20 min power

Bikeridindude said:

Hello, I just got my new powertap sl last thursday and am trying to figure out my 20 min power just like the book says. I did the warm up and time trial per specifications today and came up with 280. Last Saturday I did a hill day and my 20 min power was 298. Is it normal for hillclimb power to be higher than flatty power? Also, the longest hill I did took about 15 minutes, so there's some slop time in there too which would make my 20 min power a little bit higher if I did have a 20 minute hill. The flatty power was into a hellacious headwind. Maybe I was tired? Suggestions? Thanks.

Yes. In certain conditions, it's easier to generate more power. Slight and constant upgrades typically allow for higher power.

To which extent is this true, well that depends on various factors. A pal of mine, a mate that works with me is a spinner. A 50yo guy that still thinks it is important to ride flat course spot on 100rpm, not lower. IMHO, this self-inflicted cadence is well over what he can really handle. As a result of that, his performances on flat can't compare to mine for instance. But in a hill, where he isn't driven by this spinning madness, our performances are much more alike.

The day he installs a power meter, he will quickly realize that he generates much much more power on upgrades compared to flat. That's just an example.

My best 20m power is undoubtedly higher on flat. That's because I love riding flat, as much as I hate climbing hills.

Bikeridindude said:

Hello, I just got my new powertap sl last thursday and am trying to figure out my 20 min power just like the book says. I did the warm up and time trial per specifications today and came up with 280. Last Saturday I did a hill day and my 20 min power was 298. Is it normal for hillclimb power to be higher than flatty power? Also, the longest hill I did took about 15 minutes, so there's some slop time in there too which would make my 20 min power a little bit higher if I did have a 20 minute hill. The flatty power was into a hellacious headwind. Maybe I was tired? Suggestions? Thanks.

Hi and congratz on that new PT - nice isn't it? 🙂

I would think it is easier to get a more consistent power measurement on a steady climb, than on the flats. Simply put the variables are reduced greatly....wind becomes less of a factor, ditto aerodynamics.

You could of course have been tired, not warmed up, etc - I suspect if you do this a few times you'll get a real sense of what the number is.

So the power increase you noticed on those climbs could be a direct result of having fewer fluctuations. It could also reflect that you make better power at your climbing cadence, whatever that is, than what you use in the flats. If you analyze the files of those two 20 minutes TT's, you should be able to see what cadence yielded a given power.

Thanks for the insight guys. Looking back, I do think that cadence is a largely determining factor. The climb wasn't exactly steady, it had 2 or 3 spots at 15-20%, and 2 short flat-to-downhill spots. My cadence dropped down into the 50's at times. I'll have to look back at my file, but I think that on the flats I was mostly in the high 90's, it just feels right for me to be that high on time trials. Also, if I could keep 90 on the hills I would, but I need a compact triple with an 11-40 cassette to do that. 🙂 I guess what I'm after is this: If I do some of these training regimens with specific power ranges, do I base it on my highest ever 20 minute average power rating?

... and yes, I do like the hell out of the powertap. It's a little bit humbling though. I can't remember exactly what Floyd's numbers were, but I think his threshold is somewhere in the high 300's, and he weighs 20 or 30 pounds less than me...

Bikeridindude said:

Thanks for the insight guys. Looking back, I do think that cadence is a largely determining factor. The climb wasn't exactly steady, it had 2 or 3 spots at 15-20%, and 2 short flat-to-downhill spots. My cadence dropped down into the 50's at times. I'll have to look back at my file, but I think that on the flats I was mostly in the high 90's, it just feels right for me to be that high on time trials. Also, if I could keep 90 on the hills I would, but I need a compact triple with an 11-40 cassette to do that. 🙂 I guess what I'm after is this: If I do some of these training regimens with specific power ranges, do I base it on my highest ever 20 minute average power rating?

Heh....an 11-40 eh? Now that's a tightly spaced cassette I'm sure.... 😄

So what cassette & chainring combo are you using now btw?

The formula in the book has one possible test for FTP, whereby you would do a 20 minute TT, then multiply the AP * 0.95.

