The Powercrank Thread

[elapid]

First, it isn't my product so it really isn't my job to do any testing

So are you admitting it is your job to test your products and not rely on other researchers to do your job for you?

 

[sciguy]

First, it isn't my product so it really isn't my job to do any testing other than what an individual might do, which I am trying to do.

Frank,

Since you seem to be the main if not sole promoter of the product in the USA combined with the fact that you have been highly involved in its development I'd call that a great reason for you being involved it the products testing. You've had access to prototypes for more than a year. That has placed you in a wonderful position to do some truly excellent testing but for some reason you don't seem to find it worth doing.

Hugh

 

[FrankDay]

Frank,

Since you seem to be the main if not sole promoter of the product in the USA combined with the fact that you have been highly involved in its development I'd call that a great reason for you being involved it the products testing. You've had access to prototypes for more than a year. That has placed you in a wonderful position to do some truly excellent testing but for some reason you don't seem to find it worth doing.

Hugh

I can't test product I don't have. Having the hardware was just like having an ordinary PM. I saw no issues with reliability on the road but they have changed the design from what I have so what is going to ship and what I have are two different things.

As I said, my main interest is in the software, as I think that is where the value of the product really is. I have had their first beta iteration of that for 1 week and no capability exists yet to gather information on the road. Otherwise, without the software, you are stuck with something like the Vectors, and expensive piece of hardware with lame, almost useless, software.

 

[FrankDay]

So are you admitting it is your job to test your products and not rely on other researchers to do your job for you?

Testing is not the same as doing scientific study. Scientific proof requires the researcher to be independent to help avoid bias in the results. Anything I do could rightfully be claimed to be potentially biased and, therefore, discounted.

 

[elapid]

Testing is not the same as doing scientific study. Scientific proof requires the researcher to be independent to help avoid bias in the results. Anything I do could rightfully be claimed to be potentially biased and, therefore, discounted.

So you keep on saying and understandingly so, but expecting someone else to do your research for you when it is not one of their interests is also a copout. Furthermore, when you do your own testing and claim a 40% improvement in power (yet don't believe in a PM), then you need to be able to back this up.

 

[FrankDay]

So you keep on saying and understandingly so, but expecting someone else to do your research for you when it is not one of their interests is also a copout. Furthermore, when you do your own testing and claim a 40% improvement in power (yet don't believe in a PM), then you need to be able to back this up.

No, I have done all the "research" I need to do for my needs, i.e., to be able to tell potential customers what they might expect from using the product as we suggest. And, of course, I believe in a PM as a measurement tool. Power is an important aspect of riding fast. My problem is PM's and the claims of their advocates is there is no scientific support for the implication that using one is the best way to train and race.

If anyone wants scientific proof that my claims are correct (or incorrect) that will have to come from independent researchers. I am confident in what my product does for those who use it (on average). However, just like a power meter, there is little scientific support for my belief but, of course, there is also zero scientific evidence debunking my belief. Until then, people will have to take a leap of faith and try the product for themselves and either like what they see or send it back within 3 months for their money-back. Try and get that from a PM manufacturer.

 

[sciguy]

So let me get this straight. For months you've been hammering how revolutionary you feel Icranks will be in allowing athletes to test how the crank length they use influences the power they generate while in an aero position and then you turn around and say-

I personally don't really care much about how accurate the device is .

You've made it very clear that you think shorter cranks will allow athletes to generate more power but you don't care if the device you're selling measures that power accurately??????

Color me confused.

Hugh

 

[FrankDay]

So let me get this straight. For months you've been hammering how revolutionary you feel Icranks will be in allowing athletes to test how the crank length they use influences the power they generate while in an aero position and then you turn around and say-



You've made it very clear that you think shorter cranks will allow athletes to generate more power but you don't care if the device you're selling measures that power accurately??????

Color me confused.

Hugh

To me reliability is more important than accuracy. If my power goes up 10% does it really matter if the number I am following changes from 100 to 110 or 200 to 220? Either way I know my power has changed 10%. Hence, when I changed crank length from 145 to 130 the numbers for my typical ride changed dramatically. I could immediately associate one with the other. It is somewhat akin to putting bigger tires on your car. Immediately your speedometer is off but it is also pretty easy to learn that now when the speedometer says 50 you are really going 60. It would only be an issue if someone was driving the car that didn't know the speedometer was off. It would be better (easier) if I were able to recalibrate the cranks for each crank length because then I wouldn't have to explain this stuff to people like you but I can't so I do.

Anyhow, I am right now more interested in looking at how crank length and position on the bicycle affect pedaling technique and efficiency. It doesn't matter if the calibration of the iCranks is off because it will all be done on the Velotron where there is no controversy as to whether the power is correct or not so it will be easy to adjust any graphs of pedaling technique to correct for any calibration errors.

