Maps

[Fritzlein]

Thanks for the stats, Jeff! I was going to suggest eliminating any map with a draw rate of 10% or above; I had no idea the draw problem was so bad. Still, the relative numbers seem like a great guide for getting the worst offenders out. Also, it seems you could test new symmetrical boards and quite quickly get an idea of whether to keep them or throw them away.

Thanks again for evolving the contest to be better and better so as to keep up with the bots that are evolving to be better and better!

[Fritzlein]

More new maps. Yay! Thank you!

I notice, though, that on all the maps where both players start out on the same color square (if it were a checkerboard) have a lot of draws from the players crashing into each other. If the boards were all made even parity (e.g. 15x16) then the players could have symmetrical starting positions while being on different color squares. If the players always started on different color squares, they would not draw by crashing into each other, only by claiming equal territory. Therefore the number of draws would go down.

Just something to think about for designing those final maps.

[jeff.cameron]

Great point, Fritzlein. We're going to throw in 60 more random maps that have even-odd side lengths.

Our current method for generating the "final maps" is to throw in some random maps, watch them for a few minutes, then take out all but the ones with the lowest draw rates.

[Fritzlein]

Great, thanks for working towards a low draw rate, and thanks for the tip about how the final maps will be created!