20min vs 30min FTP test

Two questions, but first an explanation:
Previously, for some reason, I thought FTP tests were supposed to be 30mins. (I think the idea came from the Joe Friel Power Meter book) Then taking 5% off to get the hour value. That is what I have done in the past. Breaking the 30 mins down into 5 minute intervals to monitor evenness and progression. (I do a lot of 10mile TTs so it is also a more natural and real world test for me).

When I started TR (4 weeks ago) did a ramp test and that came out at 235. About right so I stuck with that.

Tuesday I did the TR 20min FTP test, but actually started the session in that initial 5 minute loading, and
bridged across the 5 minute rest in resistance mode, so that i did a whole 30 minutes in resistance mode. I averaged 249w for over the whole 30mins and 251W over 20 mins. This gave me an FTP of 238. (The 5 minutes were 244. 246, 246, 248, 255, 255W)

Question 1: I assume the 30min value is a better test. is that right?
Question 2: Does the fact that I did very similar values every 5 mins suggest I could be higher for 20 mins?
Question 3: I assume I am best taking the latter 20 minutes when the average was higher
Question 3: Is a 5% reduction used for the 20min test. Is 5% off the 30 min still one appropriate.

Thoughts please, thanks in advance

Allen and Coggan suggested that 5% less of a 20 minute average is a more reasonable way to estimate FTP than a straight 60 minute effort for various reasons. They go onto explain a power profile protocol to find limiters and strengths. Perhaps that’s where you’re getting confused and I’m not aware of a Friel 30 minute test.

Anyways, I think the most important point in all this FTP modeling is we are doing it to effectively work all three systems (ATP-PCr, Glycolitic, Oxidative). Having a test that is easy to repeat yielding consistent results over time is the point. Are you improving, flat or declining? Basically, a validation that what you’re doing is working.

I think Coggan has actually denied the 20min * 0.95 calculation has anything to do with him, although on the other hand it didn’t stop him putting his name on the book.


Second, I have never suggested that people use 95% of their maximal 20 min power to estimate FTP. Quite frankly, I am surprised that anyone with a doctorate (in any field) would make such adamant claims without thoroughly researching the topic they are discussing.

Which is from a comment on this article: http://marktallonphd.com/the-myth-of-functional-threshold-power-ftp/



Correct. From TP: “FTP is an individual’s power output in watts, or work rate in kilojoules per second, at MLSS. It can be held for approximately 30 to 70 minutes (a 30- to 70-minute TTE) at a blood lactate concentration of 2-8 mmol1.”

How to figure that w/o a lab is beyond me and not all that realistic for the masses. So, 20*.95 and/or TP ramp test will have to do.

1 Like

The Friel 30 min test is on page 55-57 of his book. Though he does do it in real world. A turbo seems just as good I would have thought.

I also got the impression reading both books, that Friel’s book was making Coggan’s a bit more readable. (Both seem to take 4 pages to explain a concept that could have been explained on a page. Coggan perhaps takes 6).

I fully appreciate the wider power profile piece. Sufferfest, WKO4 or whatever make a feature of the 4 part test. Coggan looks at the profile across the power bands. I also saw this Training peaks article https://www.trainingpeaks.com/blog/the-physiology-of-ftp-and-new-testing-protocols/

I understand the problem of a single figure test very well. And recognize the need to work on the whole power curve, depending on event type.

I might do a ramp test and a 20-30 min test. Both are useful to me. It just seemed that the 30 min test seemed a better indicator for what i was doing.

A catch with both is that you learn how to do them. Or rather the more you do them the better you are at learning how to do them better/more reliably.

I guess the key is to recognise that it is a useful benchmark. But not teh only one, and to adjust training load accordingly. the 30 minute test is a great indicator of the ability to hold a power for 30 minutes though :):grin:

Keep in mind some of the goals of knowing your FTP - 1. to guide training, and 2. for pacing during a race.

Your FTPs calculated using three different methodologies are all within 3W of each other, so not sure it’s going to matter which you pick, as they are all within a margin of error or variability you might expect.

If your FTPs from the different methodologies were more different, I’d say use the 20 or 30 minute one, as these will be more helpful for the pacing in your 10mile TTs.

In guiding training, the important thing is to start with an FTP that is close, and then you adjust the training over time anyway - up/down if too easy/hard, so you probably could use any of the FTP numbers as an appropriate starting point.

Whatever one you pick, just consistently use that approach going forward for FTP tests.


Thanks Dave - makes enormous sense.

For fun (and also because I have an FTP test coming next week), I’m going to do a Ramp Test and one of these TTE tests and see what WKO4 thinks as well. The TTE tests are supposed to be much more accurate and easier to execute compared to the 20 minute test even though they are a bit longer.

I’ll report back with what the Ramp Test said and how that tracked to the actual numbers I could hold for 35+ minutes.


I have WKO4+ and read that article as well. Maybe I missed it… I don’t understand the rate of progression in the Baseline and method 1? Is it like how TR gradually increases power say during an over-under?

EDIT: asking as I’d like to compare as well…

It’s self paced. The first ten minutes is supposed to be solidly under threshold to make sure you don’t go out too hard and can gradually bring yourself up and feel where FTP is.

The progressions assume that as you move up the progressions you are able to ride more accurately and self pace better, hence the higher percentages of FTP for longer time, etc.


Ok Super thx! If nothing else it will be an excellent real world (on the bike) race estimate I believe.