Edit: one more thing, you have misinterpreted what I think about crank length. While I believe some people will be able to generate more power with shorter cranks that is not always the case. More importantly I think that most people will be able to get into improved aerodynamic positions without compromising power if they were to go to shorter cranks.

 

[sciguy]

Edit: one more thing, you have misinterpreted what I think about crank length. While I believe some people will be able to generate more power with shorter cranks that is not always the case. More importantly I think that most people will be able to get into improved aerodynamic positions without compromising power if they were to go to shorter cranks.

Have you changed your mind since you wrote the following?

I would like to open a discussion regarding the importance of crank length to bicycle racing or cycling in general.

I have posted an "essay" of my thoughts as part of my web site here. I would be interested in constructive criticism to help me fine-tune my arguments or correct any obvious errors.

In summary I feel that shorter cranks do several things for the cyclist.

1. Shorter cranks will improve power output for most.
2. Although this goes completely against the conventional wisdom, shorter cranks can reduce knee stress
3. Shorter cranks allow better aerodynamic positioning without sacrificing power.

And, in general, we are talking substantially shorter than what most would consider to be a short crank. Our data suggests that around 100 mm crank length would be near optimum for most. I am currently riding 105 mm cranks with good feeling.

I believe there are good reasons that explain the above benefits and we can discuss them if any desire.

If possible I would like to open an actual discussion of this issue based on facts and data rather than bias and opinion.

 

[FrankDay]

Have you changed your mind since you wrote the following?

Yes, somewhat. I am not afraid to change my view as I gather more data. I suspect that optimum for most (for aero position) is longer than the 100 mm range I noted then, more in the 120-150mm range (I am now riding 125mm cranks). Further, the power improvement that most will see is probably quite small unless they are in an aerodynamic position but in an aerodynamic position I suspect power may increase considerably in many. It is one of the things I want to try to document and, if I am correct, also see if I can answer the question, why?

 

[Alex Simmons/RST]

Testing is not the same as doing scientific study. Scientific proof requires the researcher to be independent to help avoid bias in the results. Anything I do could rightfully be claimed to be potentially biased and, therefore, discounted.

A thorough study which provides details on process and the raw data can be examined by non-biased parties, and importantly, enables such an experiment to be replicated by others in order to assess if results are repeatable.

 

[Alex Simmons/RST]

Try and get that from a PM manufacturer.

All new power meters come with warranties (and is the law in many countries) and if they are not measuring power to specifications, then all the power meter manufacturers will repair or replace the unit as normal.

I imagine in some rare instances there may be a refund for faulty items but then all they are claiming is that a product that measures power will measure power, and they will supply a unit as such.

 

[FrankDay]

A thorough study which provides details on process and the raw data can be examined by non-biased parties, and importantly, enables such an experiment to be replicated by others in order to assess if results are repeatable.

Here is what you have to do to replicate my data, shouldn't be too hard.

Get a bunch of average cyclists and have them agree to train as they are used to training but doing so on PowerCranks exclusively for 9 months while periodically testing them and see if and how they change.

Ball is in your court.

 

[elapid]

Here is what you have to do to replicate my data, shouldn't be too hard.

Get a bunch of average cyclists and have them agree to train as they are used to training but doing so on PowerCranks exclusively for 9 months while periodically testing them and see if and how they change.

Ball is in your court.

1. Where's your data?
2. Your methodology is poorly described. For a study to be repeatable, not only should the results be repeatable, but those results should be based on using the same methodology. What is a bunch (i.e., define your study population)? What is an average cyclist? What time interval do you define as "periodically"? How do you "test" them? How do you define change and what statistical tests did you use to analyze these changes?

Ball is only in someone else's court if there is a study to test repeatability. Ball is in your court for the initial study.

 

[FrankDay]

1. Where's your data?
2. Your methodology is poorly described. For a study to be repeatable, not only should the results be repeatable, but those results should be based on using the same methodology. What is a bunch (i.e., define your study population)? What is an average cyclist? What time interval do you define as "periodically"? How do you "test" them? How do you define change and what statistical tests did you use to analyze these changes?

Ball is only in someone else's court if there is a study to test repeatability. Ball is in your court for the initial study.

I went to a local cycling club and asked for volunteers to participate in my study. I think I had 10 sign up. Not all finished, it is quite a commitment. If I remember right, I tested them, after the first test, on PowerCranks because, if they couldn't get to the same ending HR at "exhaustion" that was evidence they were not adequately trained on PowerCranks. Of course, subsequent to that test we have had customers report similar results when testing themselves.

