Those individuals have been evaluated against a completely different  

set of opponents.  You may want to just grab the best individual from  

the last generation, as opposed to the best individual of the run  

(which does not make that much sense in a coevolutionary setting such  

as the one you use).



On Sep 1, 2008, at 9:36 AM, Andrew Wagner wrote:



> I'm trying to figure out how to use ec.coevolve.CompetitiveEvaluator.

> I looked at the included coevolve1 problem. Running it, produces a

> stat file with some strange results. The end of it says:

>

> Generation: 99

> Best Individual:

> Evaluated: T

> Fitness: 5.0

> 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

> 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

>

> Best Individual of Run:

> Evaluated: T

> Fitness: 5.0

> 1 0 0 0 1 1 1 0 1 1 1 1 0 0 1 1 0 1 1 1 0 0 1 1 0 1 1 0 1 0 1 0 1 1 1

> 1 0 1 1 0 0 0 1 1 1 1 0 0 1 1

>

>

> This makes no sense to me. The best individual of the last generation

> is better than the best individual of the run (has more 1s), but has

> the same fitness. What gives?