Yes and no. I think it depends how a model uses NP in the calculations. Clearly i cannot sustain my max 1 hour NP, i know that from training my actual 1 hour power is around 285-290, but i have 1 hour NP's around 320 which i could not sustain for average power. But, just because i do not go out and slog out a 50-60 minute interval at 285 does not mean i am not physiologically capable of such an effort, that is where the statistical model comes in to make a calculation from my data. I am an all arounder with a continued focus on sprinting. I do a lot of 10-20 minute intervals and occasionally a 40 minuter threshold interval, i have absolutely zero reason to go out and do a 1 hour test. But with WKO's new model without a test like that it will always underestimate my FTP. Of course people will chime in and say the only real way to know is to test your 1 hour power, but if i am tapered and fresh i would rather test my 1,5,20 minute power than a 1hr power test as they are far more informative of how i will race. I want the model to extrapolate from my data to give me a good estimate to base training zones off of and then corroborate these zones during training efforts.

Thought normalised power was to give an estimation of the wattage you could have done if the effort was a sustained even effort?


So you are saying the model is wrong?

Why not extend your 40 minute effort to prove the model right or wrong?

Out of interest what is your 20 min and 40 minute average power?

I can't see how the new model would underestimate if you have intervals that long.

I do not think NP is designed to say this is what you could have averaged, but rather give an indication of the "real" effort of the ride. In the end my FTP test set is 4x12 with 2:00 rest between intervals, i can do this even when fairly tired and its much less boring than 60 minutes. I did this a few weeks ago at 286w. Interestingly golden cheetah estimates my 60 minute power at 285, pretty spot on from my testing. But WKO still says 272 even though i can do 20 minute intervals at 272 and have a HR 15-20 beats under threshold HR.

1. Note the ordering of the seven deadly sins (and realize that sin #3 could be expanded to "via mathematical modeling"):

http://www.twowheelblogs.com/...ns-seven-deadly-sins

2. See Figures 9 and 10 here:

http://home.trainingpeaks.com/...ration-model-in-wko4

3. See below:





"In god we trust - everyone else must bring data." - W. Edwards Deming

Yes an estimate of the power you could have maintained for the same physiological cost if your power output had been constant.

From TrainingPeaks

By taking these factors into account, normalized power provides a better measure of the true physiological demands of a given training session - in essence, it is an estimate of the power that you could have maintained for the same physiological "cost" if your power output had been perfectly constant (e.g., as on a stationary cycle ergometer), rather than variable. Keeping track of normalized power is therefore a more accurate way of quantifying the actual intensity of training sessions, or even races. For example, it is common for average power to be lower during criteriums than during equally-difficult road races, simply because of the time spent soft-pedaling or coasting through sharp turns during a criterium. Assuming that they are about the same duration, however, the normalized power for both types of events will generally be very similar, reflecting their equivalent intensity. In fact, normalized power during a hard ~1 hour long criterium or road race will often be similar to what a rider can average when pedaling continuously during flat 40k time trial - the normalized power from mass start races can therefore often be used to provide an initial estimate of a rider's threshold power. """

A 20 minute effort is only one third the duration. It's a totally different matter extending that effort to 60 minutes. The last 20 minutes will be considerably harder than the first.

Well, yeah, didn't I say as much when I said I never do long, steady rides, especially on a TT bike?

I'm not sure why this vexes you so much, or why you seem to think it's imperative that my ftp be some fairly arbitrary number based on one time trial. As I've said, I've done 300 NP for three hours, routinely do 330+ NP in crits and did 348 last week. I've also done 25 mins at 340 average power preseason in the mountains. Are you actually going to say that such data doesn't matter and I should only use a once-off time trial when all this other data shows suggests otherwise?

My original point all along was simply to ask whether or not a one-off, lower effort would affect the mFTP model when compared to my NP data, not whether or not you thought my FTP should be this number or that. I genuinely don't understand the point of your posts.

Very interesting. Precisely what I was wondering.

I'm trying to get to the bottom of why your TT power was so low compared to your NP powers.

it seems such a huge discrepancy. The NP numbers suggest you should be able to do so much better in a TT.

