Give me a PowerMatch for my dumb trainer!

I want to match my whatever dumb trainer VirtualWatts with my actual powermeter Watts,

A PowerMatch for dumb trainers!

So I have a dumb tacx Blue Twist trainer and a 4iiii left crank only power meter.
If you select it all perfectly in TR the VirtualPower is just wrong(aka different) compared with the PM.


The Blue Twist has 7 different resistances that TR supposedly know about. It just does not match against the 4iii readings and not just by small offsets. It differs at different rates at different speeds/powers. This is from just looking at both numbers at the same time. I’ll do a more thorough study if someone is interested. So yeah numbers can not be compared.

Though I found that it works best for me (My gatorskins, my unknown tyre preasure, my rotten chain) If I lie to TR about the resistance level I set in the dumb trainer. So I did a little search lying differently about the trainer and the result varies until it is almost satisfactory.

I wonder if the software could learn my dumb trainer power curve when using it at the same time with my crank PM. - Learn this dumb trainer at this settings. And save it.
I kind of want that TR was able to learn which trainer/trainer settings I’m using by itself (or the actual curve). And a way to save it a let me reuse it without the PM if I want.

Why this could be useful?
You have different bikes and only one PM and I want to be “close” if using VirtualPower-ML on the bike without the PM.
You want to be able to do the one leg thing when using a left only PM and a dumb trainer.
You want to “calibrate” friends dumb trainers so numbers look less uninviting when put together, re. group workouts.

We are not supposed to compare VirtualPower but you name whatever the VirtualPower calculates Watts too. So… Does anyone else think about this stuff?

I can understand the desire for data to be consistent, let alone accurate. But if you’re doing workouts with your friends and you’re actually concerned about numbers looking “uninviting”, I think you need to recalibrate your perspective.

I’m not being critical here, simply responding on the internet to someone I don’t know. Now is the time when it least matters and just get on the trainer with your friends over a group workout and enjoy the process. Enjoy their company. Save some money and buy a smart trainer when all the dust has settled.

1 Like

Hey thanks you for trying to get me back to earth! I’ll try to recalibrate my perspective.

I just peer pressed a bunch of friends to train by “intensity”, and VirtualPower. Which is better tan HR and $includeallthetext. Then I tested my dumb trainer against my PM and I realized it is a mess. I know this is a poor people oriented feature and that the solution is just to buy a PM. I just do not like to lie to my friends. I wont trust their suffering face at 105% of their VirtualFTP. And after my single quick test I can bet that 105% of VFTP is never 105% of FTP.

I can think of other uses to have learned/trustable dumb power curve thing. Help L-Only powermeter meassure L/R unbalances, on leg exercises? :shrug:. But yeah.

Perspective recalibrated! Esasier.
Stay safe.

Virtual Power is a tool. It is not guaranteed “accurate” and is impossible to apply the level of accuracy that you are hoping to get from it.

Recognize the primary reason for VP existence is as a TRAINING TOOL. It is great for one user, one bike, one trainer, when they make sure to match setup each and every time (tire, tire pressure, roller pressure, and calibration if possible). In this case, you can get consistent and repeatable info, that is useful for setting training zones and executing workouts following specific instructions of power and duration.

VP is NOT something that should be used to compare to others. The use of “watts” is unfortunate, and leads to confusion between different users and even users that get a second power data source (like you did).

I suggest reading these two articles to get a more complete understanding of what it is and is not:

2 Likes

Why is it “impossible” to get this accuracy? The temperature and the variable stuff in a dumb trainer is THAT important? (Not the initial setup but the parameters that could change over a workout)

VirtualPower is consistent and repeatable, but the curve not being similar bothers me.


[Elite Super Crono Fluid included computer vs VirtualPower]

For example the Ramp Test protocol assumes some Watts increments at each step. Or does it work at whatever changing increment? If you start doing 10W increments and after some height it starts to differ because the actual curve is wrong (not just offseted) then does the ramp thing works as expected in finding your “VirtualFTP” ?

