Stages L Sided Info: inaccurate versus others when right side dominant

Just sharing some personal experience with reference to left sided Stages power meter for info. I have been a long time user of Quarq spider PMs and have them on all my bikes… until recently when my trusty Quarq Riken (9 years old) finally died on my TT bike. Thinking I was going to get a new TT bike soon I went for a quick and dirty ‘fix’ and got a left sided stages. Quarq are dual sided and I have Dzeroes on the other bikes. I am 46/54 or 47/53 right side dominant. Stages takes left sided and multiplies x 2. My conclusion after 2 weeks was that it’s a total waste of time IF you are mixing PMs. My smart trainer and other Dzeroes are well within tolerance but the Stages is miles out (lower). So much so, it’s coming off and being replaced. Nothing wrong with the Stages approach if you are using them and them alone, but they sure don’t seem to present accurate data under these circumstances. Be interested to hear of other experiences. Sharing is caring etc etc.

1 Like

With left-sided 4iiii you can adjust the multiplier to compensate which works OK most of the time. Seems odd if Stages don’t do that since it’s an easy software fix. There are discrepancies in all of them, so best to stick to a single brand.


I heard stages is inaccurate no matter what

How is that fact established?

That exist within a brand/model as well. That is why qc and calibration is the most important step in pm production. And also the costliest step so most vendors cheap out on it.

Not much to do on consumer side other than:
-buy from vendors who are open on how they test and calibrate
-buy models that allow user calibration-slope setting
-enforce a referral systems. A pm can only be trusted after matching 2 others

It is tough to get good data. It will be even harder with the race to the bottom on prices. A 100$ pm is not a good thing.

1 Like

Do any of them do this? It’s not something I’ve ever looked into

Power2max calibrates “every single one of the NG units” according to this GCN video:

It is stated to be the main difference to the ECO model and a 100€ price difference.
I am not aware of any other dynamic calibration. Heck, I believe some vendors even skip the static calibration by the look of things.
That does not mean Power2max is still doing this or nobody else is doing this but them. We simply do not know.
SRM mentions the following:

SRM uses a 144 point calibration protocol against a known mass measured to the exact gram. With exacting calibration and our auto-offset feature, the resulting power measurement of the SRM has error of less than 1%.

That is of course falls short of a dynamic calibration.

And none of those tests against:

  • pedaling center offset (how far away is the force from measuring point, pedaling wide or narrow). Some handle this, some fall flat.
  • temperature differences up to 30 degrees (mountains anyone). some cannot do this (SRMs except last generation, old powertap hubs). Some do it using a model-wide fixed table. Some with a table generated for each unit.
  • non-stable pedaling speeds (half a stroke at 60rpm angular speed and half a stroke at 120 rpm speed, changing the leg speed within a revolution)
  • Parasitic loads. Related to above items but deserves its own title because PM has to be designed, programmed and tested for this. No static or dynamic calibration with perfect loads and pedaling styles will catch this one.

Watch how the first item masses up a Stages SB20 multipoint crankset:

I guess giving GCN a tour is a start but I’d rather have an independent body with standards that they all follow. Without that we’re dependent upon the word of the companies selling the product to tell us they’re accurate. Not exactly objective.

But for me, it’s not the biggest issue. I train on a Kickr bike for that most part and have my stages left PM on my gravel bike and it seems accurate enough. I used to have Assioma pedals but never used them on my gravel bike so sold them. But I’m not too concerned with my stages pm if it’s a little inaccurate. I’d notice if it was 50watts off.

it is not about the tour. Check out the machine they got built and the claim of dynamically calibrating “every” single unit. That is a huge difference to a single point static torque test.

Matching others isn’t really a good test. Consumers can do their own dynamic check but it’s a PITA so hardly anyone does it (unless they’re really really desperate).

Dunno about that inaccuracy. I have Stages on 3 bikes, Garmin on 1 and train Wahoo Kickr and Stages trainer at gym. All seem close 'nuf. How accurate do you really have to be? Seems a relative thing.

