Simulation senstivity when it comes to turnover and single stock performance

Just as a test, I ran simulations to see the sensitivities a simulation has when it comes to turnover (rank < x) and the effect that one or two single stocks have on the overall performance.

Here are my US test;

Here is the EU test:

And here are my global ranking system:

I think it shows how simple it is to push the performence in a simulation to the overfitted ekstreme.

Here I did the same by just adding new single-buy rules: