Oops! Looks like I skimmed past too many posts. Sorry for the duplicate info/comments.

I do not believe that the guesstimate of the Red/new Quarq is accurate enough to use this info to try to correct small imbalances with your pedal stroke. Essentially it is measuring, during half of the pedal stroke, your left leg pushing power + right leg pulling power and the respective opposite for the other half of the stroke. A rider can not differentiate whether they need to pull more with one leg or push more with the opposite leg to correct this guesstimated difference. I believe a reputable member of this forum who does many respected reviews has stated his opinion of how he believes this guesstimate system works or doesn't work. It is however more info than those with other meters without L/R measurements. I personally would just not put too much stock in that info.

I don't think you read my post very carefully.

Plus, if you're referring to Ray, then you might want to research who he talks to about power meter comparisons ;-)

---

http://bikeblather.blogspot.com/

Since I seemed to have been pulled into this by association...

It's an interesting area - and one where companies can make claims, because it's actually REALLY difficult to test/validate this, despite how easy it might seem.

Take for example, four products I've used recently: Polar/Look PM (left/right pedals), Power2Max (same PowerBalance concept as newer Quarq), Brim Brothers prototype (cleat-based, left/right power), Quarq Red.

In looking at the data from them, I see:

1) Polar/Look PM's that I'm shown as left-leg heavy (typically about L55-45R)
2) Brim Brothers almost matches the Polar, at left-leg heavy in the same ballpark
3) Power2Max shows the opposite, generally about L47-R53.
4) Quarq Red shows about L48-R52.

Which is right? Well, I don't know.

Looking at one ride, the Quarq shows L48/R52, while the Polar shows L59-R41. Was the Polar off? Perhaps. Probable. But how do you verify that? You can't. Because you can't manually calibrate it/validate it. By the same token, how would you construct a test to validate the Quarq aspect either? It's really easy (as I showed) to do simple left/right 100%/0% type tests, but it's very challenging to construct a test to detect a subtle difference of only 1-3% during an actual stroke.

(I actually just found a ride with data from both sets back in April, sending over to Tom/Robert for fun)

Then you can add in stuff like the Stages, and wonder how it all fits into the above. In fact it's funny, I've seen posts by folks over the last week that make all sorts of assumptions/accusations about my pedaling balance and how it invalidates my tests somehow, saying that I'm an outlier. From steady-state riding to stopping/start, a few folks have said I'm 'doing it wrong' when it comes to pedaling my bike Quite frankly, I don't think anyone could make that assumption given the data above at present. And even if I was 'doing it wrong' when it comes to pedaling my bike - I think that's somewhat irrelevant to power measurement.

I'm also not saying that the left-only aspect of the Stages is a deal-breaker either (that's not the area that myself/Tom/Robert were really concerned about when it came to things we were concerned about).

Anyway...

---

-
My tiny little slice of the internets: dcrainmaker.com

I'm also not saying that the left-only aspect of the Stages is a deal-breaker either (that's not the area that myself/Tom/Robert were really concerned about when it came to things we were concerned about).

I'm not saying that it invalidates any of the data, but Tom and Robert were saying that Stages was a no go before you ever testied it, precisely because of this.

---

Styrrell

Not really. We were saying it was going to be potentially less accurate due to just it's fundamental architecture (it depends on the rider, but the data shows L/R imbalances to be common AND variable both across and within rides...so, most likely as well). You might get the rare case of someone with either a spot on 50/50 balance, or who's balance doesn't ever vary, and in those cases, the concerns about the architecture would go away. However, all the data I've seen show that those types of riders are the exception, not the rule, and for Stages to claim that they don't see those sorts of variances requires, at least in my mind, some measure of proof, which hasn't been seen yet.

Of course, all of that was assuming that the components that go into making up the power calculation (i.e. an accurate measure of crank torque AND crank rotational velocity) are measured in an acceptable manner. It's in those 2 areas that the data still causes concerns.

Now, to be fair...some of Ray's data also has given me concerns about his Quarq setup too, and I've already discussed with him potential ways to alleviate those concerns.

---

http://bikeblather.blogspot.com/

Well, seeing as how the separate L/R PMs used in 1 and 2 can't (or weren't, in the case of the Brim?) be checked to see how accurately they are reading in situ, I think I'd put MY money on 3 and 4 being closer to reality. Even if 3 and 4 didn't have perfect calibrations, one would expect them to be more "internally consistent" in regards to their torque measurements on a L/R basis.

