I don't really feel like I know the right approach for this metric on this puzzle. I've gone back and forth about whether proliferation is correct, since it costs a few arms, but in practice screwing around with the fires takes so long that grabbing silver seems fine. This is adapted from a first go where I did use proliferation, landing at 430, so this scores better but feels less elegant in the end, since some of it is clearly designed around startup delays that no longer exist. Dunno how to feel good about this metric, or this puzzle in general.