vincent d warmerdam

Is there a minimum viable example that you can share? Without one it is very hard for me to understand the missing behavior.

Your example isn't complete. You didn't add the code that handles the actual logging of stats you see in your chart or the `generate_chromosome` function to generate the candidates. I...

I have a hint of what is happening. You're logging the best candidate per generation *but* you're still applying a mutate. This mutate is what is messing with your expectation....

Mhmm ... *now* I understand. I'll need to ponder this one but I think you're right that this isn't easily implemented with the current API. Maybe a "concat" verb makes...

Would it be sensible to change the `mutate` function to not change the chromosomes of the best performing candidates.

I just pushed version 0.5.2 which should have this elitism feature. @ELC let us know if it works.

This will be picked up by an example that is added to `sphinx`

This does feel like a breaking change. Might a deprecation warning be a nice thing to add here?

I can live with that. But since we're not at version 1 yet, is there a reason you immediately want to jump to 2?

@rogiervandergeer are there other changes we'd like to make for version 1? I mean ... version 1 is a milestone release.