Heh yeah I figured it would be a lot easyer to check conspecs that way...
I also though about exanding the alge a litle, let it reproduce and pass allong the saved .memloc/.memval keys, and have a lot of genes with a shoot or eye or tie comand in it and delete the genes if what it sees has less shoot/eye/tie commands, and spread them as a virus if the oponent has more...
Basicaly try to slowly addapt to any simple conspec while running away...
Release it on IM so you would either need a propper conspec or no conspec at all
Would probably only work one way though... only deleting genes... it would be easy to spread the conspec to other bots since they'll just addapt to anything they see...
Ofcourse something like alga minimalis would break that completely....
I guess it's not so easy to addapt to the realy simple conspecs...
Either way it's still fun for testing... I know excalibur dies out against it rather fast, even with the random .memloc check...
It also hit me I should probably have tested my own conspec against the alge... cause I just did now and it failed
So I have to think of something more clever... or go back to *.refshoot and *.refeye and such... funny how the simplest conspecs seem to be the best...
Think I'll just need to get the *.refage thing into .out10... it was originaly designed to add robage to the key in .out10... but something is wrong with it... theres like 16 cycles of difference somethimes... no clue why...
I'll have to run a few tests and make an update
I'd rather not use *.refshoot and *.refeye, since I want to try and fake those aswell...
But I know refage keys can work, used it in some bot I made earlyer... will just have to have a wider margin of error for the key and seperate the key values as much as possible.
May aswell do something new since it needs to be updated... so stupid to make a test bot and not run my own bot against it