Login required to started new threads

Login required to post replies

Prev Next
Re: WKO4's Power-Duration Curve Model Fundamentally Predisposed to Underestimate Power Output? [rmba] [ In reply to ]
Quote | Reply
Yes, IIRC it was reported second-hand in a book.

Regardless of source, though, it is perfectly apropos.
Quote Reply
Re: WKO4's Power-Duration Curve Model Fundamentally Predisposed to Underestimate Power Output? [Andrew Coggan] [ In reply to ]
Quote | Reply
Andrew Coggan wrote:
it's more test cases than have been used to evaluate/validate all models described in the scientific literature combined

Well it was:

https://www.ncbi.nlm.nih.gov/...articles/PMC4919094/

That said, the approach described in this paper isn't as precise as the WKO4 model, and it is a bit difficult to connect the three constants to any underlying physiological determinants of performance. If your only goal is to predict performance at duration B from performance at duration A, though, the simplicity is appealing.
Last edited by: Andrew Coggan: Nov 26, 17 7:12
Quote Reply
Re: WKO4's Power-Duration Curve Model Fundamentally Predisposed to Underestimate Power Output? [awenborn] [ In reply to ]
Quote | Reply
I still haven't had a chance to sit down and try to replicate it in WKO4, but here's a link my previous modelling efforts in case anyone else is interested or trying to do something similar for themselves; it's just an Excel file shared on OneDrive, despite the dodgy looking link!

https://1drv.ms/...ByxE7ndqunpK-HqnKOdl

I believe Excel 2013+ includes some-kind of iterative curve fitting functionality now (Goal Seek), so you might be able to do the whole lot in there (i.e. without the requirement for a third-party stats package to fit the MMP data) but I'm not particularly familiar with it, so I haven't tried as such yet.

Disclaimer: I have played around with that worksheet a fair bit, copying-and-pasting in different sets of data etc, so if there is something that looks wrong then please accept my apologies! I'm 99% sure it's all accurate and consistent, but if it's not it would be a case of genuine error rather than wilful mispresentation. And I'm still sticking with my excuse of it being the off-season for my pitiful power-PB numbers :D
Last edited by: awenborn: Nov 28, 17 8:14
Quote Reply
Re: WKO4's Power-Duration Curve Model Fundamentally Predisposed to Underestimate Power Output? [Andrew Coggan] [ In reply to ]
Quote | Reply
Andrew Coggan wrote:
Nice dodge. When are you going to back up your false statements with actual data?

You have had several years now, but haven't even attempted to do so, despite the fact that I have explained exactly how to test the WKO4 model. Instead, you keep entering threads like this one, repeat your incorrect claims, then pretend to be the injured party when I point out that you are misleading people. Clearly your agenda is not to seek the truth, but to try to sow doubt about my ideas. As such, you are no better than Trev the Troll, and no more worthy of anyone's attention.

Andy,

I'm gonna say it again. I genuinely mean't what I said. You really should seek help. You don't know how to be happy. Try counselling, role play these online arguments, explore the feelings and humiliation you feel when you can't control the argument and want so desperately to win. Talk to a professional.

It will utterly transform your life.

Mark
Quote Reply
Re: WKO4's Power-Duration Curve Model Fundamentally Predisposed to Underestimate Power Output? [liversedge] [ In reply to ]
Quote | Reply
Still waiting for you to post some actual DATA to back up your ceaseless claims...
Quote Reply
Re: WKO4's Power-Duration Curve Model Fundamentally Predisposed to Underestimate Power Output? [awenborn] [ In reply to ]
Quote | Reply
I have to admit, I'm confused by your spreadsheet, as the data it contains aren't based on either the original Perronet model or what you have labeled the "simplified" version (which actually has more adjustable parameters). How did you come up with the numbers that are shown?

As for Excel, it does contain a Solver function as one of the Add-ins. It isn't quite as full-featured as you might find in a full-blown stats or curve-fitting program, but it can get the job done. In fact, I used it to cross-check the solutions provided by other programs (e.g., GraphPad Prism) and the WKO4 solver when developing and validated the model. This proved to be really important, as there are some hidden "landmines" that could have easily gotten overlooked otherwise, at least until it was too late.

ETA1: Here is the Peronnet-Thibault model looks like when you fit it to all of the data. Parameter estimates are as follows (values in parantheses are 95% CI):

A: 13703 (13485-13922)
MAP: 226.5 (221.2-231.9)
E: -11.03 (-13.98--8.075)



ETA2: And here is what your "simplified" model looks like:

FRC: 20246 (18206-22301)
tau1: 28.6 (25.3-32.0)
FTP: 210.0 (205.6-214.4)
tau2: 101.2 (78.3-126.3)



ETA3: Finally, for completeness here is my model:

Pmax: 712.8 (685.9-739.7
FRC: 9065 (8602-9602)
FTP 241.2 (237.7-244.6)
TTE: 1976 (n/a)
Stamina: 84 (81-85)


