A few thoughts on comparisons, in general.
I certainly can understand how Chris feels, as people may make decisions and more specifically decisions not to buy his gear based on a less-than-ideal comp. Then again, people will often do dumb things, and people will use poor judgement in purchasing decisions, what can you do?
The thing is though, I don't think an imperfect comp is useless and you can only rely on perfect comps. Sure, if we could design and implement one perfect comp, that is all you need. Trouble is, even then, people might make choices on gear when it was actually the source material that was the problem. Meaning, they really didn't like the sound of the source material, and the amplifier which amplified it perfectly sounded bad to them and the amplifier that had an imperfect frequency response actually managed to correct the bad source material sound and make it sound better to that listener. Does that make the imperfect response amplifier a better amp?
To me, all comps should be taken with a grain of salt. At the same time, it is good to have comps, and 1000 imperfect comps can be very useful. So any given comp is not perfect, if on every single one, the high end sounds distorted, it gets easier to associate that distorted high end with the piece of gear (say a mic preamp), if all other variables keep changing. Location, sound material, different mics used, different venues, different soundmen, whatever. If in 1000 trials you keep getting the result that Preamp X has a harsh high end, it statistically is very unlikely that every one of those flawed tests skewed in the direction of making Preamp X's high end sound harsh, when the only variable that stays constant is the choice of that particular Preamp X. It becomes pretty apparent (likely) that the harsh high end is just coming from Preamp X.
So to me, even imperfect comps have value. People around here always say to listen to gear on the Live Music Archive to see if you like it. To me, listening to one show on the Archive tells me nothing, especially since I have no idea how it actually sounded in the venue. But, say I'm interested in a particular mic, if I can listen to a couple dozen shows with that mic fed into a Lunatec V3 which I've owned and operated for years, and for those same shows I can also listen to a mic I know well (like the km140/km184, AKG 480/ck61, Milab Vm44) also fed into a V3, I can start to get a sense of the mic I'm considering. Not perfect, and not all that useful with 1 show, but if I can hear those admittedly horrible and completely unscientific and useless comps, and I can do it 20 or 30 times -- if I start hearing the same pattern, I get more comfortable associating it with the mic and not one of the other variables.
And on the way we tend to do comps here: people simply feel more comfortable hearing comps in environments and usage patterns that reflect how they will use the gear. Taping a stereo in your living room just doesn't feel right, even though it might be a much more valid way of running a comp. Also, on that end, it may be that there is good reason for running comps the way you want to run the gear -- say recording PA-amplified music. If I test my preamp using my home stereo, I might be using high levels of gain, when in a concert setting I'm using much lower levels of gain. An amplifier circuit might sound different or have different levels of noise or distortion at very different gain levels. If I can't reproduce those sound levels in my living room, the comp isn't necessarily helping me. Also, if you have a preamp with transformers, if you expect to be running the preamp at levels where the transformers are saturating, say at the low end (and if you want that saturation effect as part of your recordings), it could be difficult to get to the needed sound pressure levels to get into saturation mode when recording your stereo in your living room.
Which all said, might mean that the "perfect comp" that someone might put out there as the model to follow, does not get at the aspects of the gear that someone is interested in. So it might be perfect scientifically, but it is useless to that person.
Bottom line -- to me, lots and lots of imperfect comps are better than none at all, and are probably even better than "the one perfect comp".