Froome's SRM data on Ventoux

[Alex Simmons/RST]

Re: Re:

I was referring to left/right imbalance with Stages.

Well we simply can't answer that question.

On the other hand I’m not fully convinced by the test they did. IMHO, we should expect power measured at the wheel using an indoor trainer to be smaller than power measured at the pedal with SRM/Stage.

The test wasn't about the difference between power at the pedals versus power at the rear wheel (i.e. drivetrain losses).

The computer controllable indoor trainer was simply there to provide a consistent resistance to the rear wheel for each of the two test scenarios being measured by the crank power meter, i.e. with (i) circular and (ii) non-circular rings fitted.

The actual resistance load was the same for each test but the power measured by the crank power meter differed depending on which chainring was fitted. This should be no surprise, since it's a well known issue for power meters which assume a constant crank rotational velocity when calculating power (since they only sample rotational velocity once per revolution).

 

[Alex Simmons/RST]

Re:

Not to oversimplify, but I always thought power to bottom bracket was just that. I'm admittedly unfamiliar with stages, but power applied should be static and easily calculated.

Well torque during a pedal stroke is highly variable, so no it's not that easy to calculate accurately, which is why it is sampled at high frequency with strain gauge technology, and it gets more complicated when to obtain accuracy with such non circular chainrings it also requires high frequency velocity sampling. Stages have enough trouble getting crank velocity right once per rev, let alone ~100 times per rev.

Then again I still run biopace on my old teesdale...

Ironically, the traditional Biopace set up will quite probably result in the opposite power measurement error of Osym rings, since they have the opposite effect on crank velocity variability, that being to slow the crank velocity when it is near vertical, and speed it up when near horizontal.

 

[bigcog]

Vayer and Salle are full of shyte when it comes to the numbers according to this:
http://cyclistsinternational.com/reports-of-froome-doping-wrong-in-so-many-ways/

 

[hrotha]

I don't know how many times it has to be said that Sallet's figure was NOT average watts per kilo but maximum aerobic power.

 

[Alex Simmons/RST]

Re:

I don't know how many times it has to be said that Sallet's figure was NOT average watts per kilo but maximum aerobic power.

extrapolated from an estimate of longer duration sustainable power and not reported with definition of the type of MAP referred to (it matters a lot how MAP is defined).

 

[Dear Wiggo]

Re:

Vayer and Salle are full of shyte when it comes to the numbers according to this:
http://cyclistsinternational.com/reports-of-froome-doping-wrong-in-so-many-ways/

The author is an idiot.

This is patently false:

Conversely, professional cyclists are tested in and out of competition at any time of day or night.

 

[Dear Wiggo]

Re:

I don't know how many times it has to be said that Sallet's figure was NOT average watts per kilo but maximum aerobic power.

Agreed. People latch on to that like it's a life saver.

 

[Dear Wiggo]

Re:

Vayer and Salle are full of shyte when it comes to the numbers according to this:
http://cyclistsinternational.com/reports-of-froome-doping-wrong-in-so-many-ways/

Did the author only read doping accusations?

Any idiot could see that Froome was struggling against Quintana’s attacks in the final mountain stages of the Tour and used the help of his lieutenants Wout Poels and Richie Porte.

Coz any idiot would have read Froome was riding with some chest infection the last week or so.

The whole article is stupid -- especially when it delves into a comparison with football.

 

[red_flanders]

Re: Re:

So it turns out that Froome's power measurement was from a left side only Stages with O-sym rings. Hardly a reliable data point.

http://www.cyclingnews.com/news/chris-froomes-tour-de-france-winning-pinarello-dogma-f8/

any thoughts on the 1.5% accounted for by the new 'aggressive' chainrings?

otherwise chimes with what you've been saying re: 4.5%-5%

There's no doubt the level of eccentricity will impact power measurement error. I can't really comment on how different the rings are as I don't know.

But as I also explain, the error isn't a consistent one, it also varies with the type of riding and I would expect, for instance, it to be larger error when climbing than on flatter terrain.

Layer on top the left crank only measurement, and that power output asymmetry is naturally variable. e.g. a modest 2% asymmetry means a 4% error in power. And that assumes the unit is accurately measuring left crank power to start with. End result is no one can ever know what the error of that power file is. It's both unknown and variable.

It cannot be infinitely variable. Again, what's the range of error?

 

[Alex Simmons/RST]

Re: Re:

It cannot be infinitely variable. Again, what's the range of error?

For a Stages with an aggressive Osym, assuming the meter itself is perfect and no user error, then anywhere from under reporting by 4% to over reporting by 10% would not surprise me in the slightest.

 

[red_flanders]

Re: Re:

It cannot be infinitely variable. Again, what's the range of error?

For a Stages with an aggressive Osym, assuming the meter itself is perfect and no user error, then anywhere from under reporting by 4% to over reporting by 10% would not surprise me in the slightest.

