after a loss is what would have let you see that number sooner.

"I had a winning position" is the saddest sentence in chess, because it's true and it indicts you in the same breath. You had it. You didn't convert. The gap between the evaluation and the result is where self-deception sets up shop. Research says the same thing in a lab coat: "the strategy should have worked." Should-have is not data. The position is the only thing that scored.

Playing on a lost position is still sometimes the right call. Over the board you keep fighting when your opponent is down to seconds and might flag, or when they're on tilt and likely to hand you a swindle. The same logic carries into work. A contract can oblige you to finish, or a dying strategy might still throw off enough cash to fund the experiment you actually believe in. Resignation is a tool for private decisions. Once other people are keeping score, playing on can be the honest move.

Before you spend another week on a position you suspect is lost, a few honest checks help. Start with the best case: if winning this game is worth less than the game you could start instead, you've already lost on the only clock that matters. Look next at the path to victory, because if it runs entirely through a specific mistake you have no way to force, then what you're calling a plan is really a prayer with a scoresheet. Take your name off the board and judge it as a stranger would, the way you'd advise a friend sitting in the same spot. Count how many results are still live, too. If the best you can reach is a draw and no line gets you a full point, you've stopped fighting for the point and started delaying the handshake. And watch whether new information is still arriving or whether you're just repeating moves. Learning is counterplay. Repetition is the sound of a game that ended a while ago.

When most of those come back empty, resign. Keep playing while you're still buying information, and stop the moment all you're doing is selling time.