To throw in some possible reasons.

1. Your position on TT bike compromises power substantially.
2. Severe lack of motivation for the TT effort.
3. The training you do has shifted your power curve to an extreme.
4. Different power meter on TT bike?
5. NP is severely flawed.
6. You are physiologically profiled in such a way that NP does not 'fit'

As Dr Coggan said, there is a massive discrepancy and perhaps you should look into why WKO4 estimates your FTP so low.

As we now have 2 people on this thread with the same example, its something worth looking into. Is it a WKO4 problem or an NP problem or two individuals with an extreme physiology?

Sorry if I gave the wrong impression, I'm asking questions to try and identify what seems to be an anomaly.

FTP is the power you can sustain during a 25 mile TT. If you can't maintain the wattage derived using NP to estimate your FTP for that distance, then that is strong evidence if not proof FTP derived from NP has overestimated your FTP.

See Coggan's No1 from his 7 deadly sins.

1. from the average power during a ~1 h TT (the best predictor of performance is performance itself).

Interesting, I have the same experience regarding the ability to generate high NP numbers in rides with a high VI, while having much lower power in a steady state situation (at FTP or higher intensity). This has gotten much worse for me the last 2-3 years, coincidentally also the years that I did not focus on shorter races (and didn't do much interval training at FTP) but have spent lots of training in Z2 and Z3. I wonder if this has something to do with it?

---

_____________________
Don't forget to attack!

For me, the biggest issue with FTP as a model is it is a big effort to go 60 minutes all out which is both hard to plan and achieve, so everyone takes shortcuts on it. Even if you do the test, it can be hard to find a road where the conditions are good and you can go for a full hour without stopping or having to slow down. I bet most people test FTP in some method of 30 minutes or under, which leads to an issue of variability. Testing with a 20 minute protocol to figure out FTP, maximizes an effort of 20 minutes, not FTP. IMO, life would be much simpler if we just adopted 20 minutes as a standard. That way we wouldn't need 20 different testing protocols to estimate and software to guess the number.

In triathlon its not as big a deal since if you're doing Oly and up those hour efforts are going to be a staple for training. For crit and road racers not so much. In racing, I haven't needed anything more than a 20 minute effort between rests, and usually its much lower (like 3 minutes). Yeah, there are solo breakaways, but they're not all that common and training for it isn't the most efficient use of time.

Well if a system was built around a 20 minute duration at least it would be based on actual performance as opposed to estimated FTP and NP which is an estimation of the sort of power one might have produced if the effort was constant.

As it is using FTP derived from a 20 minute test, or hard efforts, we have a system which quantifies training using an estimate of an estimate.

You're pretty close to actually describing the critical power concept, which I think makes a lot of sense conceptually. CP can be derived from a couple of shorter efforts with no interpolation, but FTP has so much traction that it provides a common language for folks to understand one another, and practically it doesn't seem like there's a vast difference in the real world whether you follow one concept or the other. Although some folks have argued FTP vs. CP so passionately they just got the entire Wattage group temporarily shut down.

 

I was looking more at rowing on a Concept2, where I look at 500m, 2000m, 5000m average power. Well I've done a 60 minute effort occasionally too.

I can see FTP is relevant to cyclists doing 25 mile TTs but it seems to me people doing all other durations are being forced down the FTP route when that duration isn't really of any special importance. So they end up doing tests that might not be practicable or relying on estimates.

FTP is too vague and fuzzy for me. Why use something so vague which relies on estimates then NP which is another estimate to quantify training?

In the end i just use the PM to track TSS, peak power progression, and some post race analysis. Other than that i have stopped worrying about all the details and 4 hour file analysis days, just too low of a return. I like the golden cheetah model as it is spot on for nailing down my FTP from power files without testing, huge help. NP is always taken with a grain of salt, but tends to be a good indicator of effort. Though i prefer to look at time spent in power zones + about of surges to see the real "hardness" of the race. I think many times these software developers make things more complicated that they should be. Academics only compound this problem with semantics about FTP/NP and what really is a good test. In the end its all about winning races and getting faster. If you are not getting faster and winning more who cares.

