they performed a 20-minute time trial and 95% of the mean power output (P20) was defined as FTP.
Still? After 20 years? Boy, it must suck to be Coggan.
And when someone finally talks about how to actually test ftp it becomes X’s protocol(!)
I’ve been thinking of doing an Aerotune Power Test to see what that yields. It’s a bit like the. iNSYD one but CONSIDERABLY cheaper.
https://www.instagram.com/p/CaVErO5s7Ad/?utm_source=ig_web_copy_link
Please share results. Test protocol (indoor or outdoor, 1 day or 2 days…). And if you get specific training plan/recommendations (those recommendations listed on the last side cover multiple blocks (?)).
(If testing indoor and you have power from trainer and from power meter would be interesting to compare both results (for example if trainer reads sprint lower)).
I do this test at home and the sprint is tricky to get right. Part of the problem is that some power meters are too slow to respond, so your first couple seconds of power won’t be representative of the effort. For example, the powertap hub provides much better sprint data than a P2M NGEco despite both recording at 1-second intervals. A power meter with magnets and reed switches reading at several crank positions would be best. Thats why certain power meters are required for track cycling.
The calculation is not too sensitive to VLaMax, so it’s ok if the sprint isn’t 100% perfect.
Ditto - been quite interested in this and would love to know if it yields usable training recommendations. I’d love to know what you get out of it please.
However, I have heard several far smarter people than me suggest these tests are both simply algorythms that produce data that is far from scientifically proven. The Aerotune test is cheap enough to take a punt on but I’d really like to know what actual recommendations come from it and how they may be used.
This is what the creator said in a podcast. He said “Don’t worry if your sprint is 600w”
It’s just their flavor of the CP model.
Indoor is a ramp test, outdoor is a 4 min max 
Both tests have sprint and 12min max
Ramp test and 4 minute max power both provide approximations of VO2Max. It’s hard to do a ramp test outside, so I think that’s why they use 4 minute test outdoors.
The sprint is for VLaMax and the 12 minute effort is to provide another point for CP evaluation.
I wonder if the model gets better data with the ramp
I suppose whichever model estimates closest to your actual abilities would be best.
Here’s something…check out page 10. It’s one equation to estimate VO2Max from 4-minute power. I wish I could read German. I have several papers and dissertations that I can only skim through.
So sprint then 4min/ramp then 12min?
I think I heard you could split some efforts over 2 days…or I remember wrong?
Yes, see the workout graphic I posted above…They added the one day execution, you still have the option of 2 days.
I’ve seen the same with 6 min. The problem is the bias of those coefficients. I guess nobody needs to be too precise with this stuff.
It says that ramp test measured VO2max correlates r=0,95 with 4min max VO2max calculation.
Are aerotune calculations based on that paper? It sais crossing point (MLSS) deviation up to 20watt and fatmax deviation up to 40 watts…
2 days preferred?
If 2 days…which effort which day?
Sprint + 12 min day 1? 4min day 2? Change days?
You don’t necessarily need precision if you’re doing the same test over and over. Relative differences become more meaningful than absolute values. Since metabolic carts are quite expensive, we use the tools we have available.
Using both of the equations from that paper, I get the same ballpark for VO2Max whether it’s from 4-minute max or last minute of the ramp test. They’re decent approximations in my case.
I honestly don’t know what Aerotune is based off of. The service was strikingly similar to Inscyd when I signed up a few months ago (I never used the service, I just looked).
The VO2Max and VLaMax approximations can be pretty “off”. I think I read that Weber said VLaMax from power can be off by +/-0.25 mmol/l/s. That will give a wide range if you want to do upper- and lower-bound estimates.