super(L.tsq("R Expectancy Score"), DatabankColumn.Decimal2, ValueTypes.Maximize, 0, -5, 5);
These bounds are too tight for the R Expectancy Score and (possibly other columns) during optimization as RES can be nearing 500 with a large number of trades while optimizing. Although I can see the bounds are maybe almost reasonable for a typical build task, they are not really reasonable for an optimization task. It's very touchy or even impossible to attempt to rectify this by increasing the weight multiplier when nearing or at the absolute upper bounds (where a full max score is already given or almost given for individual criteria.)
May I suggest having the upper and lower soft bounds be configurable inside the weighted fitness table instead of coded into the snippet like this? Also, maybe we can get "fitness criteria #1 fitness" and "fitness criteria #2 fitness" as columns or a single column with a text CSV breakdown of the criteria calcs so we can see when upper bounds are nearing of particular criteria. This took a long long time for me to figure out why RES wasn't working well inside a multi stat fitness for optimizations. Probably other columns are affected the same way.
BTW I want to control precisely these soft bounds for MINIMIZE/MAXIMIZE and it would be nice not to have to create a new column snippet every time I want to adjust these. I do not wish to use APPROXIMATE as this calculation actually penalizes for going OVER the target amount where MAXIMIZE would not.