Just How Far Off Is Virtual Power on Rollers

Hi All,

I’m Back on Trainer Road after a year of Track Sprint Training (kind of) and now without a Smart Trainer.

As a preface, i am ok with how training is going, albeit tough (now in week 2 of SSB LV1 after 4/5 weeks of just Z2 over December).

Yesterday i had 3x12mins at 85% and it felt really tough. HR was in to pushing up into Threshold from 8-12 mins of the effort.

This got me thinking so i had a look back to when i used TR at the start of 2022 and things look interesting.

At the time i used a Direto X Smart trainer. Now i am on Elite non Mag Rollers. Same bike.

FTP at the time was 240 (ramp on smart trainer), FTP now is 225 (AI detection after 5 weeks zone 2).

I have accepted my actual FTP wont be correct due to how virtual power works but i figured its all relative.

After yesterdays workout i had a look back to see what my power level and HR looked like during various workout types and it got interesting.

One particular WO had me at 106% FTP 5x5 with a HR of 185 at the end of each set (threshold HR). Yesterday i was almost at that after 12 mins at 85%.
Likewise a z2 ride at 75% ish with a HR of 150-160 in Jan on the smart trainer but now i wouldn’t get close to that, i’d be up over 160.

The level of training leading into that period was less than now so can virtual power really skew the power reading that much. It feels like the closer to threshold and over, the more my HR and power separate from what they were a year ok on the smart trainer.

I’m concerned i am actually at threshold even though the recorded power is high tempo/sweetspot.

I weigh 87kg and always check tyre pressures at 100psi.

How much harm could i be doing by sweetspot potentially being threshold work on SSB LV.
Might explain why after two weeks of SSB in Nov 22 i felt wrecked.
Im also pretty anaerobic which probably doesn’t help.

Any thoughts?

I don’t have any real answers, just variables that may be impacting results.

  1. Virtual Power Curves for any devices TR offers came from actual testing of a bike and power meter. Assuming you have the same exact model of rollers they used (not a generic option) the key variable in this relates to the tires.

    • The specific model they tested may well lead to a curve that differs from that of a rider using other tires. We know that rolling resistance can vary widely between the best and worst tires, and I have to imagine that the range there will impact results of VP. So, if you are using a tire that is different enough to what TR used for the generation of their device power curve, it could skew results when compared to real power.

    • A linked factor to the above is the tire pressure in use as well. Most often the recommendation is to use them at the highest rating listed on the sidewall of the tire. Presumably that is what TR did as well, but we really don’t know.

    • So this issue could well lead to different real power at any point in the speed range that drives the Virtual Power data. And the fact that it’s possible the slope and progression of your setup may differ means that there could be larger deltas at different speed/VP data points. No idea if this is the case for you, but some comments hint that it could be.

  2. Unless you have one of the roller models that includes a flywheel, another factor I could see playing here is the relatively low inertia of rollers. Most have nothing more than their own rollers, and a spindown from 20-25 mph can take mere seconds. This is compared to the usual double-digit times seen with most trainers. Some are lower around 15 seconds while others may take much closer to a minute.

    • It’s possible that your experience on the rollers differs compared to the Direto as a result of the flywheel differences between that and the rollers. We see people comment on flywheel inertia on smart trainers in ERG when using very fast vs very slow flywheel speeds (and the gearing that drives it). I have to imagine that the lower inertia of rollers is more like that of a super low gear on a smart trainer. But even that is likely to exceed the “help” that is often perceived from the flywheel.

    • It could be something that doesn’t matter at some power levels and durations, but I expect it may cause earlier fatigue and higher RPE even at the same real power values. All a theory of mine with nothing to back it other than general physics ideas, so I could be way off base.

All this to say that if the functional effect of the power curve is “off” enough, it could well be leading to poor training at one or more points within your training levels. Sadly, the only way to really know anything is to test in parallel with a power meter and your same bike, at the full range of power levels.


