Oh my… I have not looked at this at all. Thank you for the insight… Yeah if it can’t be done in a reasonable quality fairly simplistic way then it might be too much for us but hopefully we can find a reasonable solution.
Yes this is correct, it’s a calculation of the fluid resistance from every trainer (at X speed sensor speed, how much resistance is X trainer putting out – and then converted into power). I don’t think the trainer companies give much information on this, pretty sure it was all done manually by the companies using virtual power, if I remember right
Golden Cheetah has Virtual Power and it‘s open source. Maybe you can get the power curves from them if you ask nicely.
Oh interesting. I am friends with Mark at GC thanks for the tip. Yeah I see some stuff in their code but looks pretty complicated. When the time comes to look at this more deeply I will reach out to him. He’s a funny guy in that he only responds to stuff he is really interested in but he has helped me.
Back in the days, VeloReality had most dumb trainers covered. When a trainer was selected, the formula for speed to power conversion was shown for every resistance setting. You could even add your own unsupported trainer. Accuracy is not very good, but as long as it is stable, you can use it for training. Just don’t compare with other athletes.
I like this idea. More and more I was thinking that this was just too much work for us but actually doing it that way makes it more feasable. They could connect to Strava or upload a max effort GPX/TCX file from outdoor to estimate their ball park FTP and then with a ramp test in the app we could set their virtual FTP and resistance… or something like this. Needs more thought but interesting ideas. Keep them coming. These insights are great.
Virtual power through cadence/speed would be awesome.
Thanks yes, lots of people want it and I am starting to see some people without power meters might use our app outdoors too with virtual power… Hopefully we can get this.
1 Like
I trained with Trainerroad virtual power for 2 winters and after that i got my power meters i think virtual power should be forbidden. It is not accurate and that is the point of training with a power meter. As i check my rides with virtual power i see that i just got esxhausted totally, for example for a z2 ride nowadays my MAX HR is about that my average HR was with virtual power.
I am not into heart rate based training but i think in this case it would be better ( with a sutiable fan ).
I agree with you here although your idea is a very good one (you indirectly suggested it). With virtual power we should require people to have a HR monitor and guide them through the warmup up each ride to try to get it close and maybe some RPE prompts or something. It’s definitly a lot of work to get it right. As my recent strugles with timing and bugs this one feels a bit more scary at the moment.
I have used Virtual Power for multiple winters as well and I agree that it is not accurate in the sense that you should never compare your virtual power with someone elses. But if you are aware off the pitfalls, you can do some decent training with Virtual Power. The key is to keep things consistent. Some off the problems are:
- Tension between tire and trainer roller must be kept as constant as possible. If possible, keep the bike on the trainer
- Tire pressure must be checked and adjusted equal every time before starting a workout
- Understand that Virtual Power can’t track acceleration power. The Speed/Power curve only accounts for stable speed/power. Short sprints will never show acceptable results
- Understand that heat has a huge effect on resistance. The air in the tire heats up more when there is more friction (resistance) and that heat causes higher tire pressure which in turn decreases friction/resistance. At the start of a high intensity interval you have to push harder and then it get’s gradually a bit easier untill things stabilize. So the calculated Power shows a steady interval, but in reality it isn’t.
Performance on longer intervals (around 3min and longer) can be tracked quite well with Virtual power if you keep above in mind. Virtual Power is usable for training but it demands more attention from the user to keep things consistent. You may have Power numbers that are 20% off from reality, but if it is consistent, you can use it.
All these pitfalls become very obvious once you have a Power Meter. Mind you, all points I mentioned will be equally present with a Wheel-On Smart trainer too! Fortunately, Trainerday app will correct everything with PowerMatch feature
Yes but as komivesg pointed out a lot of people will try to follow plans or workouts someone else created and seriously mis judge how hard they should be working. Also people are lazy so they don’t figure out all the inner workings to get it right. They just hop on and ride… It’s like you need to educate them as they ride. "This should feel very easy like you can do it all day… " You should not be breathing harder than walking right now. And standard RPE feelings. I think your perfectionism in doing proper training with virtual power would be the minority without guidance. Again I do feel connected to the virutal power users and would love to offer them a solution but I would like to be part of improving the quality of their training not pushing them over the edge 
Not really following you…
In the end, the principle of training with Virtual Power is exactly the same as with Real Power, only less accurate.
