Solid states more accurate than tubes?

SS has higher 7th and 9th harmonic distortion than tubes and that is not as natural sounding to the ear!

Tubes have higher 2nd and 3rd harmonics, however that is generally more pleasant, so there is a subjective sonic trade off between tubes and SS.

I prefer beauty over accuracy.

When you look at your wife or GF after she rolls out of bed in the morning that is accuracy.... when she is made up for a nite on the town.....that is beauty.

Some SS has the beauty of tubes and some tubes do not have beauty.

I prefer natural to made up.

As the amp output impedance goes up, the frequency output tends to track the speaker impedance more.  That is, the frequency output looks like the impedance curve, instead of the original curve of the speaker.

As a result of this, a tube amp can be "tuned" by the speaker load.

Sometimes this is good, sometimes it is bad.

The OP is slapping paint with a 12 inch wide brush. 

Yes, ... most tube amps have higher output impedances that SS amps.  My amp, an ARC Ref 150 SE, has an output impedance of roughly .76 ohms off the 8 ohm taps.   By contrast, most SS amps have output impedances that are much, much lower, ... some close to zero.

And as other posters have noted, most speakers were voiced to be driven by a low output impedance SS amp.  So, if a tube amp has a high'ish output impedance, ... yes, ... a speaker that was voiced to be driven by a SS amp, will not produce a linear FR if driven by a high'ish output impedance tube amp. 

In the case of my ARC Ref 150 SE, I calculated that the FR variation between the low impedance point of my speakers (say 4 ohms in the bass region) and the high point (say 23 ohms at the 2.2K Hz mid-to-tweeter x-over point) is roughly 1.6 db. It's an Ohm's Law thing.  

But that does not end the story, ...ergo my point about the 12 inch brush.  As don_c55 notes, a particular tube amp/speaker combo may just sound wonderful.  Just because ...

The OP should check Ralph Karsten's (Atmasphere's) posts over the years.  He speaks at great length about why some tube amps sound so good, despite being "tone controls."

Just my humble opinion.

BIF

this is such a complex question; beginning with "is accuracy always what we want?" and "on what distortion weighting algorithm do you define accuracy?"
bell Labs and H.264/MPEG both championed subjective testing to optimize trade offs. Neither are exactly audiofools as the word goes....

Music theory tells us that there are distortions that are consonant and others that are dissonant. The finest pianos and violins have significant consonant distortion, from their cases and sounding boards. Is that a bad thing? Down with Steinway and Boesendorfer?

Obviously there is a wide spectrum within each camp as well. Old CJ stuff was lush but wildly inaccurate. I suspect, but cannot prove, that some SS gear is in many ways accurate but generates enough dissonant distortion to be no enjoyable (on some music anyway).

And then there is the SS gear explicitly designs to minimize the dissonant distortions ("SS sound") - whether you buy that logic or not. I actually do but warn that its really hard since we rarely know everything to measure. Wish I did.
G