Curiosity got the best of me, and to spice things up a bit on the indoor trainer I thought I'd give those "TMax Intervals" Paul Laursen did that study on that everyone was probably quoting and referencing a few years ago. So I started with the process of doing the peak-power output test, which looks much like Ramp Tests you see administered by national team programs and various coaches (alas, I really can't afford a coach myself at the moment).

For those unfamiliar with these:
-Do a ramp test starting at low wattage, increasing by 25w every minute. Go until you're unable to do anymore. Calculate the average power over the last 60-seconds to determine your peak power output (PPO).
-Rest for a day
-Then, ride at your PPO for as long as you're able to, which from what you see in all the stories and in the original paper itself should be "4 to 6 minutes."
-Suppose you hit 5 minutes at your PPO. You take 60% of that, which in this case would be 3min. You do 3min long intervals, and rest for 6min (double the work session). Repeat up to 8 times.

So, my results were...
I got up to 375w on the trainer for the MAP test going into the last minute, which would be 400w. At that point, I tried to keep it at 400 but I wound up going pretty hard--as in, not controlled, just gunning it basically--and averaged 415 over the last 60 seconds. It was...very hard.

Then, after a rest day, I got back on the trainer and only managed to hold 415w for 2.5 minutes! So that would mean, by the standards of this protocol, doing work sessions that are 1.5min long, on 3min rest.

Does anyone have any experience with this? Or can anyone help me figure out why exactly I can hit over 400w in a ramp test, but can't hold that wattage itself for too terribly long? :-) Thanks so much!

Two things to consider: everyone is different, and you may lack anaerobic training time.

Different interval lengths have different effects. The more usual thing is to find FTP then do intervals of a specific length at a specific percentage of FTP. So the first thing you want is to decide what you're trying to improve or optimize and then decide what interval protocol is most likely to accomplish that. It seems to me that you're going about it bassackwards.

Rather than start with a number and then use it for intervals, my usual method is to figure what interval length I want to do and how many reps, then find by experience what my output should be to be able to do the reps and then fail to achieve that output on the next attempted rep. Hopefully that output goes up after a few sessions of work and a recovery period.

I realize that many people put a lot of stock in their numbers. I'm more inclined to do the work and have less formal testing. I can tell pretty easily if I'm getting faster. But that's just me.

The Peak Power Output (PPO) measured is dependent on the test protocol. A faster ramp of 25W/min which you did will result in a higher PPO than a slower ramp of 30W/3min.

Do you have the name of the original paper?