I think preference scores are a fundamentally good starting point, but I don't personally put that much weight into them. I mean, if you have a speaker with rating of 6 and another rating of 5, I think it would then come down to personal taste. If you are comparing that same speaker rated as a 6 to something that rates a 3 or 4, I would say chances are that you would prefer the higher rated speaker.
I also don't think preference scores hold too well with something like RAAL Ribbons. There is a character to these ribbons with regards to attack and delay that would not be captured in the PS.