Stages vs. Tacx NEO, 15-20% power offset

Husum89

Hi folks

So, my Stages PM reads about 15% lower than my Tacx NEO, and I figured I'd share the results here and get your thoughts.

I've had the Stages for about a year. This year, my 20min efforts have been 280W instead of 315W last year. I wrote it off as a fitness thing (and that may be the case), although my race times are very similar.

Last week, i got the Tacx NEO, and I soon discovered that the Stages consistently reads about 15% less Watts than the NEO. I've linked to a ~50 minute Zwift ride below, where you can see the results. There's also a description of the test protocol.

I am of course rather surprised by these numbers. I'm a lefty, so it's definitely not just a leg difference. You can also see that the two track perfectly well, but with a 15% offset.

I've contacted Stages, and they say that they are absolutely certain that the Stages PM reads correct wattages within 2% if the zero offset value is in the 885-889 range (it was 885 before the first test that I emailed them).

The NEO is claimed to be accurate within 1%. I emailed them yesterday, but have yet to hear from them.

Anyway, I just wanted to publish these results, but also to hear your comments on a possible explanation. More info will follow after I hear more from Stages/Tacx.

The ride (check the graphs): https://drive.google.com/open?id=0ByU8E ... 2lqdVZmbnM

The test protocol: Both the Tacx NEO and Stages PM were firmware updated. A zero offset following Stages protocol was done for the PM, reading 883. I rode ~50 minutes of zwift, with zwift recording the power from the NEO and my garmin 520 from the Stages PM. Both .fit files were imported to Golden Cheetah and time/power values were exported to excel. The time between the two was normalized, and 30sec average values were created for smoothing of the graph.

Frank Wilson

It's all double dutch to me.

napoleond

Well I know which one I'd trust.
That said, I notice a subtle difference when the realism is set to 100% on Zwift.
My Quarq and my Kickr are pretty much bang on.

meanredspider

When I compared mine (stepped increase in power) I got figures that were very close - especially in steady state. Can't explain your difference other than one or other or both are miscalibrated. The issue is most likely with the Stages.

ben@31

You probably know this already but for beginners it's not the end of the world if the powermeter is inaccurate as long as it is consistently inaccurate.... you know a max effort figure (FTP) and from that, you can calculate the zones your working in (endurance, tempo, VO2 max, anaerobic, etc etc). The power figure could be 10 watts out, but as long as it's always 10 watts out. All you're doing is using the powermeter to work in different % zones.
Interestingly I have a scenario where I really struggle to get Threshold or VO2 max figures on the flat but I can achieve it easily on climbs almost straight away without deliberately trying? Maybe my FTP needs recalculating?

Husum89

NapoleonD: Which one?

meanredspider: I had the same thought. However, neither units have options for calibration. The NEO is calibrated at the factory, and Tacx claim that due to the design, it never needs re-calibration. Stages only have the option of a zero offset (half a calibration.)

Interestingly, Stages replied: "to be honest, there is no right and nor wrong. Each system measures power different, therefore all the values are different, one can't compare systems."

To which I replied:
"I believe i understand what you mean. But as I see it, Watts are not an arbitrary number. An accuracy claim should be in relation to a "true" number (center of the disc: http://cdn.antarcticglaciers.org/wp-con ... curacy.png). Ideally, that center should be the effect put into the system. There are of course many ways of measuring this center, hence the small variations between diffferent power meters. But with a 15% offset, at least one of the power meters is not accurate, or you and Tacx do not agree what the center of the disc is.
If there is no such standard, and the 2% claim is in regards to variation (top right on image), then the unit is precise to 2%, not accurate, and any accuracy claim would be meaningless."

Stages have kindly offered that I return the PM for testing. I'll compare the systems against another Stages and possibly a set of PM pedals first.

Husum89

ben@31 wrote:

You probably know this already but for beginners it's not the end of the world if the powermeter is inaccurate as long as it is consistently inaccurate....

For training, I agree, as long as you always use the same power meter (which people don't necessarily do). However, absolute numbers matter in comparison to others (yes, I'm that vain). Furthermore, all PM manufacturers claim "accuracy". If they are only consistent, then they are precise, not accurate. This is more than just semantics.

I could write a few percent down to different ways of measuring - but not 15-17%.

stevie63

There was a study that someone posted on here recently that showed that the Stages was the least accurate of all the power meters they tested. In addition just because you are left handed doesn't mean that you're not right leg dominant so some of that variation could be as a result of that.

Husum89

stevie63 wrote:

There was a study that someone posted on here recently that showed that the Stages was the least accurate of all the power meters they tested.

Thanks! I just found the study on pubmed. As a researcher myself, I highly appreciate such input.
Interestingly, two of the Stages meters tested more than 8% below "true".

stevie63 wrote:

In addition just because you are left handed doesn't mean that you're not right leg dominant so some of that variation could be as a result of that.

True. But I played football (soccer) for 14 years - I have no doubt that my left leg is both dominant and more powerful. This is supported by the 3rd graph (offset fix) where the Stages records relative higher left-power at lower efforts (end of graph). At cruising speed, the more powerful legs does more work, but at FTP (first half of graph) the leg difference become smaller.

thomasmorris

Just a note "i'm a lefty" do you mean you are left footed? I.e. kick a ball with your left (edit; reading all the thread I see that is the case)?

