Power Data Estimates for the climbing stages

[acoggan]

Glancing at the SRM data::calculated comparisons tweeted on a regular basis, the calculations are within claimed standard power meter accuracy of +/- 1-2% more often than not.

At least for Sorensen, w/in +/- 2.3% 68.2% of the time (but 1) it's the exceptions rather than the rule that are the issue, and 2) I don't know how much confirmation bias might exist in this dataset):

http://veloclinic.tumblr.com/post/59842116787/sorensen-srm-vs-cpl

 

[red_flanders]

Any source of error.

Are you aware of any specific error? You say GIGO like you're aware there's some error.

 

[Von Mises]

Yes, the "garbage" being that Horner is 65 kg. The power on his SRM and Vayers caculations are actually dead on when you take into account Horner's actual weight, which is about 62kg or so.

1) How do you know what Horner´s weight is?
2) Even if Horner is 62 kg Vayer still signifactntly overestimates compared to SRM. I mean, SRM shows Horner pushed 393 watts for 4,6 km - time of Horner´s attack. But Vayer calculates power for whole 7,4 km and gets same result - it is not logical.

 

[acoggan]

Are you aware of any specific error? You say GIGO like you're aware there's some error.

The physics of cycling are quite simple and well understood, and our model has been validated under conditions (i.e., high winds, rapid acceleration) far more challenging than climbing. Thus, any discrepancy between the actual and predicted power output of Horner (or anyone else, but I responded to someone who asked why there was a discrepancy in his case) would almost certainly be due to an error in the estimated or measured value of one or more of the input variables (including body mass when converting measured power in W to W/kg).

Or more succinctly: GIGO.

 

[zigmeister]

1) How do you know what Horner´s weight is?
2) Even if Horner is 62 kg Vayer still signifactntly overestimates compared to SRM. I mean, SRM shows Horner pushed 393 watts for 4,6 km - time of Horner´s attack. But Vayer calculates power for whole 7,4 km and gets same result - it is not logical.

That really isn't that amazing of power output, not sure why people think 393w is so out of this world for what 20 something minutes at most?? It is way below what 1hr record guys put out for 2x the time period.

Next, he is out of the saddle often when climbing. Hence, his body weight is being used to drive the pedals and the power numbers will give increases on every downstroke of the pedal compared to when sitting. I know SRM takes more data points compared to Quarq per second, so those downs troke segments will likely be better logged.

If you have a power meter, and climb out of the saddle, your power output is higher just from the body weight on the down stroke. You can watch it happen. Sit back down in the saddle, the effort becomes more consistent that is required while driving your legs.

Not sure why people don't consider his climbing style and effect on power numbers. People are so up in arms about a possible 3kts xwind/hwind and how it effects the numbers by x% amount, yet don't consider simple physics of standing on the pedals compared to sitting and power data based off this style of climbing.

 

[red_flanders]

The physics of cycling are quite simple and well understood, and our model has been validated under conditions (i.e., high winds, rapid acceleration) far more challenging than climbing. Thus, any discrepancy between the actual and predicted power output of Horner (or anyone else, but I responded to someone who asked why there was a discrepancy in his case) would almost certainly be due to an error in the estimated or measured value of one or more of the input variables (including body mass when converting measured power in W to W/kg).

Or more succinctly: GIGO.

So (again) what is the error in this case? What is the garbage? Or anything that contradicts the model is suspicious or garbage?

 

[acoggan]

So (again) what is the error in this case? What is the garbage? Or anything that contradicts the model is suspicious or garbage?

I have no idea. My point was simply that if the predicted and actual power outputs differ, it is almost certainly because of an error (or errors) in the input(s), and not a problem with our model itself (which, again, has been well-validated under more extreme conditions).

 

[V3R1T4S]

Next, he is out of the saddle often when climbing. Hence, his body weight is being used to drive the pedals and the power numbers will give increases on every downstroke of the pedal compared to when sitting. I know SRM takes more data points compared to Quarq per second, so those downs troke segments will likely be better logged.

If you have a power meter, and climb out of the saddle, your power output is higher just from the body weight on the down stroke. You can watch it happen. Sit back down in the saddle, the effort becomes more consistent that is required while driving your legs.

I am not sure what you are getting at here. Yes, standing allows one to use his body weight on the downstroke, but it also means one has to pull his body weight back up for the other leg's downstroke. Having to move your mass up constantly is part of the reason why large riders don't stand as much as smaller ones. Chris Horner, who is probably ~62kg, is one of the latter. You make it sound like standing always produces a higher average output, but there is no such thing as free energy.

