Agreed. Maybe add directivity (wide x narrow), dynamic compression (totally agree with Erin that this should be more relevant) as variables and that's pretty much it. But the main criteria, flat LW and low bass extension, was already discovered to be highly significant. The bulk is already there, more variables would be fine tuning. Perhaps a case of diminishing returns in research?I don't think money is the only reason.
Score can be sum up as: go as low as you can with a flat LW. I can back that up, one way to see it: when you create an EQ to optimise the score, you get 95% of the improvement by flattening the LW. You can get more if you want to improve the bass but you loose on SPL. the other parameters don't play much role. Also look at @malky68 EQ and look at the difference between LW optim and Score optim.
I think it is unlikely that you find something magical. Maybe distorsion is also a large variable in the midrange. We could find a better predictor but I am betting it will start with flat LW and low F3 and distorsion < 1% to define the usable SPL.
Would be great if @Sean Olive himself could give his opinion.