Then the dominant leg (counter intuitively) is usually weaker (unlike arms where the dominant is also usually stronger). It's because your non-dominant leg is your 'plant foot' and takes far greater forces whilst kicking. Do a quick google and this is the norm.

You say you 'know' your left leg is stronger. Have you done pistol squats with each leg? Done a power test on each leg independently (which you could do on tacx with one legged cycling).

I'm left footed, and my left foot is weaker (45-55%).

napoleond

Yep. I'm right legged. Left leg dominant on the bike.

I wouldn't trust a stages at all. Too many stories of inconsistent reading. Sky have had a lot of problems with theirs and a very high turnover of units (from a mechanic)

thomasmorris

If you are left legged, then it would be usual for your right leg to be stronger. Possibly up to 10% like me.

Stages is claims it is correct in measurement up to 2%, but it only measures half the system then doubles it. So if it's measuring out by 2% then doubles that to give you total power, that's 4% error.

Then taxc is 1%.

Plus Taxc is measuring power at the rear wheel, so you probably have 1-2% of drive train losses from where stages is measuring to where tax is. Although that should mean taxc measuring lower.

So there's 7% error there at least, plus whatever leg imbalance you have. Which unless you have tested that specifically I would imagine would be right leg favoured.

I'm not surprised the readings are out by 15%.

The question is how consistent these readings are. Does your leg imbalance stay the same with fatigue, power output, time of day etc?

Husum89

thomasmorris wrote:

Just a note "i'm a lefty" do you mean you are left footed? I.e. kick a ball with your left (edit; reading all the thread I see that is the case)? Then the dominant leg (counter intuitively) is usually weaker (unlike arms where the dominant is also usually stronger). It's because your non-dominant leg is your 'plant foot' and takes far greater forces whilst kicking. Do a quick google and this is the norm.
You say you 'know' your left leg is stronger. Have you done pistol squats with each leg? Done a power test on each leg independently (which you could do on tacx with one legged cycling).

Thanks for clarifying, I actually didn't know. Yes, I kick with the left.

Honestly, I'm not 100% sure. But I broke my right ankle 7 years ago, and my right thigh is still slightly smaller that the left, so I would be VERY suprised if the right was stronger. This is backed by the Stages left-only readings being relatively higher at endurance tempo, but again, I can't be certain.

I will test pedalling left-only tomorrow. If it's a leg difference, Stages should read very close to 2x NEO power.

Husum89

thomasmorris wrote:

Stages is claims it is correct in measurement up to 2%, but it only measures half the system then doubles it. So if it's measuring out by 2% then doubles that to give you total power, that's 4% error.

This is a common misconception; that having 2% margin two times euates to 4%.

Suppose i pedal 100W with each leg to a total of 200W. The stages measures 2% off on the leg (98W) and doubles this number to 196W. (200W-196W)/200W is still only 2%. Now these 2% are aggrevated if you have a huge leg difference, but without that, it is still only 2%.

And as you say yourself, drivetrain loss would be in favor of the Stages, so you can't include that in the 7%. So the number (without leg difference) is max 3%.

I'll test the leg difference tomorrow. Once that is out of the way, there's still 12% difference to account for.

bendertherobot

They could both be wrong

Husum89

bendertherobot wrote:

They could both be wrong

They could. I certainly feel like my 20min efforts should be 350+ Watts, so let's hope they are both way off .

I heard from Tacx. To check if the NEO is performing correctly, they suggest pedalling at 30km/h in the Tacx Utility app for 1min+. That should take around 140W.

I will perform this and the left-leg-only test tomorrow.

chippyk

Where's BikeRadars resident power meter man, Milemuncher is it?

ben@31

stevie63 wrote:

There was a study that someone posted on here recently that showed that the Stages was the least accurate of all the power meters they tested. In addition just because you are left handed doesn't mean that you're not right leg dominant so some of that variation could be as a result of that.

This maybe it. If not its still a very good article on powermeters and their accuracy and consistency
https://roadcyclinguk.com/how-to/fitnes ... s5YE4Go.97

I certainly feel like my 20min efforts should be 350+ Watts

Ha ha ha. I know that feeling... I'm so knackered surely my powermeter is under reading.

Interestingly I have a scenario where I really struggle to get Threshold or VO2 max figures on the flat but I can achieve it easily on climbs almost straight away without deliberately trying? Maybe my FTP needs recalculating?

I did a 20min FTP again today, because I couldn't get out for long and it's been way way overdue. And yes my FTP has noticeably dropped. Which I expected as I haven't been cycling as much as I was doing a few months ago, thanks crappy weather and shorter days.

thomasmorris

Husum89 wrote:

This is a common misconception; that having 2% margin two times euates to 4%.

Suppose i pedal 100W with each leg to a total of 200W. The stages measures 2% off on the leg (98W) and doubles this number to 196W. (200W-196W)/200W is still only 2%. Now these 2% are aggrevated if you have a huge leg difference, but without that, it is still only 2%.