I can see you’ll fail at the same VWatt if you repeate that test. But I do not know what is that line you have just found. Is that actually my VFTP? Would that be the max VPower I can maintain for an hour if I’m chased by zombies ? (the fast kind)


I want to be able to learn the curve using a source of truth like a well known crank based powermeter. I doubt the environmental variations would affect the obtained curve so much. You can “trust” some aspects of the world like you not changing the tire in the middle of a workout and other controllable stuff.

Anyway … stuff I think about in the trainer.

It’s not impossible as a pure statement, and that’s not what I said.

These 3 scenarios (numbers added by me) are attempting to use the VP trainer as a “power meter” for purposes of comparison:

  1. This won’t work unless you have exactly the same tire model on both bikes. Even if you do, there is likely wear difference in tread and casing wear, that are likely to skew the virtual power reported.
  2. This is reasonable in essence, assuming this is on the same exact bike and setup each time. But regardless, the issue with one leg and single sided power are minor at best, and not worth the concern.
  3. This appears to be a way to “handicap” and equalize power between riders. I don’t really think it’s necessary and is again, somewhat beyond the scope of what VP was intended to do.

The point is that people often want to make VP act as far more than is practical. Yes, we have clear examples of rides who added power meters, and found it aligns with their VP. But we see far more examples where they diverge, and greatly at times. The variables in tires alone is enough to make this a guess at best. Most offer no calibration, which means there is little to any chance that you have a tire and roller pressure combo that “matches” the setup used by TR or Z when they generated their power curve.

It is all too variable to use in a wide reaching way as proposed. It is best used as a trainer tool for one rider and setup, nothing more.

Understood. Not giving the user the ability to build their dumb trainer (+ setup) power curves.
(In retrospective there are specific options to select supported trainers and even the resistance levels in each. That’s what threw me off. )

±15W? This is the same problem that all the indirect but smart trainers are already trying to solve. I’m thinking about that webpage that tests all the rolling resistances. :shrugh:

Cheers.

That is attempted to be “fixed” via the calibration operation. Getting the tire and trainer to operating temps first. Then do a spin down calibration, and the trainer reads that time, applies whatever magic the manufacturer has determined, in an attempt to make the reported values “accurate”.

The one example I know that is effectively calibration for virtual power is the Kinetic Inride sensor. It is still virtual power in essence. It improves in a way since it measures the speed directly on the trainer roller vs wheel speed. But most importantly, the Kinetic app allows for calibration of the Inride, similar to wheel-on smart trainers. So it is better than most virtual power that doesn’t include any calibration feature (which is most that I have ever seen).

So, there is a massive difference in virtual power and wheel-on smart trainers simply from the calibration option.

The time and money spent obtaining tires and trainers for that type of testing would be far more cost than a person could recoup. It’s old tech that serves a purpose, but is not worth deeper investment suggested, IMHO.

1 Like

So that’s what I want. I want a spindown calibration of my VirtualPower so the curve is better represented and (why not?) if I happen to have a separated source of truth I also want it to calibrate to that power. IMO It’s better than selecting names and resistance levels.

In-ride calibration? You could even know how each workout affects the virtual power readings.
Baxter does not do much but some “Insane” one is -2W at the end or something like that. #bigdata

I also think another consideration is that fact that PMs (Stages / Quarq/ 4iii/ and etc) as well as smart trainers themselves have issues. Just look at the number of forums or posts of certain smart trainers being ±15 watts or more compared to others. Then if you dont calibrate every ride same problem exists. I have a smart trainer and a PM and they track close but sometimes they are way off. Think about all the variables… Wheel on, direct drive, tire choice, pressure, temperature, type of PM, inherent difference between PMs and Smart Trainers, power curves (kinda like MPG in cars). I have also seen it where my Quarq and TR are tracking fine and inline with what my Garmin says but on Z the watts are 15-20 higher.

Each setup and however you determine your baseline is what matters. After that then as long as you keep your variables to your setup consistent you have a tool to base your training on. But you cant expect much more. Just my opinion based on personal perception.

1 Like