News for hackers. Stuff that matters.
Re:HTML5 Genetic Cars by Kiro on Wednesday June 26, 2013 @09:27AM (#5945837)
I'm using 5% mutation rate and have 1 elite clone. I thought that meant that the whole population would be based on the winner with a 5% mutation chance (so all 19 are children of the winner). However, I have a track with an "impossible" hill and every round there's one car that almost makes it over but it's never that one which is blue next round.

How does this work exactly?

EDIT: It seems the one that almost made it was a replay ghost or something and not part of the population. Still wondering if I understood it correct though.

Re:HTML5 Genetic Cars by meandthebean on Wednesday June 26, 2013 @01:35PM (#5947551)
I recommend a smaller mutation rate. You see the inheritance of the previous winner a lot better.