That‘s like complaining that you can‘t put together your new IKEA shelf, but you refuse to read the manual. The importance of adapting your intensity/FTP to match your abilities or even your daily form is emphasized in the workout texts. It is a topic on many blog posts and has been covered many, many times in the podcast.
Even if you come from another training software or had a coach, you‘d be taught the same thing: there exists no simple formula to infer FTP from whatever test you use that holds for everyone, be it ramp test, a 20-minute test or an 8-minute test. You can‘t leave your brain at the door when training.
That‘s implying that TR is not basing their approach on evidence, which is evidently not the case. AFAIK you are using plans from Fascatcoaching (correct me if I am wrong). On one of their first podcasts their boss Frank explained this point quite clearly: in his mind, most coaches are (for various good reasons) ahead of the science, i. e. successful coaches often use approaches that are not scientifically validated — yet.
To me TR‘s AT is just such an approach: it wouldn‘t be possible to compile such a dataset for a scientific study, and it seems to me that they are very much evidence-based, simply because their approach requires them to be evidence-based. It doesn‘t mean their method works, they have to be careful not to simply reproduce their own biases, etc. etc. But to imply TR‘s approach isn‘t evidence-based isn‘t corroborated by the evidence 
Now that doesn‘t mean AT and TR’s training plans work or that they work better. But the numbers that they have released last week seem to suggest they improve on what TR had before: much higher completion rates, higher gains, etc. Yes, there are caveats like selection bias you need to consider, but that was made explicit in the podcast. That is, TR is aware of them. Would I prefer more openness when it comes to benefits and results? As a scientist and nerd, absolutely. But I understand why they are being cagey.
Coggan‘s definition of FTP is that it is an indirect measurement of your lactate threshold in a field test; he distinguishes between FTP and LT2, because they are obtained through different measurement protocols and thus, aren‘t directly comparable.
In AFAIK his very first work where he developed this, he already said that while a 40k TT is a good way to infer your LT2, but I think it is a misunderstanding to claim that “this is Coggan‘s definition of FTP”. Coggan‘s definition evolved over time and when you look at the scientific literature, you often see FTP20 or so to indicate the testing protocol they have used. The idea being that numbers obtained with different protocols aren‘t directly comparable. That was Coggan‘s impetus to distinguish between FTP and LT2 in the first place.
Coggan immediately recognized that a 40k TT is impractical for a number of reasons, one of them being the mental and physical fatigue incurred by a long, all-out effort. But another is that not necessarily all cyclists train for TTs, as a dedicated crit racer it might not make sense to focus on long TT-style efforts. That‘s how the 20-minute and later 8-minute tests were born. And eventually the ramp test. AFAIK many exercise scientists these days sidestep the issue by just measuring MAP and not bothering with FTP, because the ramp test is simpler and more consistent. I reckon consistency is an issue with less experienced cyclists who haven‘t done 20-minute FTP tests before, perhaps some of the gains are then due to them getting better at pacing and testing.
IMHO I don‘t see any clear winners or losers. If you find a testing methodology that you like/prefer, and from which you can determine “your FTP” (as in the base power level by which to scale your workouts), then stick to that and remain consistent. But all methods have error bars that are big enough to produce “FTPs” that do not quite scale your workouts right. That‘s simply because the adjustments you might need to make are small, comparable even to the accuracy of your power meter. Like I wrote before, recommendations for the FTP20 test range from subtracting 5–15 %. Is it worth discussing what the “correct number is”? Personally, I‘d focus on emphasizing that you should find out by yourself what the right number for you is.