The same methodology involves asking people to train on the cranks exclusively and testing them at monthly intervals using a step test (20 watt increments every 2 minutes) to exhaustion. In my data 40% max power improvement occurred as early as 6 months so to "replicate" my data you have to go at least that long. Post testing must be done on PowerCranks to demonstrate adequate training stimulus, as noted above. It doesn't matter what my raw data shows, all that matters is whether the outcome is similar or not.

 

[elapid]

I went to a local cycling club and asked for volunteers to participate in my study. I think I had 10 sign up. Not all finished, it is quite a commitment. If I remember right, I tested them, after the first test, on PowerCranks because, if they couldn't get to the same ending HR at "exhaustion" that was evidence they were not adequately trained on PowerCranks. Of course, subsequent to that test we have had customers report similar results when testing themselves.

The same methodology involves asking people to train on the cranks exclusively and testing them at monthly intervals using a step test (20 watt increments every 2 minutes) to exhaustion. In my data 40% max power improvement occurred as early as 6 months so to "replicate" my data you have to go at least that long. Post testing must be done on PowerCranks to demonstrate adequate training stimulus, as noted above. It doesn't matter what my raw data shows, all that matters is whether the outcome is similar or not.

Thanks for the update. But there are still too many holes. You think 10, but how many actually was it? Some dropped out, but how many is "some"? How many were excluded because of so-called inadequate training? Out of a possible 10, your sample size is looking awfully small with dropouts and exclusions. That's one immediate criticism.

You should also have a control group - a similar group and number of cyclists of similar "average" standard with a similar training program and followup testing, but on "standard" cranks. No use saying PowerCranks result in a 40% increase in max power when you don't know if those that use standard cranks would or would not get a similar increase in max power over the same time period with the same training regimen.

Raw data is raw data, but it needs to be analyzed statistically. Statistics can always be criticized, but there is no validity to a study that does not include sufficient population size and adequate statistical analysis.

 

[FrankDay]

Thanks for the update. But there are still too many holes. You think 10, but how many actually was it? Some dropped out, but how many is "some"? How many were excluded because of so-called inadequate training? Out of a possible 10, your sample size is looking awfully small with dropouts and exclusions. That's one immediate criticism.

It doesn't matter how many there were. It is what we got and subsequent data has supported that number. I am quite confident that if anyone even comes close to replicating what was done with a reasonable number of people that the results will be similar. If you do this study it isn't important what results I got, all that counts are the results you get.

You should also have a control group - a similar group and number of cyclists of similar "average" standard with a similar training program and followup testing, but on "standard" cranks. No use saying PowerCranks result in a 40% increase in max power when you don't know if those that use standard cranks would or would not get a similar increase in max power over the same time period with the same training regimen.

Yes, a good study should have a control group but I wasn't doing a study. I was simply trying to gather data so I could tell people what they might expect if they purchased my product. One good thing about including a control group is one might be able to say PC group increased 40% and control group increased 20% therefore, PC's are worth an extra 20%. Instead, we are stuck lumping all the reasons a person might improve together. Anyhow, good luck with that. It is hard enough to get people to commit to a 6 week study. Try your luck getting commitments to a 6 month study.

Raw data is raw data, but it needs to be analyzed statistically. Statistics can always be criticized, but there is no validity to a study that does not include sufficient population size and adequate statistical analysis.

My guess is that since you have characterized 40% as bull**** that a statistical analysis of a 40% change, even in a small population, when an expected improvement from training effect might be about 10% will result in a P<.01. Anyhow, you are welcome to do a statistical analysis of your large pool data when you get it. Good luck, have fun.

 

[elapid]

It doesn't matter how many there were. It is what we got and subsequent data has supported that number. I am quite confident that if anyone even comes close to replicating what was done with a reasonable number of people that the results will be similar. If you do this study it isn't important what results I got, all that counts are the results you get.Yes, a good study should have a control group but I wasn't doing a study. I was simply trying to gather data so I could tell people what they might expect if they purchased my product. One good thing about including a control group is one might be able to say PC group increased 40% and control group increased 20% therefore, PC's are worth an extra 20%. Instead, we are stuck lumping all the reasons a person might improve together. Anyhow, good luck with that. It is hard enough to get people to commit to a 6 week study. Try your luck getting commitments to a 6 month study.My guess is that since you have characterized 40% as bull**** that a statistical analysis of a 40% change, even in a small population, when an expected improvement from training effect might be about 10% will result in a P<.01. Anyhow, you are welcome to do a statistical analysis of your large pool data when you get it. Good luck, have fun.

Sample size does matter. If there are only two cyclists left in your so-called study, then they may be both outliers. The bigger the sample size, the more robust your findings.

