[arh68]

I would agree all models are bad, some models are useful even though it goes unstated. I feel like the author’s point is that asking for “the right model” is the wrong question. I don’t mind argument-by-analogy too much, but I think it would help if we knew what the aliens were trying to do with their model. People criticizing "inaccuracy" seem to just have high standards for their models (perhaps you want to simulate 100B chess games in 10 seconds.. how's that coin flip idea sound now?). Just imagine you're dreaming about backgammon: do you know all the rules? no, but you can roughly model what's going on. Your model could be said "worse" than standard rules, but it's good enough for a dream sequence.

More than once have I seen a nice chess set in a home, set up completely wrong. They’ll be close! Their model isn’t too far off from standard, but just so (K & Q reversed, B & N reversed, &c). I presume the owners were quite content. Honestly, I don’t even know if most people would notice, so why point it out (is my model even the “preferred” one)?