> Grimmer found that the fastest sequences always had one thing in common: The middle step was always a big one. So... it's like binary-search, but you don't k

> Grimmer found that the fastest sequences always had one thing in common: The middle step was always a big one. So... it's like binary-search, but you don't know the maximum before hand? (i.e. you double the step size until you overshoot, then do regular binary search from there)