[seq opt] Make sequential optimizer actually optimize for highest scoring AREA and choose the center point

Currently, it doesn't actually optimize to find the best area, it just optimizes to find the best settings that give the highest fitness and then checks the area to see if it "passes" the stable area filter.

In doing manual heat map style optimizations, it's best to find the best area using the heat map and use the settings from the center of the area. That's the whole point of the heat map style manual optimization, you can check Rob Pardo books for more info.  So, it would be better do this automatically. I suggest we pick the "best/most stable" area by scoring each area and then select the center of the highest scoring area and use the settings from the center of the highest scoring area.

And furthermore the current one-size-fits-all-filter system is problematic as some variables are way more sensitive and or have higher variance than others and we still want to pick the best area in relationship to other areas for each variable. An optimizer should be for optimizing first and foremost. 

In order to score the area of best fitness we need a scoring method.  I'd like to hear more ideas for scoring the areas but here are mine so far:

-Best average fitness of the entire area
-Best lowest fitness (find the lowest individual fitness of an area and that will be that areas score, if there is a tie, the tiebreaker can be the best average fitness of the entire area)
-Allow scripting our own selection/scoring mechanism of the area
-Median fitness (agentpot)
-Sum(gazorpazorpius)
(examples and screens bellow...)