Interesting post / thought - I doubt anybody would view me as racing below where I should be, I always have to try 
Just to be clear: I am talking about pace group, not rider category based on ZRS.
Well, I was demoted and have since achieved new 2 and 5 min PBs but remain in pace group C. My racing score has however continued to rise, although not by the level I might expect based on how I have seen other riders’ scores increase.
Not sure what you mean by category enforcement as that only applies if racing, at least that’s my understanding. But yes I assume the FTP drop is the reason as in 230w puts me above 3.36 w/kg while my current 226w puts me below 3.36 w/kg.
I still find it very strange as 5 min, 20 min and average watts have all gone up; I am certain my energy output (using a variety of measures) is higher and more consistent than when I qualified for group B.
When I say Category Enforcement I’m talking about the pre-racing-score system that combined both the pace group and pen enforcement of that. Under that system, I have done 3.48W/kg 20 min and ended up at 3.35W/kg zFTP, remaining in C. It’s possible to stay in C with slightly higher numbers but it gets increasingly rare. It has to be combined with high zMAP but not too high, because at 4.1W/kg zMAP you are B, but with lower zMAP your zFTP goes up.
Yup, me too
Power curve modelling (like zFTP) tries to separate aerobic from anaerobic power at a given duration. Since it knows your anaerobic engine to be more powerful it must mean your aerobic contribution at a given duration is less. There’s actually yet another energy system (ultrashort pmax, phosphocreatine, short, frc, anaerobic glycolysis, ftp aerobic glycolys. So pmax goes up, frc gets modelled down, ftp gets modelled up (shorter durations only affect modelling of longer durations, not vice versa)
(For what it is worth wko5, the gold standard for power curve modelling models my FTP at 299W, intervals.icu at 300W and zFTP at 302W but obviously zFTP is the most correct)