What results you get does matter, because the aim of any further studies are to validate your results. How can your results, and hence your impartiality, be validated if we don't know your results?

If it is hard to get people to commit to testing PowerCranks, then doesn't that tell you something about your product?

Ball is in your court again. At this stage you do not have a study for someone to replicate.

 

[FrankDay]

Sample size does matter. If there are only two cyclists left in your so-called study, then they may be both outliers. The bigger the sample size, the more robust your findings.

Large sample sizes are necessary to show small differences. They are not to show large differences. I suspect you consider a 40% improvement a "large difference" since you call if bull****.

What results you get does matter, because the aim of any further studies are to validate your results. How can your results, and hence your impartiality, be validated if we don't know your results?

No, the object of further study is to confirm or refute prior studies. What is important is whether you follow the same protocol. Regardless of what Fergie thinks a study lasting 5-6 weeks of part-time intervention cannot refute data coming from 6-9 months of exclusive intervention. If your data shows no difference then that really puts my data into question. If your data supports my data then one might conclude I was right all along. If your data is somewhere in the middle then maybe a third person needs to do the study to help decide where the answer really is. And, ugh, you know my results, about 40% power improvement using step test to exhaustion after 6-9 months of training on the device. Your getting the raw data doesn't change that (I am not even sure I know where it is now.)

If it is hard to get people to commit to testing PowerCranks, then doesn't that tell you something about your product?

No.

Ball is in your court again. At this stage you do not have a study for someone to replicate.

If you say so.

 

[CoachFergie]

Sample size does matter. If there are only two cyclists left in your so-called study, then they may be both outliers. The bigger the sample size, the more robust your findings.

What results you get does matter, because the aim of any further studies are to validate your results. How can your results, and hence your impartiality, be validated if we don't know your results?

If it is hard to get people to commit to testing PowerCranks, then doesn't that tell you something about your product?

Ball is in your court again. At this stage you do not have a study for someone to replicate.

Could do a meta analysis of all the previous studies.

The good thing about these studies is that they reflect the actual way people incorporate independent cranks into their riding and racing. No racing cyclist has 6 months to dedicate to total immersion training and as we see in Pro cycling the likes of Cadel Evans and Taylor Phinney would rather ride with a SRM than there independent cranks so real use is minimal sessions within the off season and most likely no use in season.

I think it is notable that those who claim to be able to totally immerse themselves in independent crank use and claim to make gains around the levels Frank claims are usually found to be using an estimate of power, are shown to be using "Doctored" or don't even know how to zero a power meter.

But when the "creator" can't even comprehend why one should strive for validity in their power meter data as well as reliability you do have to wonder.

I see the Powercrank web page now talks about gains in speed after their use. Absolutely hopeless. What can we conclude from that? As anyone will tell you, riding your bike makes you better at riding your bike.

 

[coapman]

Large sample sizes are necessary to show small differences. They are not to show large differences. I suspect you consider a 40% improvement a "large difference" since you call if bull****.No, the object of further study is to confirm or refute prior studies. What is important is whether you follow the same protocol. Regardless of what Fergie thinks a study lasting 5-6 weeks of part-time intervention cannot refute data coming from 6-9 months of exclusive intervention. If your data shows no difference then that really puts my data into question. If your data supports my data then one might conclude I was right all along. If your data is somewhere in the middle then maybe a third person needs to do the study to help decide where the answer really is. And, ugh, you know my results, about 40% power improvement using step test to exhaustion after 6-9 months of training on the device. Your getting the raw data doesn't change that (I am not even sure I know where it is now.)No.If you say so.

You stated in an earlier PC thread, if a rider is already using the perfected circular technique he will not get any power gains from PC's, is that still the case

 

[Alex Simmons/RST]

Here is what you have to do to replicate my data, shouldn't be too hard.

Get a bunch of average cyclists and have them agree to train as they are used to training but doing so on PowerCranks exclusively for 9 months while periodically testing them and see if and how they change.

Ball is in your court.

How did the control group with the same characteristics perform?

 

[FrankDay]

How did the control group with the same characteristics perform?

As I stated, there was no control group. But, our expectation was that it was unlikely that any group of experienced cyclists would see a 40% increase in one season of ordinary training so it seems reasonable for us to make a positive claim. If we only saw a 5% change one would have to wonder if it was PC's or training effect. But, it wasn't 5%, it was 40%. Our claim is what we observed after people trained on the device for a long period of time.

 

[coapman]

You stated in an earlier PC thread, if a rider is already using the perfected circular technique he will not get any power gains from PC's, is that still the case

Still waiting for an answer.

 

[FrankDay]

Still waiting for an answer.

Well, improvement is always possible but I don't think that any improvement being seen would be technique related.