You could also collect data for 5, 10, 20, and 30 minutes, then build a critical power curve. So if you wanted to do L4 intervals for say 10, 20 or 30 minutes, you can just base your watts on the power you make for 10, 20, or 30 minutes on the curve you built.

About Landis, yeah his numbers are higher than yours (and most people's) but then, he seems to have used more than just a powertap to get them 😉

Bikeridindude said:

Thanks for the insight guys. Looking back, I do think that cadence is a largely determining factor. The climb wasn't exactly steady, it had 2 or 3 spots at 15-20%, and 2 short flat-to-downhill spots. My cadence dropped down into the 50's at times. I'll have to look back at my file, but I think that on the flats I was mostly in the high 90's, it just feels right for me to be that high on time trials. Also, if I could keep 90 on the hills I would, but I need a compact triple with an 11-40 cassette to do that. 🙂 I guess what I'm after is this: If I do some of these training regimens with specific power ranges, do I base it on my highest ever 20 minute average power rating?

2 questions:
1. What was normalized power from each effort?
2. Was the "flat" course really flat or was it rolling?

Lucy_Aspenwind said:

.....You could also collect data for 5, 10, 20, and 30 minutes, then build a critical power curve. So if you wanted to do L4 intervals for say 10, 20 or 30 minutes, you can just base your watts on the power you make for 10, 20, or 30 minutes on the curve you built....

IMHO this is the 'best' way to determine your FTP and it takes a lot of the guesswork out of the mix 😄

MY02_STi said:

IMHO this is the 'best' way to determine your FTP and it takes a lot of the guesswork out of the mix 😄

Yeah, some others who know their stuff have come to the same conclusion. The nice thing I'm figuring out is that you can then make your L4 intervals whatever length you want and just ride at say 90% of the power curve for that duration. No having to worry about lacking a 15 min or 25min FTP test for example.

I think for the reasons mentioned about testing on inclines, it is important to test on both gradients and flattish roads. If you use numbers gained from a 20 minute climb to decide what power to use for a 10 mile time trial on flattish roads, you may run into problems sustaining your power.

IOW, a 10 mile TT (with an official clock running) is probably the best way to test for 20 minute power on the flats, as it is the only way to put the right level of pressure on yourself to use 100% effort. Even if you take longer than 20 minutes to complete the distance, you will have a very good indication of 20 minute power in there.

My reason for putting this point accross is because there are no gradients longer than 12-15 minutes where I live in the UK, just shorter, sharper inclines.

Lucy_Aspenwind said:

If you analyze the files of those two 20 minutes TT's, you should be able to see what cadence yielded a given power.

and

Bikeridindude said:

I do think that cadence is a largely determining factor

Cadence is a red herring. Since you have PMs, focus on power, not cadence.

Lucy_Aspenwind said:

You could also collect data for 5, 10, 20, and 30 minutes, then build a critical power curve.

You might want to do that, but once you've settled on a protocol what's important is to repeat it exactly. The exact test length isn't so very critical so if you do only two tests at, say, 4 and 18 minutes, but you do them exactly the same way each time you'll get more consistent results than if you do 4 haphazard tests at varying lengths of time. There's not much difference between doing a 20 minute and a 30 minute test in terms of CP estimate, so I wouldn't do both.

BTW, using this method, the estimate of CP will be more robust than the estimate of AWC.

RChung said:

...
The exact test length isn't so very critical so if you do only two tests at, say, 4 and 18 minutes, but you do them exactly the same way each time you'll get more consistent results than if you do 4 haphazard tests at varying lengths of time. There's not much difference between doing a 20 minute and a 30 minute test in terms of CP estimate, so I wouldn't do both.

BTW, using this method, the estimate of CP will be more robust than the estimate of AWC.

When you have 2 data points (for ex. 5 and 20 min) the estimate will be any better if you add a 3rd test ?

What would be a more robust method to estimate AWC ?

Thanks!

Ale Martinez.

amartinez said:

When you have 2 data points (for ex. 5 and 20 min) the estimate will be any better if you add a 3rd test ?

He was saying consistent gives better results than haphazard, regardless of the number of points.

amartinez said:

What would be a more robust method to estimate AWC ?

