I did "Vandever" today - I am on a KK Rock and Roll, using the Garmin GSC-10 speed/cadence sensor. Power is measured by a Stages Power Meter.

I noticed that during each of the 4 minute intervals, I had to gradually increase speed (spin faster) in order to maintain constant power. My theory - during a hard, sustained effort, the fluid in the resistance unit heats up - viscosity decreases - resistance decreases. Therefore, I have to spin a bit faster to achieve the desired resistance - and hence wattage.

You can see in my Strava data that power is relatively constant - but both speed and cadence (same gear throughout the interval) had to be increased in order to achieve that constant power.

https://app.strava.com/activities/103549189

This could affect those that use Virtual Power. As their units heat up, the resistance will decrease at any given speed, making it easier for one to achieve the target wattage as the workout progresses.

Just found it interesting....

On your Strava activity when you graph speed on the "performance" graph it looks like it's tracking power almost exactly. I don't think there's anything comparable to a real power meter when it comes to doing intervals.

Just out of curiosity what value are you using for your FTP?

The change in speed is very subtle - about 1.5km/h increase during the 4 min interval.

My FTP is 301w.

I noticed that as well with my cyclops fluid 2. Virtual power is by no means perfect but it is still an improvement over only using a HRM. IMHO