"Well," Keltham says, speaking less rapidly than usual, because some of his processing just got diverted to a subthread, "we can and do play it for fun, and do that our whole lives, in fact. It just involves a mindset where - you can try to do as well as possible, each time you confront a prediction challenge, without feeling like you're losing something in virtue of doing less than perfectly."
"The most obvious thing to match yourself against, there, is other people's predictions. Pilar predicts one number, Yaisa predicts another, we see who did better, just like if they were playing some other - competitive game whose positive extra is the fun people have from playing it, or their pride in showing their skills. There's a version of that game which dath ilani play, before they're ready to go all the way to prediction markets, where we put up a sheet of paper on the wall - or just use walls and Prestidigitation, I guess - and write down a question, and then people who think the current probability is wrong can write down a different probability underneath and be scored by how much they gained or lost relative to the previous guess."
"I'm not quite sure we're ready yet, but once we are, we'll probably just start doing that over all the place, whenever somebody comes up with an interesting question that will actually settle in a few days."
"But I think the most basic point - lost 2s aren't actually like sending out the merchant ships to the wrong place. They're a measure of how well you do; the fact that the numbers always look negative don't mean you're doing poorly. If you're losing few enough 2s, you can send your merchant ships where they need to go, and that's the reward for playing and losing some 2s."