I was referring to the power zones he gives in his "bible". Apparently
he had only been using a PowerTap for a few months before he completed
the book, and didn't have sufficient experience with power measurement
to have a good feel for what people can and cannot do. Trying to work
"backwards" from HR, he came up with zones that are way, way too
high...for example, IIRC in the book he labels anything less than 80% of
"threshold" power as being recovery, when in fact you have to be riding
pretty damn hard to get your average power up that high. Indeed, as a
general rule I don't think it is possible to come up with precise power
zones or levels by working backwards from HR, simply because of
variability in HR not only within, but between, individuals.

My criticism of the "critical power paradigm". at least as he
implements it, is his use of subscripts or numbers to denote the
exercise duration, e.g., "CP12" being the maximum power you can sustain
for 12 min. The problem with this approach is that it implies that you
have multiple critical powers, when in fact you really have only one.
That is, the critical power concept is well-developed in the scientific
literature, and it has been shown that beyond the first couple of
seconds of exercise, the entire power-duration curve can be described
using just two parameter, one representing/correlating closely with
anaerobic capacity, and one representing/correlating closely with
lactate threshold. The advantage of presenting things this way, in
addition to just avoiding confusion from trying to reinvent the wheel,
is that it helps people better understand the physiology underlying
performance in various events. For example, presented the Friel way, one
might be led to believe that the factors determining performance in a
pursuit (power just above "CP6") and performance in a 40k (power just
above "CP60"...for most people) are significantly different. Presented
the established/accepted way, however, it is clear that both are largely
determined by lactate threshold, since one's "critical power" has a
significant influence on performance (power) even at the far left end of
the power-duration curve. From a training perspective, that makes it
easier for people to understand, e.g., why to ride a fast pursuit you
still need to put in significant volume*, or why even masters racers
whose events might be less than one hour can still benefit from
multi-hour training rides.

*Under Charlie Walsh, the Australian Institute of Sport pursuiters
would train up to 35,000 km/year, whereas according to a recent
interview with Bradley Wiggins on cyclingnews.com, the endurance track
riders in the British World Class Performance program (directed by Peter
Keen) would, at times, train up to 32-35 hours/week. All for an event
less than 5 min in duration!