1 Like

How? Those test rigs cost mid 6 figures to get built.
Matching pms is the only real-life dynamic test consumers can do.

I did a hill test once.

I had three PMs on my bike. I rode up a hill I knew the distance and elevation. I knew my weight and bike weight. And used a calculator to estimate the workload. Then I rode up the hill and compared the results.

That’s a free dynamic test. You only need one power meter to conduct this test. No need to compare two or more.

Power2Max gives instructions: Accuracy – Power2Max North America

1 Like

I’m a stages guy.

My PM history has been 2x cyclops rear hub, 3 stages left side, and a cyclops hammer (smart trainer).

All my stages have been consistently accurate (I have them on 3 bikes right now).

Riding my stages on the trainer with the hammer and stages hooked up to different head units gives accurate readings. The stages tracks like 10w lower at lower power (>300w), and tracks higher by a similar amount at high power (500w+). So it’s non-linear and depends on the wattage. I had this in a spreadsheet a while back.

I did not have a noticeable change in power numbers when I had my original switch from the rear hub cyclops to a left only stages.

Maybe I’m more balanced? You might consider contacting stages warranty and try to get a new one. If mine is reading low, maybe I should think about going pro! Lol

1 Like

@GarageLab gave the right answer. You don’t need a 6 figure test rig. You don’t even need a hill, but whether you have a hill or not, in order to get accuracy and precision it’s a PITA so hardly anyone does it. Climbing the hill at different speeds allows you to validate across the entire range of power, cadence, and crank torque.

If you are using an online calculator and calling that a validation you are better off with Strava estimations as well.
Unless you are a god or something that is not a confirmation.
Mind you I am not talking about happiness or satisfaction given by a power device. If you are happy I should rather stay out of it.
I am being a realist and that is the reality.

I appreciate your opinion, however I disagree. Just because you don’t believe this can be done with some degree of precision doesn’t mean it can’t be done. As Robert so truly stated, it’s a PITA to conduct.

The hill test is a field test that can be used to validate your device. This isn’t Strava guessing many of parameters. Ideally you do this on a no wind day on a good surface with little traffic and eliminate/control as many variables as best as possible. It’s a PITA to do.

Same with field aero testing. It works but it’s a PITA to do.

Same here. 3x Stages L crank and one P2Max. They are a guide for me, not a piece of metrology equipment where I’m measuring hundredths of a micro inch for silicon wafer fabrication.

Single crank is always going to be less complete info than dual, a spider or pedals.


Exactly. On those occasions when we have done the PITA controls and protocols, we validated not against an online calculator but against a full power dissipation model, then altered drag by a known amount to see whether we could capture that accurately. Drag can be altered by a known amount either by adding objects like spheres of known size and measuring the change in CdA or adding objects of known weight and measuring the change in rolling resistance.

The point is, you don’t need a $100,000 test rig, and you don’t have to settle for “two out of three” comparisons against other power meters. You do have to be willing to do PITA tests.

As an historical aside, VE and modern ways of estimating CdA from field tests were a side effect of trying to come up with a way to do dynamic testing of on-bike power meters.


that is a million miles away from calculators or any other “soft” validation.

full power dissipation model
I am guessing that is not a back of the napkin calculation.

Coming back to the user testing pms. Unless you have a test rig or have the intellectual skills your best bet is comparing pms.

lets be clear. You can model anything with enough brain and computer power. It is not easy. It is not even hard. It is extremely difficult.

And both a good model or a test rig misses on other pitfalls I listed above. What if rider is pedaling wonky? Test rigs are pedaling perfectly with a stable PCO for example. A model looks at the total power required and compares that with the measured power. What is powermeter is reading high on the left and low on the right but gives out perfect number due to rider imbalance? That is exactly what happened to me with SB20 and Assioma pedals.

But if you’re just using them for “good enough” why do you need power at all? I believe @RChung has referred to them at expensive speedometers on other forums