Hi Chad,

that is all a huge help and gives some context to what i have been experiencing.

As i am so new, just the extra riding will, im sure provide a benefit, but as you say, there are so many variables at play, its hard to know how close, or far i am vs where i should be for any given session, or even effort within a session.

Given my weight, tyre choice (Continental gatorskin 25), pressure choice (100psi), rollers, i feel i may be stacking the variables against me, and probably to the harder side of things.

I feel i may need to invest in a power meter for more optimal training given i may be quite far off where i should be.

What you said about inertia of rollers (or lack there of), is there any benefit or detriment to using rollers.
For the record my roller manufacturer and model match what TR have in their VP list.

Also, do yu have any tips for how to better gauge the efforts in my sessions to try and minimise the impact of VP perhaps putting me to high, or low. Can HR be used as a sanity check?

1 Like
  • That is a common choice generally speaking. We don’t know what TR used, but I’d guess it was something that might have a lower rolling resistance compared to the Gators. Could be worth an email to them to see if they can share that info, but I’ve never seen it mentioned.
  • Great question and one I have wondered about for a decade… and have no answer. I don’t think there are any studies on the training impact difference between inertia levels. It may well be individual with respect to a rider’s training needs (high inertia like road riding vs low inertia like climbing or off-road riding) as well as their individual physiology.

  • I’ve seen comments that high vs low inertia feel easier/harder or leads to better results. Just enough N=1 stuff along with my own experiences… and all I can safely say is that it usually “feels different”. I have no idea if it will lead to less than ideal results of someone trains in the “wrong/opposite” inertia vs their needs. But in a pure simple sense, I would aim to try an mimic my outside needs if at all possible with my inside workouts.

  • We may be talking about less than marginal differences here, so I would try not to lose sleep over it if this is the setup you have to use.

  • Excellent. That at least removes one of the possible variables from consideration.
  • Without power, it’s all a fair bit of guessing. HR could be a decent double-check, but we all know the variability it contains, so I’d use it with a grain of salt. The duration & intensity of different workout intervals may lead to HR being less ideal, but it’s not entirely terrible either. I look at my HR at times, but it’s largely just me cross-referencing it to my power data since I have it.

  • Adding in your RPE is at least as good as HR. Depending on your experience with training at various zones, you may have a feel for what those levels feel like and tweak your speed & Virtual Power to get more “accurate” training load. But it’s also far from concrete.

Thanks again for all the help.

I have little experience across power zones so that might a tricky one, although i can probably make a guestimate on whether i can hold certain powers for 20 mins which may be something.

At 85% sweet spot 3x12mins, the power felt like i may be able to squeeze it to 20 mins but that would’ve been a big push. I was fairly rested, had a good nights sleep and have been eating well too. That plus my HR being at threshold was a hint that i might be off on VP, and so lead to this thread. I may try and increase tyre pressure ever so slightly to see if i can get things to line up against expectations a little better. 12 mins low SS shouldnt feel like a fight to complete should it.

Regarding the low inertia of rollers, i too dont know if its good or bad. Compared to my Lemond, there’s more resistance, or rather, i need to engage more across the top and bottom of the stroke but i don’t actually mind that. I actually prefer rollers than a fixed trainer BUT for the power thing.

I will email TR and see if they can offer any help. If i can at least match pressure, or understand the ride/bike weight/tyre width they used, i can gauge which side of the higher/lower i may be.

Absolute figures are of no concern but trying to get as close as i can in terms of deviation from the power curve across the curve would be great.

1 Like
  • Yup, and I personally prefer this lower inertia. I use the 34t x 17t on my smart trainer for that lower inertia. Matches my needs outside more, but I also like the feel of having to engage more, especially in ERG mode. No idea if it matters at all, but it’s what I like.
  • Yup, I spent a season on motion rollers before making my first rocker plate in 2015, so I could still have some motion while gaining the benefits of smart trainers.

