Personally none of my live successful strategies use standard period values for any of the indicators. Overall I've noticed in 13 years of live trading that the standard period values of almost any indicator have no value for successful live trading. Hence I am not using the option and don't think it will be any useful for EA Studio / EG. Just my 2 cents.

Overfitting is an issue indeed, it's with any kind of datamining, not just in trading, and with any platform. But it's not avoided by restricting indicator periods to some standard values, but by giving the creation process full freedom and using a large enough data set to stimulate model generalization, stability tests and multi-market validation and huge OOS and hold out periods for the final verification.

As said, my strategies that pass these conditions (which are less than 0.1%) and are profitable live, do not use any standard indicator periods, so I know it's a dead end on this one after 13 years of 24/7 strategy building.

It absolutely is, I wouldn't trade a strategy which only works on one pair and falls apart on other pairs completely. It means that it is too curve matched if it does that and chances are 99% that it won't work live. Think of the other markets as an extra big OOS verification. Less than 0.1% of the generated strategies achieve that, but the once that do, are almost guaranteed to be profitable live.

I am building on data from 1987 to 2017 on H1 and validate with OOS from 2018 to 2023. I am using my own data for this which I've compiled and cleaned over the last decade of working in this field. It's available for others too, PM me if interested.