agreed.
in a 2003 paper which predates the Training and Racing book, he noted it is not a precise physiological response. And to avoid a false sense of precision in all methods of estimating. 
In any case he offered several practical ways of estimating FTP, for the express purpose of establishing power based training zones described in the paper.
While a 40km time trial is âprobably the easiest and most direct way of estimating a riderâs functional threshold powerâ the emphasis is on functional. That statement is backed by data, for all you data fans, but it is not functional and therefore not a good choice if you canât push or donât want to push threshold for such a long effort. And the word âdirectâ - as opposed to indirect methods like ramp tests or 3-min all-out critical power tests. Direct because riding at maximal lactate steady state is something well trained athletes can sustain for more than 30 minutes and less than 90 minutes. But hey, Iâve only attempted one or two handfuls of these efforts so Iâm not going to suggest it in general.
Which brings us to Cogganâs second suggestion for estimating ftp:
âAn even easier way of estimating an athleteâs threshold power is to just measure the power that they can routinely produce in training during long intervals or repeats aimed at raising LT (e.g., 2 x 20 min at level 4). Typically, this will be very close (within perhaps 5 percent) to what can be sustained during a 40k TT, with the shorter duration and recovery period(s) between efforts compensating for the generally lower motivational level in training vs. competition. ⌠The primary advantage to this approach is the ease of measurement, which in some cases may make it preferable to more formal testing.â
Coggan circa 2003!
(and when ramp test underestimated FTP that approach worked great for me on TrainerRoad)
Noted what is not a precise physiological response? FTP or Lactate Threshold? I am stating FTP is a proxy for finding LT, the former not being to the same accuracy/precision standards as the latter.
Anyways, for practical purposes, you want to be in a range near LT and there are many ways to get there.
1 Like
Lactate Threshold. Top of page 27 in the 3rd edition of Training and Racing with a Power Meter which concludes the lab determination of MLSS via blood lactate (an indirect marker) with â⌠more difficult than many people realize and the data obtained are not more accurate or precise than those obtained using much simpler field testsâŚâ For example I typed âlactate curve mlssâ and here is one of many papers that discuss the topic https://www.researchgate.net/publication/326693764_Manipulating_graded_exercise_test_variables_affects_the_validity_of_the_lactate_threshold_and_VO2peak (graded exercise test = ramp test).
And yes, you want FTP estimate to be in a range close to MLSS. I rely on two methods: triangulating using interval efforts around FTP, and when doing max efforts on Wed night worlds looking at power frequency distribution. I happen to get good triangulation data from Garmin sub-max efforts on my 530 computer (doesnât work for everyone), and if enough hard efforts from WKO/Xert while Intervals.icu requires occasional tweaking due to the shape of my power curve.
Thatâs cool to dig that out - nice.
I canât comment on what Coggan defines as acceptable accuracy or precision for the field test compared to blood test, nor do I know how the state of the art in measuring LT has progressed since that quote. I donât doubt that Coggan had it pretty well pinned with field testing and relating FTP to LT. But I do doubt that the broad community has it as well understood.
Definition of LT is fairly crisp, while the definition of FTP is not. And it gets more blurry with âFTPâ when anyone but Coggan is talking about it.
So back to what weâre getting at - for practical purposes you want to know where ones threshold zone is and circle out from that. And oneâs body is an excellent sensor for detecting threshold (and all other zones). Just like any sensor, the more samples you take with it, the better you can pin things down. Thatâs basically what youâre doing and itâs gonna be really hard to beat that approach with anything else.
1 Like
LT is not MLSS
Well, letâs see how things go 
Nateâs back! 
3 Likes
This is such an important point - TR FTP determination is so important to get the TR workouts at the right level. So many people get into knots because they want the TR FTP to be exactly mapping to their ârealâ FTP⌠whatever that is. I find it hard to believe that itâs essential in the real world. Maybe if youâre racing really long events like Grand Tours, but Iâm sure that for the majority of us, then getting within a few % is good enough.