You do a 20min or a ramp test to eyeball your FTP and follow the exact same training plan. Virtual Power is accurate enough to ride Endurance/Tempo/Treshold. Not HIIT though…
Or are you talking about simple “health” training without any performance goals?
No, it is no accurate.
Maybe i feel it can cause overtrain because of trainerroad sweetspot fetish.
I train with power meter now and trainerroad plan that focus on sst, i thinking about this and this is for who have a basis and can manage this load.
But who try to train with virtual power i think needed guide and help not to overtrain.
My experience says a lot of people are over achievers and try to do more… If they are 20% under the target it will be easy and they will re-calibrate it higher. If they are 20% over they will just think they are out of shape and kill themselves trying to finish that “105%” sweetspot workout (that should have been 87%). As you can see I take the opposite approach of “sufferfest” and feel that people should not suffer too much, just the perfect amount of suffer on the right days even serious cyclists.
I fully appreciate what you are saying regarding FTP test and I realize this same problem exists with wheel-on-trainers to a smaller degree but at least with a wheel-on-trainer you do have a spin down process to get it in the ball park each ride. I know I basically know nothing about this and you have a lot of experience just voicing my concern.
I agree with what you say on Over Achievers. It’s part of the FTP hunting for bragging matters… Sad but very common.
But I’m still convinced that Virtual Power can be used fairly good. If you have a Power/Speed curve from the manufacturer, it will be within about +/- 10% accurate. That’s not really acceptable, I know. But the important part is that it is consistent. If you can borrow a Power Meter, you can make the Speed/Power curve yourself. It only takes about 5 different speeds per resistance setting. Note the Power for those speeds and fit a curve to it in Excell. Most older resistance trainers have an exponentiel curve. I had a Elite Mag trainer back in the days and thought my FTP was 230W. When I bought my Vortex it turned out te be 205W. When I bought my Assioma pedals I knew it was 190. So the more equipment I bought, the lower my FTP was 
Another possibility is to find someone with a Power Meter and the same make/model of your resistance trainer, willing to do it for you. Most of the formula’s you find in software using Virtual Power were made like this. If you happen to find 2 different formula’s for the same trainer and plot them, you will see the exact same curve but shifted up/down for a number of watts. It’s simply the manufacturing tolerance that causes this differences.
What the spin-down calibration on wheel-on smart trainers does is exactly that: it checks the time between two speeds during spin-down and shifts the curve up or down according to the result. It is a digital way of compensating manufacturing tolerances. On a Vortex smart, the curve is linear. I checked that while the trainer was unplugged. Don’t know about other trainers. But the lower priced Smart trainers still use the same principles of Speed/Power conversion be it in a smarter way. The tolerances are better and the spin-down test allows to offset the Power number they spit out.
The more expensive trainers (mostly wheel off) are factory calibrated using a real Power Meter and an automated process. That + the absence of a wheel - roller interface make them much more accurate.
Sorry for the slow response. I am on vacation so doing minimal. Your insights really help me. You are right in the fact that wrong is wrong and regarding FTP / trainer accuracy. Probably most people have are ±10% which if you are trying to follow a workout prescribed by someone else 10% is a lot when it is around threshold. When it is over threshold W’ comes into account so FTP based is wrong anyway So everythning is just wrong… and FTP generally just gets you in the ball park. As @Jeremy complains ramp test is off for many people which screws up sweet spot workouts. Makes them too hard. I did get the manufacture power curves for a lot of trainers from a friend.
So based on what I am hearing you say is if I want to help people it’s not about making their trainer accurate as much as it is more about guiding them how to relate power to RPE (and possibly HR) and they should make sure they are getting the right RPE. Secondly maybe trying to get their trainer and FTP testing fairly consistent so once they found their right workouts they can just focus on increasing power.
Hi Alex. Taking some time off from “work” is just as important as taking the right amount of recovery during training, so no need to apologize. Feel free to read this quite long post when you feel ready for it
It would be a pitty if interested users move away after reading that Virtual Power is useless, should be forbidden, can’t be used, etc. After all low budget people is part of your intended audience. So i’m taking the time to put things in perspective
I’m afraid you’re still not getting my point about Virtual Power. For your personal training, the accuracy, and I mean the real definition of the word, is not important. Consistency, on the other hand is extremely important. If your dumb trainer reports 600W Virtual Power as FTP, so be it. As long as you can reproduce the same number, it gives you a reference point. If after training hard FTP based for a couple weeks, you retest and have 610 watts: congrats, you have improved.
