Hey guys I have a fairly technical question about optimizing parameters in Ninja. Should I optimize the parameters separately, for example, one at a time, or optimize them as a group? I wonder if the parameters would inadvertently interfere with each other in optimization. I may try it both ways and see what results I get, esp. for the ones that aren't really related to the same indicator.

There's no strict right or wrong answer.

I think if you do them as a group you may get better looking results but they will have a greater probability of being over-fitted as every variable is able to adapt to both the data and each other. Doing them one at a time, with the not yet optimised variables set to sensible values, will possibly result in a less fitted result.

It takes more time to do them one at a time though and you may miss out on potentially better solutions since the end result will somewhat be dertermined by the 'sensible' starting values you pick for the other variables.

Personally, I do them all at once but take measures to avoid over-fitting like using large amounts of backtesting data, minimising the possible range of each variable as much as possible and just having an eye and a feel for likely curve-fit results through hard won experience.