Also, since crank based power meters assume a constant angular velocity, and during the power phase of a standing downstroke the crank will be moving at a higher angular velocity than the dead part, I would think standing power is actually under-sampled in comparison to the ideal case - sort of the opposite of why oval/harmonic rings inflate power numbers on crank based systems.

 

[Dear Wiggo]

That really isn't that amazing of power output, not sure why people think 393w is so out of this world for what 20 something minutes at most?? It is way below what 1hr record guys put out for 2x the time period.

1. Power to weight is what you should be comparing, not absolute power. In this instance, it is not way below. It is essentially the same.

2. Physiological preparation for efforts should also be compared. In this instance, the build, taper and subsequent peak of an hour record attempt, vs climbing a mountain of a stage in the second week of a GT at the tail end of the season. No way known the efforts can be compared.

3. Not specific to this example, but consider also: no hour record attempt is riding with the thought he/she has to get up tomorrow and do it again, or worse still, descend the current climb at 70-100km/hr, concentrating intensely for 20 minutes so he/she doesn't go off the mountain, and then TT through a valley and climb again.

It's been shown fairly consistently that what someone can sustain for 20 minutes approaches what they can sustain for 1 hour; +/- 5-10%. For a single, solitary TT effort. So to do it mid-stage is even more telling, and probably close to what that rider could do for an hour.

That's why it's a big deal.

What's a big deal for me is when Mr "Facts" sees a post as illogical as yours and lets it slide. Logic tells me that's because it sides with his tenor. Only disagreeing with posters who hold a different position reinforces the notion that his claim to defending "facts" is dubious.

 

[Alex Simmons/RST]

3. Not specific to this example, but consider also: no hour record attempt is riding with the thought he/she has to get up tomorrow and do it again,

Graeme Obree's hour record was set less than 24 hours after his initial attempt. I'll agree he didn't plan to do that, but he did. At 8am the following day.

 

[acoggan]

What's a big deal for me is when Mr "Facts" sees a post as illogical as yours and lets it slide. Logic tells me that's because it sides with his tenor. Only disagreeing with posters who hold a different position reinforces the notion that his claim to defending "facts" is dubious.

Or I just don't visit here on a regular basis.

(BTW, speaking of facts: I converted Horner's power in W to W/kg using 65 kg as his mass simply because that is what was provided in the article to which I originally linked.)

 

[Race Radio]

Ross Tucker on Twitter today

I too would like to call out Tim Maloney (& a few others) to engage, not sling. Bring Coggan & Coyle, let's @veloclinic & I explain nuance

 

[131313]

I have no idea. My point was simply that if the predicted and actual power outputs differ, it is almost certainly because of an error (or errors) in the input(s), and not a problem with our model itself (which, again, has been well-validated under more extreme conditions).

Sure, but in this case power outputs don't actually differ...so your Socratic pontifications don't seem all that relevant.

 

[acoggan]

Sure, but in this case power outputs don't actually differ...so your Socratic pontifications don't seem all that relevant.

Indeed, they would agree* if Horner actually weighed just under 62 kg (and his kit weighed exactly whatever was assumed as well). If he doesn't, though, then they don't (and the reason would be almost certainly be GIGO). Maybe you can tell us precisely how much he weighed at the start of that ~15 min effort? (ETA: I used 65 kg to be internally consistent with the linked article.)

*Ignoring the fact that somebody posted that they were actually over different distances/durations.

 

[acoggan]

Ross Tucker on Twitter today

The twitterati speaketh!

 

[BigBoat]

So I hear of 393 watts average for 15 mins for a 64 kg. The altitude is not very high (3,000 ft average). This would indicate an FTP of around 5.9 w/kg if Horner gave us the true file and didn't alter the data (which is nearly effortless to do and not mentioned too often as a possibility).

This however, is pretty consistent with what we've seen over the last 6-7 years in cycling, especially at the Vuelta which typically has the "slowest" winner of all 3 Grand Tours. It certainly is a strong possibility that that is the UN-adulterated race winning file.

Is Horner riding clean, I think not! The empirical evidence against him is a lot higher than his current performances would show.

 

[Dear Wiggo]

Or I just don't visit here on a regular basis.

Yeah Simmons uses the same defense. Another irony, given your collective tenor is also in harmony.

 

[131313]

Indeed, they would agree* if Horner actually weighed just under 62 kg (and his kit weighed exactly whatever was assumed as well). If he doesn't, though, then they don't (and the reason would be almost certainly be GIGO). Maybe you can tell us precisely how much he weighed at the start of that ~15 min effort? (ETA: I used 65 kg to be internally consistent with the linked article.)

