Virtual power reads 75-100 watts higher than perceived. 🤔

I am using two bikes for trainer road. One has a power meter, the other not.

I have a standard wheel setup with a giant fluid trainer. The watts readings are much higher than perceived effort. This is correlated with low heart rate so I know I’m definitely not pushing that hard. Does anyone have any solutions that doesn’t require me to buy a power meter.

I haven’t done an virtual power FTP test but I know it will be off.

Oh. One other thing. Using a Garmin non magnet speed sensor. Which I just took battery out and that didn’t help either.

Is there a reason you are using two bikes? I would recommend sticking to just using the bike with the power meter. If you do need to use two bikes, I would recommend taking an FTP test with your VirtualPower bike and having two FTP’s. Then you can change your FTP based on what bike you are using each day.


One is a TT that has the power meter and the other is a road. I really prefer to switch it up.

There has to be a reason that it’s off. I’ve been troubleshooting everything and I Think it’s either defective speed sensor or trainer vs possibly TR power curve.

Hoping someone has something else to offer.

VirtualPower is not accurate but it is precise allowing for you to use it to get faster. There is no way to get VirtualPower to be as accurate as a power meter as it is based on a speed to power ratio. There is most likely nothing wrong with your speed sensor or your trainer but you can try to change resistance level settings within the VirtualPower settings to get your power readings closer to what your power meter readings are. That said, there is no way to make VirtualPower match what your power meter reads.

I would still just recommend having a TT bike FTP and a road bike FTP to make sure you get the most out of your training on each bike.


Where can I dial in the resistance settings? I think I remember reading that for other smart trainers but I don’t recall that setting for my standard giant fluid trainer. I don’t think there is a drop down selection for that.

The reason I thought that it may be a speed sensor prob is bc when outdoors pedaling 30+ mph is about 350-400 watts. On trainer, that wattage is like 23-24mph.

Agreed on the dual ftp. That may be my fall back…booooo.

Thanks!Any other advice would be appreciated!

Ah that’s a bummer… Unfortunately, you will just have to have two FTPs that are pretty different then. The power curve for that trainer came from the manufacturer so we don’t have control over what the power numbers are unfortunately.

Sorry I don’t have a better option!

Even though I have the Giant trainer, I ended up randomly selecting the Tacx model fluid trainer with adjustable resistance option. Selected higher resistance level and that seems better. Will do an FTP on it and see but it ‘feels’ close. There’s hope maybe.

The Giant trainer power curve is def way way off. At least with my trainer.

There is a workaround. You can measure power with physical powermeter and with Trainerroad’s virtual power simultaneously, and then adjust wheel size parameter. You should always set the same tyre pressure. I’ve noticed maximum 3% error in average power to the Favero Assioma Dual readings. It can be implemented in software, I wrote about it in feature request forum.


Excellent advice @tadzikpl . I don’t know why TR wouldn’t offer somekind of calibration value to adjust virtual power. It’s not like with other platforms where you’re competing against others.

I did as @tadzikpl did and found that the percentage error changed according to whereabouts on the power curve I was. ie when set accurately for Sweet Spot it was inaccurate for VO2max workouts. I thought about inputting different wheel diameters for the different type of workouts but then realised I was getting a bit anal about it and just averaged it out.

My Elite Quick Motion rollers have linear power curve, and I’m only talking about accuracy of average power readings.

As some said above, I used to use virtual power with kinetic road machine before got a power meter and did some experiments when first got it to access how accurate virtual power was. For me it was pretty darn close at lower level power up to around sweet spot / threshold but virtual power massively over estimated at vo2 and higher.