That actually points out one of the main "issues" with separate L/R power meter solutions, be they pedals, cleats, crankarms (i.e. Rotor), whatever...you now have TWO PMs that you need to make sure are calibrated and working properly. Twice as many things to go wrong as well.

The interesting thing about that, especially since we're talking about the Stages product here, is that I don't think any of those separate L/R implementations have put in any sort of error detection/ correction so that if they detect that one side or the other is kaput, that they would then start transmitting the other side power measure as 2X power. Sort of like turning themselves into a Stages for "limp home" mode ;-)

---

http://bikeblather.blogspot.com/

Hmmm. I don't believe I said that at all. I did say that the Stages would be less accurate -- but I've also emphasized that the important thing isn't how close two power meters are on average, what's important is knowing the conditions under which they differ and by how much. Anyway, even if I had said that doubling the left was a no go (but I didn't), Ray took it off the table so our analysis didn't focus on that.

And...just to preempt the question as to why I 'took it off the table'.


Simply put, doing any sort of left/right leg variation analysis that would be of any value would require more than one riders balance*. I'd think at least a dozen+ riders, in many, many, many different use cases (steady-state, sprints, fatigue-factors, etc...).


But, even if all of that was done, it would still likely show what we already know: People are different.


Given that there wouldn't be a 100% agreement, even if 90%, it doesn't really help since no given person would know if they were a 10-percenter. Which, brings me back to focusing on that being a 'known', and instead looking at everything else (which is what we did).


*Nitpickers corner: Wouldn't the same apply to my testing in whole? Sure, which I said in the very first line of my accuracy section. Though, our concerns weren't actually to do with a given person, but rather some of the technical aspects of it.

---

-
My tiny little slice of the internets: dcrainmaker.com

I believe that the guys at trainer road did some sort of testing using an engine hooked up to a bike.

This was likely direct drive and wouldn't work for pedal based systems.

But I think the only way you could really test the accuracy would be something that you could dynamically, and accurately control how much preasure was used on each pedal.

Start training on powercranks and you will be balanced in a few months time :-)
Sam
samgyde.com

So if neither you or Tom had preconceived notions about Stages then what was this exchange referring to? (From the first thread on ST discussing Stages)Quote | Reply

---

Tom A. wrote:

RChung wrote:

That's exactly the second thing I thought of.

What was the first? :-)

"Here we go again."

---

Styrrell

That people tout every newly announced power meter as the best thing since sliced bread before any data come in, and that most of those people don't know how to do proper comparisons anyway so even when the data do come in there are arguments about what they mean. So my "here we go again" wasn't about the Stages power meter -- it was about the arguments that I was predicting would be made by unknowledgeable people. BTW, I notice in that thread you claimed the SRM and Quarq measure only on one side, too. Hmmm. Seems like it's turning out pretty much like I thought.

That was a mistake on my part but we all make them. Nonetheless in that thread and others both you and Tom claimed that other PMs on the market were calibratable and trust worthy. Subsequent testing has certainly called that into question as either the PT or the Quark or both can't be within 2% of each, given the data from DCRainmakers test.

At this point it would be hard to make a recommendation to someone for any of those three if they wanted accurate repeatable data. But thats one reason rigorous scientific testing calls for double blind testing, impartial investigators. Both you and Tom expressed doubts about Stages from the beginning and both have been involved in the Secret Thing thats been guerrilla marketed on ST a few times, and both requested to be involved with DCRs testing. Thats not the best environment for critical analysis, although I have no issues with either the test procedures, nor the data analysis, but I do thinkl some of the recomendations and conclusions are pretty biased, but like Tom has stated in the past, when someone has purchased something they have a tendency to want to believe in its worth.

---

Styrrell

styrrell wrote:

...and both requested to be involved with DCRs testing.

Just as a minor (tiny) bit of clarification, I asked them to help analyze data. They didn't ask me.

I do believe that there's valid questions around which of the PM's is 'the most accurate' that were on my bike. Which is one reason I actually avoid that question (it's not something anyone can solve). My interest lies within variability between them over a ride (or rides), as well as aspects related to cadence (where non-PM data such as external sensors were used).

But...this is getting off-topic for the titled thread of "left right" distribution...

---

-
My tiny little slice of the internets: dcrainmaker.com

I should know better by now than to respond, but what the heck...Just because it's possible for a device to be calibrated (correctly) and be "trustworthy", it doesn't mean that in all cases it's used that way. Besides, the issues Ray, Robert, and I have been bringing up have nothing to do with the calibration, either torque slope or zeroing, of the Stages.




