Amir,
A good hearing model will tell you statistically what most people would prefer, from a distortion perspective, and if that is then put into practice correctly with an amp design, it should guarantee that at least a fairly large number of people will find that amp excellent sounding. This is the case with Lamm, where most people who have heard the products think they are excellent sounding. It doesn't mean everyone will agree. Human psychology also guarantees that a good correlation might be 50-60% agreement to the model. That leaves room for a significant amount of naysayers. So, please don't compare this to the proof or absence of proof to the existence of God.
Finally, as a student of the philosophy of science (I was a chemistry major and philosophy minor in college...later Ph.D in Analytical Chemistry), Popper came up with the now adopted concept of hypothesis falsifiability. He didn't believe anything was provable...only falsifiable. You can have 1000 positive data points on a hypothesis and it only takes one negative to deem a hypothesis invalid. Nothing is ultimately provable but the best hypotheses are those that withstand efforts to invalidate them.
In his view, it wasn't science if you couldn't formulate a hypothesis that was testable and falsifiable. It doesn't mean the thing you espouse is necessarily false...it simply can't be explored by science. Human hearing and determining preference does not fall into that category, it is testable and at least potentially falsifiable.
Psychology is probably somewhere between science and non-science because there is a lack of precision in the test methods and responses that yield correlations, which at least by 'hard' science are weak, leave a lot of room for people to think differently than those that correlate well with the hypothesis or model.
So, what this means is that, if the story of Lamm using a model to design his amps is true, he was in a way kind of a pioneer in trying to merge the human psychology of hearing with circuits that deliver the best sound based on the model. This doesn't mean they will be universally hailed as the best ever but it does mean that they are likely to sound better than most to a large percentage of audiophiles.