If you have enough data to estimate FTP without a ramp test, then I guess thatâs good. What I came here to say is that I LIKE Ramp Tests! Itâs like tabata intervals in that itâs one of the few times you really can go all out and give everything. No need to keep anything back, no need to plan beyond the next 60s. In terms of effort management, mental toughness and self-knowledge I consider Ramp Tests to be extremely valuable.
If an athlete isnât getting accurate results form a Ramp Test (over OR under) then I suspect that it says more about their mindset than about the test protocol.
Everything you said was ok until this point. There is no need to judge or point fingers. Also, one big problem in general - and I am definitely guilty of this - is the word accurate has different meaning for different people. Does accurate mean less than 1% error? Less than 5% error?
If you like ramp tests, great! They are a good idea and have a definite purpose. The problem only comes in with how the results are interpreted.
1 Like
Not being judgemental, apologies if it came across that way. I was just trying to say that there are going to be people who quit before their limit, and people who push themselves way past what they should - in both cases the FTP test result isnât going to correctly set the wattage percentages in the TrainerRoad workouts.
Both those groups would benefit the most from a predicted FTP (if the accuracy is at least as good as the output from a Ramp Test - which I imagine has been validated with the big TR dataset)
Cool. But the whole point of the ramp test is not to think at all. You canât go past your limit on it. And, if one gets an underestimate, it could easily be just that their max aerobic power is close in line with their sustainable power. There is just no way around inherent uncertainty in the test that has nothing to do with how the user implements it. Good news is that a lot of the problems from that are well-addressed with AT.
1 Like
Itâs got nothing to do with desire for folk who are out liars from the accuracy of the Ramp test. It is accurate for 85% of the population and has a lot of benefits in being repeatable without incurring too much TSS. Some folk however (around 7.5 %) have a higher MAP (or whatever its called, short term power) and can comfortably on a ramp go beyond their FTP. So a ramp test will overestimate for them without them actually desiring to push too far. Other folk have a better long term power and push harder and longer over s longer period and have plenty of desire to suffer but in another way. For those 7.5% of the population the Ramp will tend to underestimate their FTP. Both categories may be better off with another style of test to estimate FTP but AT should be able to fill the void and catch those outliers.
1 Like
Do you have a source for the 85% claim?
It been talked about in a podcast or two, the ramp is a good estimator for but follows the standard âBell Curveâ. A standard Bell Curve has an 85% accuracy 
I think that was an off hand comment without data to back it up. The guy that popularized using a ramp test for cycling, more than 20 years ago, found that athletes he studied (many time trialists) had multipliers between 72-78% IIRC. That doesnât imply a bell curve distribution. Beyond that I havenât seen much data, and when I started using the (public beta) ramp test, TR made no attempts to have me do a 20-min or longer test to compare.
1 Like
Perhaps Its only really the principle which I think is valid. Ramp tests are valid for the majority and have the benefits of low TSS and ease of pacing but arenât suitable for everyone when it comes to FTP prediction. It would be interesting as you say to get a more comprehensive number. If accurate FTP prediction becomes a reality I would hope TR can quite easily put something together.
The basic principle is if you donât have the data to back it up, donât make broad statements and give precise numbers.
HYPOTHETICAL example⌠lets say you run an experiment with 100 athletes. You accurately measure FTP and within a day or two use the TR ramp protocol to estimate MAP. You could possibly end up with a distribution like this:
where 12% of athletes have an FTP that is 75% of TR Ramp MAP (1-min power). You can make two immediate statements from that graph:
- majority have valid FTP estimate using 75% multiplier
- it is not a bell curve distribution
Again its HYPOTHETICAL example meant to illustrate the old saying âlies, damned lies, and statisticsâ
1 Like
Yes, damn lies and statistics, I wouldnât get too hung up on a %; 90% or is it 99% of them are made up on the the spot. Even when the research is done thereâll be sample errors etc. The broad principle is correct a ramp test whilst good for the majority, doesnât fit all 
In a very recent podcast Nate mentioned again that it was possible internally for TR staff to see what the AI thinks would be your FTP after a training plan/phase assuming you completed the prescribed plan. Is there any update on when this would be available for users? This would be a very motivating (or demotivating ) feature.
1 Like