Thursday, May 25, 2017

FTP Testing - Coggan or Friel? Is the 5% discount too much?

Dr. Justin Choo on the Dragon's ITT course, North Lantau

Functional Threshold Power (FTP) is defined as the average power that you can maintain on the bike for a one-hour, all out, race effort.

Knowing your FTP allows you to set zones that form the basis for your training. There are many different protocols used to test for FTP. Two of the most widely used methods are the Coggan test, and the Friel test.

Presented in his book Training and Racing with a Power Meter (widely considered to be the power training bible), the Coggan Protocol is essentially as follows:
  1. 20 minutes easy warm up 
  2. 3 x 1-minute wind ups with a minute rest between (100 RPM pedal cadence)
  3. 5 minutes easy
  4. 5 minutes all out (hard at first, but not so hard that you can't complete the effort)
  5. 10 minutes easy
  6. 20-minute time trial effort (this is the test - like the previous 5-minute all out effort, keep in control, hard but steady, you don't want to over cook it and die at the end)
  7. 10 to 15 minute cool down
You then take your average power over the 20-minute hard effort, and subract 5% to get an estimate of your FTP.

Another popular and much simpler test is what I've always referred to as the Friel Protocol. It is:
  1. Warm up
  2. 30-minute time trial (this is the test) all by yourself (no training partners and not in a race - it should be done as if it was a race for the entire 30 minutes)
  3. 10 minutes into the test click the lap button on your heart rate / GPS device 
  4. When done look to see what your average heart rate was for the last 20 minutes - that number is an approximation of your Lactate Threshhold Heart Rate (LTHR)
  5. When done, look to see what your average power was for the entire 30 minutes - that number is an approximation of your FTP
It is elegant in that it yields an FTP in 30 minutes with no correction. And in the same test you also get you get a nice LTHR estimate from the last 20 minutes, and again no correction is applied. Essentially you are chopping off the first 10 minutes in which the heart rate climbs more rapidly up to homeostasis, and then settles to climb more gradually around the LTHR, around the last 20 minutes (see pic below, with the initial heart rate ramp up highlighted).

FTP Test, with initial heart rate ramp up highlighted

The main differences between Friel and Coggan then are:
  • Duration of test - the Friel test uses average power over 30-minutes, the Coggan test over 20-minutes.
  • Discount to calculate FTP - the Coggan average power is discounted by 5%, there is no correction on the Friel test.
  • Warm up - the Coggan method tires you out with a hard warm up, the Friel warm up is easier (but the test is longer).
Friel has argued that no correction is needed because when you do a test alone, you feel sorry for yourself, and without the adrenaline of a race situation to push you, you simply can't push yourself hard enough.

And I totally agree with him!

I propose that the 5% correction factor is excessive for most non-elite athletes.  Age group athletes just don't have the chronic tolerance for pain that elite and professional athletes do. These people, as part of their job, need to habitually push themselves to levels that they race at. The have formed a mental discipline and self-knowledge that allows them to dig deep and simulate a race situation within themselves. Psychologist and researcher Sian Beilock describes this ability well in her book Choke. Being able to simulate a race situation in your head, visualize all the accouterments of the competitive atmosphere, and then be able to push your body to suffer as though you were in it, is a honed skill. And it's a skill that pro and elite athletes usually possess, but that many age groupers do not!

Thus, I think the 5% correction is excessive for non-elite athletes, and a better estimate of FTP for age groupers is the Friel estimate, or the Coggan estimate, with the hard warm up but without the 5% reduction.

I've seen this for my own performances and those of nearly all my power meter customers. The long race time trials yield 1-hour average power levels within 1-2 watts of the predicted ones without correction. And I suspect this would be the case for most age groupers. I can back this up with numerous race time trial results and step test lab test results from the Institute of Human Performance at HKU from me and my customers!

FTP ramp test at HKU
One very negative repercussion of this then is that if we are discounting the FTP test number by 5%, then we are setting our training zones too low, and thus not getting the full effect that we should be getting from our training.

So with that I was surprised while reading Joe Friel's latest book, Fast After 50, that the FTP test he is now recommending is the Coggan test and not the 30 min test that most of us refer to as the "Friel" test!  In all his previous books (there are many!) and on his blog, he has always advocated the 30-minute test. So I asked him in an email about the change. His reply:
Hi Tony, 
Thanks for your note.  Good insights. I have used both methods (and more) but the Coggan 20-min test has become the standard, it seems. Do whatever works best for you and your clients. There is no right or wrong here. I also tell coaches to trust themselves when it comes to making an opinionated change in FTP based on recent training data. FTP changes a bit on a daily basis due to fatigue and other lifestyle factors not to mention psychological variables also. It even changes during a ride. Weather can affect it. So there are many variables. What we are aiming for is to get a ballpark number—albeit a small ballpark.
Good luck! 
Fair enough. I could find no research to validate the use of the 5% Coggan discount, and none that considers that it might be different for non-elite athletes.

So until I find some validation, I won't be discounting my own FTP test numbers when setting my power training zones, and I don't recommend that my customers do either, unless, that is, they are elite athletes.


  1. I think you're right: for the typical amateur, taking 95% of a 20-minute trial is likely to yield training zones that are a few percent too low.

    A point worth noting is that Coggan himself doesn't recommend the 20-minute test.

    Since lactate threshold itself is only a vague, approximate concept, perhaps we shouldn't expect too much precision in calculating FTP. For practical purposes, the key point might be for individual athletes to get a feel for the difference between zone 4 and zone 5 training intensities. That can be developed by starting with a 30' trial, as you suggest, and then experimenting with a range of typical zone 4 and zone 5 interval efforts, such as 4 x 12', 2 x 20', or 2 x 30' and 5 x 3' or 3 x 5'.

  2. Good points. Yes, the important thing is just setting a baseline and setting zones from that. And that's how Carmichael gets away with an 8 minute test!

  3. Hello,

    A few comments:

    1) It's really Hunter's test protocol and correction factor, not mine;

    2) As I mentioned in the chapter on power-based training that I wrote for USA Cycling 15 y ago, based on the scientific literature the difference between 20 min power and FTP (actually, 10 and 25 mi TT power) would, *on average*, be closer to 7%;

    3) that said, in a MUCH larger number of subjects there is close agreement (again, on average) between 95% of maximal 20 min power and FTP as determined using the WKO4 power-duration model (see my short article about the model on the Trainingpeaks website); and finally

    4) the exact relationship between 20 min power and FTP can and does vary somewhat between individuals, but this more closely related to inherent strengths/weaknesses and how somebody trains, and not really to elite/non-elite status per session.

  4. One further comment: there is no magic to training at very tightly defined intensities, which means there is really no point in worrying about a couple percent one way or the other. Indeed, that is partially why I chose to term my original power-based intensity guidelines "levels" and *not* "zones."