Of course, how silly of me.

It'll be interesting to here back after your one legged test.

It was me who posted the study on here. Stages had often more than the claimed error, however, that included leg discrepancies of riders as it was a practical test.

napoleond

But to add to the problem of the inaccuracy, it is increased if you have a discrepancy between legs. My left right balance is 52:48
So at 300w on my quarq, my left is doing 156, the right 144, that gets doubled on a stages though to 312w.
12w at my ftp is a big difference.
May not be an issue as long as it's consistent and the stages is all you use, but then when you use a smart trainer or another power meter you'll get problems.

stueys

Borrow another PM ideally to test it. I've a P2M, Vector and a Neo, they all track very closely but I'm comparing full power meters to full power meters. Once you introduce a single sided PM into that mix it's another variable which will no doubt impact accuracy.

And to add, I'm right footed. My vectors (and P2M though that's guessing) tell me that I'm 52:48 left dominant. That split varies between 51:49 and 53:47 on different rides and at different stages of rides. I suspect I'm not unusual which means it's always going to be tricky to reconcile a full PM to a single side only one.

Husum89

I just finished analyzing todays test - and I'm 98% sure I found the answer!!

Drumroll.....
....
....
....

Crap!
You were right, and I was wrong! It seems I have a huuuuuge leg discrepancy.

For the entire ride, there's still a 15% offset between the two. However, during 2x30 second left-only pedalling, the Stages (divided by two for left-only) measured a measly 2,5% lower than the NEO! This is well withing the 2 and 1% accuracy claims respectively.

Using an average 30min near-FTP NEO effort * 0.975 (offset between the two) as the "true" stages effort (left and right), my L/R balance for that particular interval was actually a whopping 43.5%/56.5%!*

I must admit, that I would never have believed this was the case. I will have to do a longer left-only test to verify (especially as the left-only Stages/NEO offset was 5% for the first 30 sec and 0% for the second), but both 30sec intervals track very close between NEO/Stages.

So, the good news is that the initial test indicate that both power meters measures correctly. Bad news is I have a huge discrepancy, which makes the Stages inaccurate. I will have to to a lot more testing to see if my discrepancy is at least consistent. Also, I have a whole new training topic to read up on (which I guess is good news for a data geek), what this means and if I should train left more.

Interestingly, Tacx say that pedalling at 30km/h in their utility app should require about 140W, but it took me more like 160W to maintain. I'm slightly uncertain what this means, especially since the two track well - I'll have to ask Tacx what they think.

Thank you all for some great input! You lead me on to the discrepancy-line-of-thought, which was a big help!

*Left percentage was calculated as (Stages/2)/(NEO*0.975). Right percentage was calculated as ((NEO*0.975)-(Stages/2))/(NEO*0.975)

Husum89

Also, I'm rather curious whether my power drop from last season is real, or whether it's a results of increasing discrepancy. I did have some issues with my left knee for the first half of this season, and this may subconciously have lead my to distribute more power to the right to avoid aggrevating the knee problem.

ben@31

Husum89 wrote:

Also, I'm rather curious whether my power drop from last season is real, or whether it's a results of increasing discrepancy. I did have some issues with my left knee for the first half of this season, and this may subconciously have lead my to distribute more power to the right to avoid aggrevating the knee problem.

In the past 2 months I've still been cycling regularly. No more than 4 or 5 days off but not the big distances and climbs I was doing at start of summer (training load?) According to the Strava data, like the fitness and freshness curve, I've still lost fitness despite still cycling every week. A drop in FTP too.
Makes me wonder how much fitness we would loose if we were off the bike altogether for 2 months?

Husum89

Way more than we would gain in two months . Are you resting enough?
But if the leg-difference was the same last year, my 20min efforts were 310W/0.85 = 365W, and I was certainly not that good.

penski

Interesting to read this thread and the subsequent updates.

I noticed a 20% difference between the Tacx Vortex trainer and the Stages powermeter when I tested a few weeks back. I put it down to the inaccuracy of the trainer, which is a huge 10%, but reflecting now, my discrepancy was twice that.

foggymike

I don't know about the Neo but the Vortex isn't that accurate. I've tried to match mine to my p2m but it's impossible to do it properly. They match ok at my 1 minute power once the Vortex is warmed up (it under measures from cold) but the vortex measures increasingly under as measurement times go down and increasingly over at longer time periods. The vortex is about 30 watts over at 20 minutes and 30 watts under at 10 seconds. As you'd suspect the average isn't usually a million miles off after I've matched them as best I can, but the average is not an especially useful number.

I used to have a Stages and it measured consistently more than my P2M. I now know from wattbike sessions and the less acccurate L/R split the P2M gives that I'm right handed but left leg dominant when pedalling (like the poster above) so that makes sense. Not really a problem for training as long as you just use the same power meter.

My experience is you can't really expect a turbo (at least a cheaper one that calculates/estimates power rather than measures it), a 1 sided and a 2 sided power meter to match up. They might if you are very lucky I suppose, but it's unlikely!