Your comments on the SS workout may mean either the power curve is off, or your baseline FTP is off (or both?). SS should be some work but I don’t think it should be pushing you where it seems you are landing.

Let us know what you learn from TR. Hoping they can offer some guidance here.

1 Like

Put this in the wrong place

Not much of a useful response from TR via the chat but they are going away to see if they can find any additional info.

I also sent an email asking the same questions so lets see what they come back with.

Any thoughts on tyre pressures to use, 100psi felt harder than it should.

sorry, i meant the effort level felt harder, as per my original post.

Is it worth trying 105 or 110psi to gauge how it feels at threshold and if its closer to what i would expect for that fort of effort.

I can make a better estimate of what i can hold for 20-ish mins vs an hour or longer.

Im thinking adjust the pressure up so power at my FTP feels about right. And then adaptive training take care of the above and below over time

Ah, ok.

I think you can consider tweaking pressure to find what “feels right”. But it’s guessing on top of an unknown, so hard to know what to expect. Probably can’t hurt to try and see how it goes. Reset at any point of you feel you are headed in the wrong direction.

1 Like


And sorry to the OP.

For some reason I replied to my thread on here.

I tried to fix this and missed one reply. But it’s harder to fix now, so we will just move forward from here.

1 Like

Would there be an issue with leaving it as is. Sure workouts may be really hard now, but at some point i assume my fitness/ability will catch up and things will move forward.

I just wont see much progression level change.

It may mean sweetspot workouts are more like threshold, assuming VP over reads so how bad would it be if i ended up with 3x threshold-y workouts a week on LV?

So hard to know. It may parallel the results from training with an FTP that is too high. Could mean that you are getting too much load, but that depends LOTS on your nutrition, recover and that long list.

I just am not the right person to answer that question.

1 Like

n=1 but rollers with a flywheel are an improvement that make it feel like the best of both worlds. The main detriment to rollers is the potential lack of resistance (or, well, falling off) and the flywheel helps with that without a mag strip.

1 Like

Unfortunately I am where I am with what I have so I’m trying to better understand how to make it work as well for me as I possibly can.

I don’t want to stunt progress by pushing too hard and being in the wrong zone but I also don’t want to undersell myself by making it too easy.

My instinct says put the pressure up a little, so sub threshold actually “feels” sub threshold and stick with that tyre pressure for the rest of the plan.

Given how hard 3x12 @ 85% was with 100psi (threshold 184bpm HR near the end. If I could’ve held on for 20 mins I think I would’ve been up near 90% max HR 195).

There’s no way I’d complete an O/U session like that.

Not always a fun thought, but you may want to perform an actual FTP test on the rollers. AIFTPD is great and works well for my use, but it’s entirely possible that your current FTP from it is not appropriate for the rollers.

Could be time to bite the bullet and see what you can really get on them to make sure you have the right foundation FTP.

1 Like

I’ve been doing everything to avoid that :joy:.

But you may be right. In fact, you probably are right.

Whilst the ftp from a test won’t be my real correct ftp, it will be relevant to the equipment.

Worth replacing an o/u session to do a 20min test?

  • Yeah, that seems like a decent swap to me.
  • This x 1000
1 Like

I wonder if the VP curve accounts for weight. Resistance is proportional to weight for rollers only, and not just rider weight but system ie. bike+rider weight. If not, this is the most likely culpit - OP is a bit heavier than the average bear.

@Trackdom, here’s Chad fielding essentially the same question from 3 years ago:

Get after that FTP test!


Indeed, all else being equal (same tires & pressure, rollers & diameter), riders of different weight may well experience different results from VP on rollers. I suspect TR may have just done a sample of one, but have not idea if they have done more testing with different tires and rider weights for any given rollers. It would take some effort to get that data and then contrast that for any potential differences.

Perhaps they did that and found it’s close enough to ignore, but I sure think there would be differences at the extreme ends at least.

1 Like