Thanks. I don't understand the calculations, so I'm happy to assume that's correct. As such, power data with ellipticals is basically useless at this point. No surprise to me since we don't even know the riders' actual weights either.

It does seem odd that the Science of Sport guys got it within 1%.

 

[Random Direction]

Re: Re:

It cannot be infinitely variable. Again, what's the range of error?

For a Stages with an aggressive Osym, assuming the meter itself is perfect and no user error, then anywhere from under reporting by 4% to over reporting by 10% would not surprise me in the slightest.

So why then does Froome stare at his power readings like it is Lady Godiva? Yes, it could be consistent, but inaccurate, which is hard to correct for in race situations - oh yes, 400 is actually 418 with this gearing. Oops, just changed the bike, now 400 is 388. Seems odd.

 

[Alex Simmons/RST]

Re: Re:

It cannot be infinitely variable. Again, what's the range of error?

For a Stages with an aggressive Osym, assuming the meter itself is perfect and no user error, then anywhere from under reporting by 4% to over reporting by 10% would not surprise me in the slightest.

Thanks. I don't understand the calculations, so I'm happy to assume that's correct.

Yeah, some of it is a bit complex, which is why I outlined many of the issues already earlier in this thread and on other parallel threads, and also provided links to more detailed assessments with enough information for others so inclined to verify.

I just finished reading some comments from a couple of power meter users. One is ditching his Stages because units consistently read around 10% low for him when compared with a verified meter (just an asymmetry issue). Another upgraded his Vectors and can't understand why his power readings are now 10% higher. When using such instrumentation, it helps to have some understanding of the quality of the data when assessing in what ways one can validly use it.

As such, power data with ellipticals is basically useless at this point.

Uselessness is in the eye of the beholder. It might be useless for what you wish to use the data for, but it may still provide some value for someone else, especially if they understand its limitations. It's a bit like heart rate data. It's useless for many things too, but can still be helpful for some, provided they understand its limitations and don't try to read things into it that it's not suitable for.

Also, such rings affect power meters that assume a constant crank rotational velocity per pedal rev to calculate power. Not all meters are affected by this, but many spider/crank based meters are (including Stages).

No surprise to me since we don't even know the riders' actual weights either.

Well it's just another unknown variable. To a reasonable extent use of W/kg estimates rather than wattage estimates normalises the body weight variable. Still some wiggle room, e.g. at 20km/h up an 8% grade, there is about a 0.1W/kg difference in W/kg estimates for a body mass difference of 5kg.

It does seem odd that the Science of Sport guys got it within 1%.

Well I don't think the SOS guys got anything. They just report what others say without doing a level of validation I'd expect from scientists.

How do we know their sources got it within 1%?
A: by comparing modelled power of a different rider using an unverified power meter.
See the problem?


Personally I think there is over confidence in the precision of these estimates. Overall I think accuracy of better modellers is probably pretty good, but level of precision means one needs to be careful with any individual data point.

 

[Alex Simmons/RST]

Re: Re:

It cannot be infinitely variable. Again, what's the range of error?

For a Stages with an aggressive Osym, assuming the meter itself is perfect and no user error, then anywhere from under reporting by 4% to over reporting by 10% would not surprise me in the slightest.

So why then does Froome stare at his power readings like it is Lady Godiva? Yes, it could be consistent, but inaccurate, which is hard to correct for in race situations - oh yes, 400 is actually 418 with this gearing. Oops, just changed the bike, now 400 is 388. Seems odd.

Well I don't know what he's staring at. He drops his head a lot but I've no idea what he's looking at.

 

[observer]

Re: Re:

It cannot be infinitely variable. Again, what's the range of error?

For a Stages with an aggressive Osym, assuming the meter itself is perfect and no user error, then anywhere from under reporting by 4% to over reporting by 10% would not surprise me in the slightest.

So why then does Froome stare at his power readings like it is Lady Godiva? Yes, it could be consistent, but inaccurate, which is hard to correct for in race situations - oh yes, 400 is actually 418 with this gearing. Oops, just changed the bike, now 400 is 388. Seems odd.

Well I don't know what he's staring at. He drops his head a lot but I've no idea what he's looking at.

The whole thing is crap. If they are horribly inaccurate, why use them at all. All of this is just excuses from the team. They use them, and they use the numbers, the numbers aren't wrong, they are just insanely high.

 

[red_flanders]

Re: Re:

Uselessness is in the eye of the beholder. It might be useless for what you wish to use the data for, but it may still provide some value for someone else, especially if they understand its limitations. It's a bit like heart rate data. It's useless for many things too, but can still be helpful for some, provided they understand its limitations and don't try to read things into it that it's not suitable for.

I don't want to use the data for anything, as I find it useless anyway, since there is currently no way to know a rider's actual weight during the race. I just said that, so I don't get why you're assuming the opposite. Strange comments.

Well it's just another unknown variable. To a reasonable extent use of W/kg estimates rather than wattage estimates normalises the body weight variable. Still some wiggle room, e.g. at 20km/h up an 8% grade, there is about a 0.1W/kg difference in W/kg estimates for a body mass difference of 5kg.

