I have to say that my subs are doing a better job than my towers. Well calibrated, they punch very cleanly and very powerfully.
Like the Audioholic's BM article state, bass needs a lot of power to be reproduce, so why don't let your powered sub do it and leave more power to your AVR for mids and highs? Specially if you have 2 powerfull SW?
The problem is a decent speaker competently designed should be allowed to perform the way the designer intended. The bass tuning is an important part of any speakers voicing.
So if you have a speaker that rolls off acoustically at 40 Hz say, then rolling it off at 80 Hz is making nonsense of the designers intent. So if you then supplement the last octave with a sub as the speaker rolls off, then you have minimal violence done in terms of phase etc. You always want to use acoustic crossovers as much as possible.
Now we get this myth again that bass takes all this energy to reproduce. It doesn't. There is very little acoustical energy below 60 Hz, I see that all the time on my energy spectrum meter.
The reason you think it takes a lot of energy is that subs are very insensitive and inefficient because they are too small. If they had the same efficiency and sensitivity as the mains, they would take very little power. My bass amps don't break a sweat, not even in HT.
The sub Matt is building will only need about one or two watts and threaten to destroy the house, but its not small.
The acoustical energy is really concentrated between 80 and 1500 Hz. That is where you need power handling capability. That is where the acoustic energy is that is radiated into the room. That is why small pipsqueak mains and centers are not the way to go, for really realistic reproduction.