Last edited by: Andrew Coggan: Nov 29, 17 6:24
Quote Reply
Re: WKO4's Power-Duration Curve Model Fundamentally Predisposed to Underestimate Power Output? [Andrew Coggan] [ In reply to ]
Quote | Reply
Try the PT drop off at 420s as an adjustable parameter, maybe call it TTE,
Quote Reply
Re: WKO4's Power-Duration Curve Model Fundamentally Predisposed to Underestimate Power Output? [liversedge] [ In reply to ]
Quote | Reply
In case you missed it, Perronet and Thibault picked 420 s because their model is based on maximal aerobic power, i.e., MAP or VO2max. Everything else in their model flows directly from that.
Last edited by: Andrew Coggan: Dec 3, 17 5:39
Quote Reply
Re: WKO4's Power-Duration Curve Model Fundamentally Predisposed to Underestimate Power Output? [liversedge] [ In reply to ]
Quote | Reply
liversedge wrote:
Try the PT drop off at 420s as an adjustable parameter, maybe call it TTE,

Okay, to humor you I got off my phone, fired up my laptop, and did as you suggest above, even though it doesn't make any sense physiologically. Here is the result:

A; 13778 (13616-13941)
MAP: 222.3 (218.9-225.7)
E: -39.92 (-48.77--31.77)
TTE: 3194 (2520-3900)


Last edited by: Andrew Coggan: Dec 3, 17 6:27
Quote Reply
Re: WKO4's Power-Duration Curve Model Fundamentally Predisposed to Underestimate Power Output? [Andrew Coggan] [ In reply to ]
Quote | Reply
Hey all,

I know this is an old thread but i read through it over the past few days and it's a super interesting discussion. I am new to WKO4 and i at first had the same reaction: i thought my FTP was 95% of my 20 minute power and i was shocked and appalled when my actual performance data was telling me that it was not!

I'm kidding, but you know what i mean. However, i've come around since then, partially because i started using a muscle oxygen sensor to more directly measure what's happening physiologically. There's a lot of noise in the day to day variation that i do not understand yet (e.g., fatigue, nutrition, hours of sleep the night before, etc. all seem to be factors), but the lesson seems to be that my FTP is substantially overestimated by a 20 minute test: 20 minute test suggests it should be about 330, the PDC says it should be between 305 and 310, but the muscle oxygen sensor is giving me anywhere between 270ish and 290ish depending on the day.

However, i love the iZones! It's given a great new tool / insight into designing training, building up weaknesses but racing strengths.

So question is, given all this, what drives the iZones? Are they all governed by the output of the PDC? Or, are they built around sFTP, but informed by the PDC data? What i would like to do is use the PDC tool, but be able to "peg" FTP at a lower value than mFTP, rather using hte value given from ramp testing using hte muscle oxygen sensor.

It hurts the ego, but i think it will get better results.

what do ya'll think??
Quote Reply
Re: WKO4's Power-Duration Curve Model Fundamentally Predisposed to Underestimate Power Output? [devolikewhoa83] [ In reply to ]
Quote | Reply
devolikewhoa83 wrote:
what drives the iZones? Are they all governed by the output of the PDC? Or, are they built around sFTP, but informed by the PDC data?

The former.

devolikewhoa83 wrote:
What i would like to do is use the PDC tool, but be able to "peg" FTP at a lower value than mFTP, rather using hte value given from ramp testing using hte muscle oxygen sensor.

Sorry, but that's not possible in WKO4.
Quote Reply
Re: WKO4's Power-Duration Curve Model Fundamentally Predisposed to Underestimate Power Output? [Andrew Coggan] [ In reply to ]
Quote | Reply
Thanks for the quick reply!

Okay, got it. I’ll compare next time I do a ramp test, but fact is mFTP and ramp test FTP will probably be close enough that it doesn’t make a real difference anyway.

Thanks again!
Quote Reply
Re: WKO4's Power-Duration Curve Model Fundamentally Predisposed to Underestimate Power Output? [devolikewhoa83] [ In reply to ]
Quote | Reply
If you're using a NIRS device (Humon?), you might be interested in the various experiments that I did using the Moxy 3-4 y ago...the post start here:

https://www.facebook.com/.../?type=3&theater

and end (mostly) here:

https://www.facebook.com/.../?type=3&theater

Edit: Or to just cut to the chase:

https://www.facebook.com/.../?type=3&theater
Last edited by: Andrew Coggan: Oct 2, 18 9:50
Quote Reply
Re: WKO4's Power-Duration Curve Model Fundamentally Predisposed to Underestimate Power Output? [Andrew Coggan] [ In reply to ]
Quote | Reply
Thank you!

I will definitely check out those studies. And yes, it's a Humon Hex. So far it hasn't been super useful for the ramp testing because of lack of off-the-bike analytical tools, but that's supposed to change in the fairly near future. I've just been using it to guide warmups and what people refer to as "tempo" rides, but I want to learn as much as possible about it.
Quote Reply
Re: WKO4's Power-Duration Curve Model Fundamentally Predisposed to Underestimate Power Output? [Andrew Coggan] [ In reply to ]
Quote | Reply
Dr. Coggan,

If you don't mind me asking, what is the duration that the PDC "weights" the most? I'm not racing currently, and i've been using "spot testing" of sorts to keep the PDC valid. But, i just crushed a five minute and 20 minute tests by about 30 watts or so in each case and the zones barely moved. What other durations of maximal efforts ought I to throw in? Hour test? 45 mins? Mind you, i'm not trying to move the mFTP for its own sake, but rather to keep it valid.

Thank you!

Also, does anyone else use the Humon Hex here? Is it worth its own thread to compare notes?
Quote Reply

Prev Next