To be honest, I'd like to hear that figure confirmed from someone else, as I'm beginning to suspect you're looking to create noise around these figures rather than get to the heart of things.

Well I don't think the SOS guys got anything. They just report what others say without doing a level of validation I'd expect from scientists.

How do we know their sources got it within 1%?
A: by comparing modelled power of a different rider using an unverified power meter.
See the problem?

The point is that when the supposedly verified data came in, it showed their estimates, which they don't pretend are based on hard facts, to be very accurate. I'm sure that's clear to you.

 

[red_flanders]

Re: Re:

It cannot be infinitely variable. Again, what's the range of error?

For a Stages with an aggressive Osym, assuming the meter itself is perfect and no user error, then anywhere from under reporting by 4% to over reporting by 10% would not surprise me in the slightest.

So why then does Froome stare at his power readings like it is Lady Godiva? Yes, it could be consistent, but inaccurate, which is hard to correct for in race situations - oh yes, 400 is actually 418 with this gearing. Oops, just changed the bike, now 400 is 388. Seems odd.

Well I don't know what he's staring at. He drops his head a lot but I've no idea what he's looking at.

The team has stated on innumerable occasions that they use power meters as part of their approach. So it really doesn't matter if he's staring at the power meter or not, the team itself claims they use them to race to a number.

I tend to agree the constant looking down is more his awkward riding style than anything else.

 

[Dear Wiggo]

Re: Re:

How do we know their sources got it within 1%?
A: by comparing modelled power of a different rider using an unverified power meter.
See the problem?

lolnope. Try again.

 

[Alex Simmons/RST]

Re: Re:

How do we know their sources got it within 1%?
A: by comparing modelled power of a different rider using an unverified power meter.
See the problem?

lolnope. Try again.

OK then. Answer the first question.

 

[Dear Wiggo]

They calculate the power and then the riders in question give them their power files. They match to within 1%.

Repeatedly.

Just random coincidence, no doubt. But that's a whole lot of coincidence.

 

[Dear Wiggo]

Please tell me the power meters aren't calibrated or they don't know what the wind is doing, etc, etc.

That just makes the coincidence even more ridiculous.

 

[Alex Simmons/RST]

Re: Re:

The whole thing is crap. If they are horribly inaccurate, why use them at all. All of this is just excuses from the team. They use them, and they use the numbers, the numbers aren't wrong, they are just insanely high.

You'd have to ask them, I don't know what they were thinking in using them. Not my recommended choice.

I can only speculate.

If you believe Stages are always accurate, then you are mistaken. Keep in mind not all uses of such data need high quality/accuracy. Perhaps they recognise the limitations and know what they can and can't do with the data. Maybe they have no idea. Like I said, ask them. I'm just pointing out the reality of the equipment. Not believing it doesn't change the fact.

Of course Stages are a sponsor of Sky and they are paid to use them so there is that. Wouldn't exactly be the first time a team used sub-optimal product because of a sponsor.

 

[Alex Simmons/RST]

Re:

They calculate the power and then the riders in question give them their power files. They match to within 1%.

Repeatedly.

Just random coincidence, no doubt. But that's a whole lot of coincidence.

A Pioneer power meter. Have you read any data accuracy reviews on them lately?

I looked at some recent test review data from a DC Rainmaker test in May 2015. 2 bike rides comparing with a Powertap and Garmin Vectors.

Ride 1 average power of Pioneer data expressed as:
% of Powertap: 92%
% of Vector: 88%

Ride 2 average power of Pioneer as:
% of Powertap: 99%
% of Vector: 91%

 

[Alex Simmons/RST]

Then why is it when these models are actually put through some scientific rigour to assess their performance versus verified accurate meters do we find they are in general accurate, but lack precision? Like an SD of 10%, and 6% when windy climbs are removed.

http://www.fredericgrappe.com/wp-content/uploads/2015/01/Millet.pdf

People are over confident in the precision of these climbing power estimates.

 

[bigcog]

Re: Re:

It cannot be infinitely variable. Again, what's the range of error?

For a Stages with an aggressive Osym, assuming the meter itself is perfect and no user error, then anywhere from under reporting by 4% to over reporting by 10% would not surprise me in the slightest.

So why then does Froome stare at his power readings like it is Lady Godiva? Yes, it could be consistent, but inaccurate, which is hard to correct for in race situations - oh yes, 400 is actually 418 with this gearing. Oops, just changed the bike, now 400 is 388. Seems odd.

Well I don't know what he's staring at. He drops his head a lot but I've no idea what he's looking at.

The team has stated on innumerable occasions that they use power meters as part of their approach. So it really doesn't matter if he's staring at the power meter or not, the team itself claims they use them to race to a number.

I tend to agree the constant looking down is more his awkward riding style than anything else.

He's said he does it because his neck hurts if he doesn't, others have suggested it's a breathing technique, ferrari mentioned it in an article somewhere.