On Thu, Nov 6, 2008 at 5:52 PM, Ed Murphy <[EMAIL PROTECTED]> wrote: > No, e would go to -1 points and stay there. In general, though, > even small complex scores would inflate way too quickly, e.g. > > (S = 0) 2+i -> 3+4i -> -7+24i -> -527-336i > (S = 1) 2+i -> 4+4i -> 1+32i -> -1022+64i > > A better analogue to the Mandelbrot set would scale the winning > condition down from 100 to 2, thus do something with P/50 and S*50, > but I'm having trouble figuring out how to tune it so it doesn't > pull too hard in one direction or the other.
Ah, I misunderstood the proto and thought that each axis was being updated independently. Now it makes sense to me what you're trying to do. Note the only possible score index that's actually in the Mandelbrot set is 0. There would have to be some clever mapping for the other possible score index values to keep them within the set. Another potential problem is that being in the Mandelbrot set only means that the sequence starting with 0 will remain bounded. As far as I know, there's no guarantee that a sequence starting from an arbitrary score will remain bounded even if the score index is in the Mandelbrot set. -root
