How much can be measured -- and how much cannot?

"The number of times that I and many others have referred to John Atkinson's measurements in helping Audiogon members to make purchase decisions or to diagnose problems is practically countless."
Actually I have never found measurements to help me in any audio decision to by a certain product. Measurements are helpful in the design process but do not tell you anything how a component will sound. Use your ears and they will guide you on the right path

If we spent more time listening than thinking we would be better off.

And I will add that learning to read and understand measurements will provide a great education to an audiophile. When I compare the specs I looked at originally compared to the factors that I try to take into account now, I realize how much I have learned from this hobby.

Off the top of my head, factors that I now take into account:
output voltage
output impedance
input impedance
input sensitivity
gain
harmonics distortion (2nd, 3rd order, etc)
damping factor
speaker distortion (hard to find in general)
crossover point
crossover slope (1st, 2nd, 4th order etc.)

I'd also add to that list the standard plots vs. frequency graphs that one sees in Stereophile's measurements section.

And when I see many of Al and others' posts, I realize that there is much more to learn.

BUT all of that having been said, I too ultimately come down on the side of letting your ears make the final decision. Psychoacoustics are pretty wild, can be tough to get your head around, and aspects of them are difficult/impossible to measure. Heck there are very likely aspects to psychoacoustics that remain to be identified. But trust your ears, they are excellent psychoacoustic instruments connected to all the bits of the brain that matter for a satisfying experience of music.

>But others say there are things in high end audio that will never me measurable and that measurements are really not that important.

Speaker preference has a very strong correlation (if you weren't a scientist you'd say it was causal) with with uniformity of amplitude response curves at angles representative of a direct listening window and where front-wall, side-wall, and floor/ceiling reflections come from in average rooms.

Sean Olive has reportedly taken that to the next level producing a formula that predicts peoples' speaker ratings based on such measurements.

Stored energy is important too, although that shows up in polar response curves (ripples in the on-axis response can come from diffraction effects that aren't that audible, but ripples at many angles tend to be resonances which are) with fine enough frequency resolution.

>How important do you think measurements are?

Very. Good enough to predict that you're not going to like a speaker before you go to the trouble of listening.

>Are the ears really the only true arbiter?

If you can hear it you can measure it and if you can measure it you can fix it or decide it's not worth the price tag which goes with the fix.

You'd do well to read _Sound Reproduction: Loudspeakers and Rooms_ by Floyd Toole.