Accuracy becomes important if you want to compare to someone else or if you want to train on other equipment. I doubt that you will be able to use the 600w as a reference FTP on a Tacx Neo 
And I know from experience that it is possible to get consistent results from a dumb trainer if you are willing to set up and follow a routine before riding. I don’t have to tell you how important such a routine can be because you developped a HRV app. Anyone using HRV seriously is following, religiously, a routine for morning measuremant. If you don’t, the data is useless.
And that is what you should do for your trainer. Setup the bike always in the same way. Same tension on the wheel axle, same tension of the resistance unit against the wheel, same tire pressure. If you keep those three things consistent, you will be able to measure progress and train almost as good with that old second-hand trainer compared to a mid-range smart trainer. After all, the only thing that these smart trainers do better is adjust the resistance for you, by means of electromagnets. While it is up to you to keep your speed correct on a dumb trainer.
People need to understand that smart trainers all work with the same principle of speed/power conversion. There are, as far as I know, no trainers with a built-in Power Meter. Not even a single strain gauge… And no, the Neo hasn’t got a power meter eighter. You can check that with your friend at Tacx. What the Neo has is:
- A consistent bike-trainer interface (chain - cassette - axle)
- A unit specific calibrated speed/power curve
- Better (more expensive) mechanics (bearings - low tolerance parts)
- Low tolerance electromagnets
And believe me, the first two are making the big difference.
Taking my Vortex as an example, the formula for speed/power conversion would be something like:
P=av + bv + c where P = power, v= velocity (speed or rpm of the roller on the unit)
The factor a is accounting for the resistance created by the permanent magnet. And that is influenced by the distance between the flywheel and the permanent magnet. Making it a first source of error because there will be tolerances during manufacturing…
The factor b is accounting for the resistance created by the electromagnet(s). It is the adjustable part because current sent through the electromagnets regulates magnetic field and thus resistance. Second source for error because the tolerances on these electromagnets make for difference between the units.
Factor c is a simple offset calculated by the spindown test. Sort of X seconds spindown equals Y Watts more or less than “standard”.
And “standard” is almost surely an average of a few dozen units tested with a real power meter. That value is accepted as the “spec” for this model with a tolerance of +/-5%.
You see that it isn’t too difficult to accumulate a number of errors because there is no loopback to check any of those.
This doesn’t mean that the Vortex is a bad trainer. It simply means that here too you have to pay attention to set it up as consistent as possible. Certainly for people who train by Power with only the trainer as Power source. If you have a separate Power Meter and use PowerMatch, it becomes less important because the app corrects it for you 
So to wrap up: if you take the route of Virtual Power, make sure to Pin a post with some guidelines how to properly use it. People following those guidelines should not be affraid of FTP based training on a dumb resistance trainer.
And I hope I’m not upsetting to many people with this post 
1 Like
I’d like to +1 for virtual power. I’m currently considering switching from Trainerroad (dumb trainer, speed and cadence sensor) to Trainerday for financial reasons; literally the only reason why I haven’t is because buying a £100+ power meter would completely negate the financial benefits of switching 
Fyi yes I do understand all the limitations of using virtual power (my masters degree is in physics) but for me, it’s helped me to get into structured training on a very low budget, where I wouldn’t have been able to otherwise. I don’t really care about my FTP in absolute terms, I just want to be able to do an interval at e.g. roughly 120% of FTP, which I can do using “virtual power” and making sure my bike stays on the turbo so the tension is the same (I rebuilt a very cheap old bike frame specifically for this) and keeping the tyre pressure constant.
1 Like
Hello there, yes I fully agree with you that absolute FTP does not matter and relative FTP is fine. I have fully warmed up to the idea of adding virtual power and see it’s virtues. I started this project because I felt there needed to be more affordable ways to train so virtual power fully fits that.
I estimate it to be done in the next 60-90 days (before winter for sure). If the stars align and this takes priority over another big feature then it really will be closer to 60 and possibly even less. I would love to have you as a customer, and love for you to not have to spend extra dollars to do so.
1 Like