Return for January is: +2.65%
I am very confident with this model for now, wich behaves exactly (even better in fact) than the initial model on demo.
Since, it is not because there's a lot of magic newcomers, that the old ones are not as good as before.
For this reason, to grow in size, I will NOT replace DARWINs, but add DARWINs to my portfolio. And so, with an annualy expected return, and a half-yearly review of the assets - as @CavaliereVerde proposed - the DARWINs that tend to lose in strength will be just... removed gradually.
This way the portfolio should keep a consistent strength and be automatically "self-cleaned" by (times + rules)
The philosophy is:
"not to be confident in the DARWINs, but confident in the process"