[Simple Optimization] Fitness Stability Method

I'm not even sure it is possible. You could find "stable" area of fitness (which is just one input), and then look at similarity of parameters for the individual fitness results, but I'm not sure how to evaluate this and translate it to stable area of parameters.

Hii Mark,

i would suggest going that route:

as shown in the screenshot,

we make some optimizations 1000 max as shown in the screenshot,

33% up & down with 10 steps,

we then should get a histogram chart just like with SPP or with SeqOpt,

this chart will include ALL the optimizations results and find the stable area just like with SeqOpt, only this time this will include ALL the parameters,not just 1,

we would like to sort the histogram with Fitness, but it would be cool if we had a Ranking tab that we can sort by more rules like:

Profit Factor (Simple Opt's Stable area) > 1

# Trades (Simple Opt's Stable area) > 30

thanks.

-best lowest score of area

-best highest score of area

etc.

Picking the first area that passes filters is not optimization. OPTIMIZATION requires some kind of a SCORE to compare all areas and select the OPTIMAL area. 

Picking the most optimal area that also passes filters is the goal.

Probably requires much testing to find a decent solution, so it may be a good idea to leave the "area scoring" to be defined in a snipet as well.

https://roadmap.strategyquant.com/tasks/sq4_9484

thanks.

Since we have build 136 stable version right now, this would be a great feature to add to the next build,

Simple as described!, again..:

As with SeqOpt but this method will optimize ALL the parameters of the strategy at once, (not each parameter of the strategy sequentially as with-in the current SeqOpt)

and then pick the stable region..,

Very simple and to the point, imagine SeqOpt but rather than optimizing all of the parameters one by one, one after another,

we will just optimize all the combinations at once.

Attached a screenshot of my vision of this feature.

Thanks.