*Ignoring the fact that somebody posted that they were actually over different distances/durations.

I can't tell you with 100% certain what Chris Horner weighed at the start of the climb. I'm sure he can't tell you that either. However, here's what I can tell you: I know how much I weigh. I know how much most of my teammates weigh. I can tell you what Horner looked like at Utah, as compared to what he looked like in years past (when his weight was listed also at 64-65K). I can also tell you he looks absolutely nothing like he did then. I can also tell you what I weigh at the end of a 100 mile-ish stage (generally 1.5-2kg less than when I started). That's pretty typical among other riders, at least in hot weather.

So, I have enough information to made a pretty well-educated guess that if he was 65kg at the start of that climb, I'm moving over to body building.

 

[horsinabout]

I can't tell you with 100% certain what Chris Horner weighed at the start of the climb. I'm sure he can't tell you that either. However, here's what I can tell you: I know how much I weigh. I know how much most of my teammates weigh. I can tell you what Horner looked like at Utah, as compared to what he looked like in years past (when his weight was listed also at 64-65K). I can also tell you he looks absolutely nothing like he did then. I can also tell you what I weigh at the end of a 100 mile-ish stage (generally 1.5-2kg less than when I started). That's pretty typical among other riders, at least in hot weather.

So, I have enough information to made a pretty well-educated guess that if he was 65kg at the start of that climb, I'm moving over to body building.

There will be even more doping in body building, trust me. Though second thoughts in pro cycling I'm not so sure?

Na stick to what you are doing and keep it....well as real as you can. I will respect you more, and that is one at least.

 

[Von Mises]

I can't tell you with 100% certain what Chris Horner weighed at the start of the climb. I'm sure he can't tell you that either. However, here's what I can tell you: I know how much I weigh. I know how much most of my teammates weigh. I can tell you what Horner looked like at Utah, as compared to what he looked like in years past (when his weight was listed also at 64-65K). I can also tell you he looks absolutely nothing like he did then. I can also tell you what I weigh at the end of a 100 mile-ish stage (generally 1.5-2kg less than when I started). That's pretty typical among other riders, at least in hot weather.

So, I have enough information to made a pretty well-educated guess that if he was 65kg at the start of that climb, I'm moving over to body building.

But you still havent explained second factor. How can power outputs be the same if one (SRM) calculates it for 4,6 km - time of Horner´s attack, but other (Vayer) are calculating for 7,4 km climb. Distances are different, average outputs cannot be same.

 

[MizunoMX20]

But you still havent explained second factor. How can power outputs be the same if one (SRM) calculates it for 4,6 km - time of Horner´s attack, but other (Vayer) are calculating for 7,4 km climb. Distances are different, average outputs cannot be same.

Power output is stated per unit of bodyweight for a certain period of time. Why could an average power output over 4,6 km not be the same as an average power output over 7,6 km?

Compare it with running speed: you can run 10 km per hour for 1 km but you can also run 10 km per hour for 5 km. The average speed is the same.

So to conclude, your presumption is false.

 

[acoggan]

I can't tell you with 100% certain what Chris Horner weighed at the start of the climb. I'm sure he can't tell you that either.

Well there you go: any difference between the estimated and actual power outputs could be due to using the wrong mass in the calculation (i.e., GIGO).

 

[Netserk]

Well there you go: any difference between the estimated and actual power outputs could be due to using the wrong mass in the calculation (i.e., GIGO).

If you're only 1kg wrong it will result in less than 0.5% error. Not exactly garbage....

 

[Ferminal]

Not sure what the practical applications are if it's garbage unless you have wind tunnel. Of course there are going to be differences when tactics/drafting and wind/rolling resistance are involved. Anyone with a quarter of a brain understands that zero difference isn't actually "perfect" and that the real key in all of this is in the descriptive analysis. This line of discussion is pointless and irrelevant in the context of actual racing.

 

[zigmeister]

I guess posting his power meter numbers isn't enough....people screamed forever about Froome releasing his, he never did...Horner has been putting his out there and talking wattage for 4+ months now during his training pretty openly and what the power numbers that were being put down by the top riders at the Tour of Utah during stages. Specifically what he was seeing on his PM during some of the stage climbs.

But of course, people have some need to translate that into a w/kg, which nobody can get right because no formula (math) can understand the reality of the ride that happened on that day, at that time under specific conditions.

Continue on though...entertaining all of the "I don't know" and "Can't say for sure with certainty" responses.

Einstein may not have been the best mathematician, but one of his quotes is often very true:

"As far as the laws of mathematics refer to reality, they are not certain, and as far as they are certain, they do not refer to reality."