"Buying a product using a trial and error method is not really 21st century standard, is it? But that's what we've got."
There is only trial and error.
Several different groups create computer models of the atmosphere. Each group is sure they have it right. The one that predicts the weather most accurately is chosen. It's trial and error - even for "21st century" computer modelling.
If there was only trial and error, then the scope of science would be limited accordingly and there would be no need for engineering as a field of endeavour.
My views on the value of models that can be used to guide consumer choice has already been made:
http://yabb.jriver.com/interact/index.php?topic=77129.msg523173#msg523173To use your weather reference, meteorologists don't stop using models because they're wrong sometimes. They seek to understand the errors and improve it where they can. If they didn't, we'd be stuck with a wetted finger in the air and poring over animal entrails to divine the weather.
Ultimately, in my view, some kind of tool based on scientific principles is better than nothing.
However, as I've said before, any measure can be corrupted for marketing purposes so the end consumer needs to beware. That said, I do find it reassuring that a number of apparently independent sources comment on the strong correlation between the DR measure and their own aural perception and preferences. Each to their own.
I'll stick with the group who value the DR database. I'll continue to use it as a guide when I purchase music and avoid relying solely on trial and error.