(preliminary warning: this post was written partly by AI - Claude - after a discussion about tactics and decision theory. The situation I was in a few weeks ago: 12-12 at the end of a very tight game, one boule in each team's hands: me, the shooter, and my opponent was pointing but had shot very well when needed. They have a point at 80cm on a very, very difficult pitch. I decide to shoot and let the difficult terrain take care of my opponent. I take out his boule, but he manages to point for the win. No need to say there was a very long talk about that particular tactical choice after the game 😉. Not with my partner though, we did agree on this beforehand) Picture a tight end in a game of pétanque. Two teammates can't agree: one wants to shoot, the other wants to point. The one who wants to point has the boule, so he points. The opponent shoots, misses, and the team keeps the point. The pointer turns to his partner: "See? I was right!" Was he right? Maybe. But what just happened doesn't prove it. The problem is that we'll never know what the other choice would have produced. If he had shot, he might have hit a carreau and the opponent player might have missed his last boule. Or he might have missed, leaving the opponent to point for 2. We only saw one version of the story, the one that actually played out. The other one stays forever in the realm of "what if…". Above all, the outcome depends partly on luck. An opposing shooter who succeeds seven times out of ten also misses now and then. A badly placed pebble, an unexpected bounce, and everything changes. Saying "it worked, so it was the right choice" is like saying it's fine to drive home after a few drinks because you got home safely. This mistake has a name: *outcome bias*. It's the tendency to judge a decision by what happens afterwards, rather than by what we knew when we made it. It shows up everywhere: in games, in business, in medicine, in everyday life. To judge a choice fairly, you have to put yourself back before the throw. What did we know? Was the ground smooth or stony? Was the opposing shooter on form? How many boules did each side have left? A good decision is the one that had the best chance of working *given that information*, even if, on that particular day, it went wrong.