You're new here, aren't you?

Yay, my first pink post.

LOL...

I am just tired of the endless discussions that seem to go nowhere. As a competitive cyclist i look at something like WKO 4.0 and think "how can this make me a faster cyclist?" Will the new model allow me to train differently and more effectively? Right now i think it is a very polished program with nice features, but nothing that will make me faster. If i was coaching i think it would be really nice, in fact if i was a coach i would only have WKO 4.0 as it would make my job very easy.

As far as calculating FTP and the discrepancy with golden cheetah and my own testing, i realized that one of my 60 minute crit races was not in the software where i had a 60' power of 279 (three man break for most of the race so similar to a TT), golden cheetah did have this file, so that might have caused some of the discrepancy. In that case WKO 4.0 would have been very close the golden cheetah, but still lower by about 5w, which is fine with me as field testing can do the rest.

Also in regards to how to define FTP, i know on here it is debated like crazy, but every single person i ride with defines it as max 60' power and not one person had a different definition.

TSS is totally flawed. Do you really want me to list those flaws?

Except that per se, 20-minutes is insufficient as a means to parse out the underlying contributors to that performance, and that matters for several flow on considerations when using data to help track training loads and guide training decisions*. Hence why using means to parse out the different metabolic contributors (which can be expressed in terms of power and energy capacity) is a much more useful approach.

That of course can be done by including shorter duration mean maximal power data - the question then become which model do you use to assess this? Model choice affects estimates as well. e.g. I chose a random client and ran their 2015 data through each of 5 CP models (using the default settings in GC) and come up with answers spread over a 26% range. 173W - 218W. So which is right?

As for the NP buster thing, sometimes those people exist and that itself is instructive of their unique capabilities and is helpful information, but keep in mind anomalies with input data can affect things as well, and when examining such discrepancies it helps to examine the quality of data first.

The NP algorithm assumes the data it's fed is correct, and in situations with a lot of stop/start pedalling (e.g. crits, CX), that's not always the case. e.g. check the way head units stores/reports power meter data where there is a lot of stop/start pedalling action. Check to see if there are falsely repeated power data points and how much that inflates power. It can be substantial, I've seen AP/NP inflated by between 5-10% in some crits due to this data artefact. This sort of thing doesn't impact steady state efforts.

So before declaring an NP buster, one must pass the "is the data correct?" test. It may very well be, in which case go back to what I said already, i.e. this information is instructive about the rider's capabilities and how to interpret NP data accordingly.


* less so for setting basic training levels since they are fairly broad anyway, the ones I use overlap, and really specific interval work should be a range based on what you can do and have done recently.

---

http://www.cyclecoach.com
http://www.aerocoach.com.au

What did the 2 and 3 parameter models say ?
Each of the 5 models are OBVIOUSLY very different (otherwise we wouldn't bother with them).

Mark

Ask Bradley Wiggins.

Funny...and here you got the wattage list put on moderation by calling me a liar and telling me to stop by my bullshit for pointing out the exact same thing.

So, which is it, Mark: does the three parameter model provide a different (lower) value for CP than the original two parameter model, or does it not? If it does not, then how can you argue that the latter does not overestimate the limit of metabolic control (as based on MLSS, the NIRS breakpoint, VT2, the iEMG breakpoint, etc.)?

EDIT: Here's the study you probably want to start with:

http://www.ncbi.nlm.nih.gov/pubmed/10694142

Sorry, I meant "If it does".

For christ's sake, get a life.
Or get counselling.

You didn't answer my questions.

Funny you should say that. Here is a quote from one of your recent posts.


""1. I'm not sure why, since "the best predictor of performance is performance itself." IOW, when collecting raw data is so easy (relatively speaking, anyway), why rely on a model-based prediction instead of actual data? (E.g., Wiggins certainly didn't for his recent hour record ride...instead, he did extended efforts in training to determine what power/pace he could expect to be able to maintain for 1 h). ""