I've recently come across with these two articles from Cesar Alvarez:

http://alvarezquanttrading.com/blog/equity-curve-monte-carlo-analysis/
http://alvarezquanttrading.com/blog/connorsrsi-strategy-sensitivity-analysis/

The first one introduces an interesting idea about the way we use to perform Montecarlo analysis and the problems associated with it.

I don't know if Zorro could handle a similar analysis, maybe CBI does it already. But if not, i think that simply assuming your 50th percentile CAR to be 50% lower and your 50th percentile MDD to be 50% higuer could suffix this. That is, putting on the extremely precautious side.

For the second one, i've thought it'd be interesting if optimize function could have a parameter for controlling sensitivity.

Although optimize does a real good job selecting robust parameters, i've found a few situations where a relatively narrow peak has been selected instead of a broader range with worse performance.

Theorically that broader range should be more live trading proof.

This new feature could determine that a certain parameter value and a percentage range of other values surrounding it have an objective value range similar. Exactly what optimize is currently doing but letting the user to control the length of the range.

Maybe this can actually be made with current Zorro capabilities, if that's the case then i'd like to know about it. If not, i think it could be handy to have this.

Thanks.