By using a test shorter than the 5-min minimum that she suggested, say at 3 minutes. The longer you go, the more aerobic-dependent the result will be.

RChung said:

Cadence is a red herring. Since you have PMs, focus on power, not cadence.

Robert,

I agree that, for the most part, you should never focus on cadence. You should pedal at your "self-selected cadence", the cadence that is most comfortable for you.

However, I will say that I typically put out more power below that self-selected cadence. I don't go much faster, and I tire quicker, but I do put out more power...

Hence, this is one of the reasons that I feel some people will make more power going up hills. Their cadence drops and they become less efficient. I am not saying they are traveling up the hill the fastest...

I can see this in my power files as a short, sharp "bounce" in my power graph.

If I am wrong, please correct me...

Jim

otb4evr said:

However, I will say that I typically put out more power below that self-selected cadence. I don't go much faster, and I tire quicker, but I do put out more power...

If you tire quicker, then would you still say you can produce more power over a 20-min duration (like we're talking about)? Or are you really just ruining your pacing by jumping hard out of the saddle and blowing up 3 minutes later? If you're blowing up quickly, then that's not really the same as saying you can put out more power, since the duration is much different.

At low cadence, it's easy to yank on the bars and produce a lot of power (ie, standing starts), but it's not really sustainable so cadence isn't the issue.

frenchyge said:

If you tire quicker, then would you still say you can produce more power over a 20-min duration (like we're talking about)? Or are you really just ruining your pacing by jumping hard out of the saddle and blowing up 3 minutes later? If you're blowing up quickly, then that's not really the same as saying you can put out more power, since the duration is much different.

At low cadence, it's easy to yank on the bars and produce a lot of power (ie, standing starts), but it's not really sustainable so cadence isn't the issue.

Hmmm...

Interesting points...

However, I am not saying that I am standing and sprinting, however, I am only putting out 15 - 20 watts more at a cadence that is ~10 rpm lower for the same speed.

Don't you see the same phenomenon?

Jim

frenchyge said:

He was saying consistent gives better results than haphazard, regardless of the number of points.

Oh yes, I understand that point but muy question is: If you take consistent measures the estimate is any better if you use 3 or 4 points instead of 2 ?

Lucy_Aspenwind said:

...... No having to worry about lacking a 15 min or 25min FTP test for example..

😄 😄 😄

amartinez said:

Oh yes, I understand that point but muy question is: If you take consistent measures the estimate is any better if you use 3 or 4 points instead of 2 ?

Only using 2 data points may not give an 'accurate' regression analysis for predicting values other than the 2 points you've tested. Using 3 or 4 points gives a better prediction of how 'accuate' the estimate is

amartinez said:

Oh yes, I understand that point but my question is: If you take consistent measures the estimate is any better if you use 3 or 4 points instead of 2 ?

Maybe, but very possibly not. By the way it's constructed, this test produces highly collinear points. In that case, you have to balance off the statistical advantage of marginally greater sample size vs. the stress and strain of doing a hard test. First, that's why I'd generally not recommend doing multiple tests on the same day. Second, the marginal improvement in doing three or four tests is truly miniscule once you've gotten comfortable with the protocol. So, once you've settled on the protocol, there's really no reason to do extra tests.

The reason why the AWC estimate is less robust than the CP estimate is purely statistical: it's because estimates of y-intercepts are less robust than estimates of slope. However, mostly we use this method cuz we care about getting an estimate of FTP, and an estimate of AWC is just gravy. So the fact that it's less robust isn't a deal-killer.

As I said earlier, the exact test lengths aren't too critical. Note, however, that there's some evidence that a too-short time interval for the short test may lead to a CP that (slightly) overestimates FTP. I wouldn't go below 3 minutes, and maybe 4 or 5 or 6 are better. For the right-hand test point, there's not much extra to be gained by going beyond 20 or maybe 30 minutes at the max, depending on the person -- I wouldn't go under maybe 15 minutes for the right-hand point. There's a trade-off in all of these choices. Whatever you do, do it exactly the same way each time. Frenchyge is right: fewer tests more carefully done trump more tests [censored] done.