You act as if Ray's ride data is the only stuff "out there" on PTs and Quarqs...interesting.



Yes, I did...but from a fundamental architecture standpoint and understanding the known variability of L/R balance across and within rides/riders that exists. You seem to be forgetting this statement of mine from post #129 of that original thread you mentioned above:

Quote:

Hi Jesse! IMHO, that view of it being some sort of "game changer" (as compared to your PTs) is probably predicated on the weight and "wheel flexibility" factors...just a guess though.

I'm still waiting to see a more comprehensive look at the data before "passing judgement" though ;-)

Yeah, sounds as if I had my mind ALL made up back then :-\



Relevancy?



Not true. See Ray's response above.

---

http://bikeblather.blogspot.com/

To be precise, Tom asked you to ask Robert for help.

This is no intended slight on Ray, and I think he'll admit to not exactly knowing the "ins and outs" of comparing PM accuracy, but I'd prefer someone more like RChung to take a look at it to compare with other PMs.

---

Styrrell

Hmmm. Well, first, the reason why almost all of the analyses I have done have included a public posting of the data (instead of, for example, simply posting plots of the data) is so others can re-analyze the actual orginal data sets. Ray has posted the data sets he's collected. That's not the behavior of someone who thinks that the analysis is biased -- it's an invitation to others to delve into the data themselves. But second, you continue to claim that it is impossible to determine whether a power meter is accurate. That's just about as correct as your belief that the SRM and Quarq measure on one side. What's more accurate is to say that *you* don't know how to do it. It is a sign of hubris to think that because you don't know how to do something that no one else does, either.

No one thinks their data is biased, but time and again it has been shown that researchers introduce bias when they dont mean to. Hence the use when ever possible of doubleblind studies.

But second, you continue to claim that it is impossible to determine whether a power meter is accurate.

No you continue to claim that I've said something I haven't ever said. Both you and Tom have said that its easy to do so. So I'm waiting to see a similar comparison of a Quark/PT/SRM where you have done so.

What's more accurate is to say that *you* don't know how to do it.

I don't know how to do it. I know what the manuals say to do, I'm pretty sure I could do it with 2 SRMs and probably 2 Quarks and 2 PT, as long as both are kept in pretty controlled enviroments. I doubt anyone can do it with 2 different brands of PM of their choosing on a "real" ride. But I'd like to be proven wrong.

---

Styrrell

WHAT!?!? WTF are you talking about? What product do you think I'm involved with that competes with the Stages? I'd like to know because I'm unaware of it and if I am involved in something like that I'd like to get paid for it.

That's pretty funny.

What ever the product is that has been teased as a virtual windtunnel in a few prior threads. I stated this earlier, you missed it.

---

Styrrell

Dang, you really need to figure out how to make your replies more readable...I had to search through that thing like I was looking for a needle in a haystack. Doh!



In that quote of mine you show above, I was referring to variances they showed plus or minus as compared to their "reference" of 20W or greater in the span of seconds! Not any sort of fixed offset.

The differences shown between the PT and Quarq in Ray's data appear to be mostly the result of a mis-zeroing event and/or some possible drift issues with one or the other PMs. Like I said above, I've already related to Ray how he could go about reducing the variations on the Quarq side for future testing.

So far, however, the variance of the Stages to the other 2 meters have been indicated to NOT be just a torque slope calibration or zeroing issue. Instead, other more fundamental measuring issues are at work, the whole L/R imbalance thing notwithstanding.

---

http://bikeblather.blogspot.com/

I'm not involved with it, I don't make any money from it, and it's not a competitor to any power meter (least of all the Stages). I did think up virtual elevation but I haven't made any money on it. I have received the thanks of several people, a couple of free dinners, and a really nice bottle of wine. You know, the number of mistaken ideas you have is pretty impressive. Do you also think that global warming is a hoax, that the Earth is 8000 years old, and that Obama was born in Kenya? Just wondering.

The differences shown between the PT and Quarq in Ray's data appear to be mostly the result of a mis-zeroing event and/or some possible drift issues with one or the other PMs. Like I said above, I've already related to Ray how he could go about reducing the variations on the Quarq side for future testing.

Sure its possible, but thats just a theory and from recent posts, drift of more than one of the more popular PMs seems to be a big issue. Maybe you have a fix maybe not, maybe Stages has a fix for their issues, maybe not. One of the good things that Ray seem to find is that the Stages doesn't seem to drift or need to be rezeroed nearly as much as the PT and Quark do. Again that needs to be confirmed.

---

Styrrell