4iiii Power Meter vs Suito trainer: The results are...interesting

Very long post coming at you:
I’ve been meaning to do this for a while, but I finally got around to comparing the data from my left crank 4iiii PM to my Suito. In the past, I used the 4iiii to control the trainer via Power Match, however this season I have moved to having a dedicated trainer bike, and the 4iiii mounted to my outdoor only bike.
My expectation based on RPE was that the Suito was over reporting power compared to the PM, however, that isn’t the case…mostly.
The setup for this test:
Trainer Road iPhone app, latest bata controlling the Suito in ERG mode via bluetooth.
4iiii Left crank based PM recorded on Wahoo Elemnt Bolt via ANT+
Data uploaded and analysed in DCR Analyzer Tool
The first screenshot shows the stats:

Interestingly, there are some variations in the heart rate data when zoomed. I guess this has something to do with the difference between Bluetooth and ANT+ recording?

So, here is the full power graph:

As you can see, there is a substantial gap between the two power meters. It seems that the gap is most extreme at low power outputs around Z1 and low Z2 (for me). At these levels, the smart trainer to my surprise reports lower numbers than the crank based PM. The difference is as much as 20W which seems extreme. The average is 10W off.

At my Z2ish numbers things sync up a bit better, with the crank based meter now only reporting an average of 5w higher.

As power continues to rise, the difference narrows further, however at this point, the 4iiii is now reporting slightly lower numbers. Close enough to pace a TT based on my indoor training I hope.

I threw in a form sprint, and you can see how drastically the Suito over reports power vs the crank:

What does all this tell us? I don’t know… It certainly raises some questions for how I approach training and pacing. I ride indoors all week, all year, and outdoors on the weekends basically April-November. My key workouts are indoors, and my outdoor rides are largely by feel, so I’ll continue to base my training on the Suito. However, I wonder how the data reported from my outdoor rides will under-report TSS. I prefer very long days, so a difference of 8 hours averaging 180w, but with the data reporting 170 w could make a substantial TSS difference. The most interesting question to me though is, if my 4iiii is over reporting Z2 relative to the trainer, why do I feel so much more fatigued after 2 hours outside than I do after 2 hours inside? Cognitive load? Full-body engagement? Why is my heart rate at 160w outside always a good 10-15bpm higher than it is inside, when the power appears to actually be lower? Because motorists are terrifying? Perhaps.

You’re comparing apples to oranges. Well… I guess it’s 1/2 an apple x2 and another apple. The missing part here is your right side power from the bike. If power numbers are really important then a single sided meter isn’t going to cut it, especially if you’re digging this deep into data. This is time better spent getting a meter that’ll measure both L/R or total power and you’re done.

*then if the Suito is still off you can dive right back into the rabbit hole of optimising your drivetrain, testing protocol, etc.


One interesting thing is that, compared to the Suito, his L/R balance doesn’t appear to be constant. I’ve looked at a fair number of data files collected from riders using a left-side crank and a Power Tap hub, and this appears pretty common: L/R balance isn’t constant.


Pretty big difference in average cadence, too. Which is noteworthy, because it is essential for calculating power.


Been a while since I last looked at my Suito but for me it was a little lower over all and lagged behind the powermeter until the flywheel got up to speed and like the poster above says the Suito power calculation is directly related to cadence and depending upon the steadiness of it determined for me how fast it would catch up. After a week of testing I ignored the Suito and just used the pm. Occasionally though when Ant + drops out on TR but the head unit still reads it. TR defaults back to the Suito, whereas the Edge still reads the pm and I think I can see the same thing.

My thinking is that strength imbalances are to some degree rooted in differences of muscle activation. Personally my left leg definitely doesn’t fire as well as my right. This is evident in single leg squats and glute work like clamshells etc. I’m thinking that as effort level changes, so does the way the body fires those muscles, creating variability between the two legs. This does sort of eliminate the utility of the offset settings provided in power meters though, doesn’t it?

I’ve found the cadence measurement on the Suito to be a bit of a joke. I’ve got a good feel for cadence, so generally know where I’m at within a couple RPMs. The Suito is within 5 RPM for steady mid-power efforts. At recovery efforts, it is wildly inaccurate, like off by 50%. It also drops to about 50% if I take a hand off the bars. During all those scenarios power stays consistent (by value and feel), so the two don’t seem to be particularly linked. I think the cadence calculation has something to do with the smoothness of your pedal stroke. The 4iiii is pretty bang on for cadence.

Given your comment, can I assume that you feel the power data from the trainer is more valuable, as it is “total power”? In your video where you demonstrate your method for testing different meters against each other, you made some comments about the importance of accuracy for e-sports. Would you say that single sided PMs should not be allowed in competitive e-sports?
I agree that the numbers aren’t really important relative to each other, since I do most of my training indoors all that matters is that I use the same source consistently, and that that source is consistent to itself, which it is.

That depends who’s running the show. If it’s a cycling body running championships (BC, UCI, etc), then I don’t believe a single sided meter would be suitable for that level of competition.

1 Like

Thats only true for the power meter.

The trainer is NOT using cadence for power calculations. It is using a different sensor that more accurately measures the rpm of one of the internal wheels/axles.

The cadence guesses output by the trainer are secondary and the results of them using power fluctuations to estimate rider cadence. It’s not driving power data in any way.


Right, this is like how the PowerTap hub measures power – it doesn’t need cadence at all. It does produce an estimate of cadence but that’s for informational purposes, not because its power measurement depends on it. With the PowerTap hub, you can verify the accuracy of cadence, if you need to by comparing wheel speed and gear ratio to cadence. If you can rely on the wheel speed given by direct drive trainers, you can do the same thing.


Re L/R imbalance:

  1. At up to c.75% FTP and when I’m not at all fatigued, then according to my Assiomas, I’m L leg dominant: 53-55/47-45
  2. As FTP moves up, when I’m not fatigued, or only a little, the balance shifts to more like 51-52/49-48, though still L dominant.
  3. At VO2 max, or when I’m pretty tired, it’s around 50/50.
  4. Towards the end of a really hard set of intervals, or when I’m really struggling, I’m slightly R leg dominant: 51-52/49-48.

ALL imbalances are most pronounced at higher cadences.

I suppose the point of all of this is that power imbalances - for me at least - vary by power, fatigue, and cadence, and unpicking all of that to compare accuracy across 2 devices might well be a very tricky job.

I just have the Assiomas as my power source on the trainer as well. It may not be ‘right’, but it is consistent.

Edit: a further observation to muddy the waters even more: out of the saddle, I’m slightly R leg dominant at every power level and stage of fatigue.

I’m sure all of this is somehow connected to recurring L ham tendinitis and lower R back pain, but it’s not my thread…


I don’t think you are that unusual. One of the reasons I tend to discourage people from buying single-sided power meters, especially as an additional power meter on a second bike.

Looking at my power data (Assioma duo’s) I see a lot of short term variation on L/R, even though longer term it is pretty close to 50% and power numbers are reasonably in line with the trainer. After looking at lots of logs that show this, I have little interest in left-only power.