In Dec 2011 I compared Strava power estimates to Powertap measurements for the Cortland Hurl, a previously popular climb for SF2G. Unfortunately the Hurl has gone out of favor with the Google riders, but recently I did a ride with a leisurely spin up the climb, which has been an occasional fitness test for me.
New to the ride was my use of Garmin Vector, mounted on my heavy (10.5 kg) Winter Allaban. You'd expect a relatively high watts at a given VAM with the big bike with fat tires at relatively low pressure. So I was curious to see how the watts I reported on the climb compared to my previous trend.
Here's the result:
The data are sparse, and it would have been a lot better had my Vector ride been at a higher VAM. But it is strongly compelling that the Vector is reading low. The Strava estimates are certainly low in particular due to an underestimation of mass. The Strava method of asking for "bike weight" and "rider weight" omits the weight of equipment, which in my case typically includes a loaded backpack.
Here's the same data, but converted to an effective mass, which is calculated assuming all power goes into moving mass against gravity. Since I'm typically 57 to a historic maximum of 60 kg, obviously a substantial fraction of power is going into wind resistance and rolling resistance and, perhaps, some acceleration.
Fortunately, I finally got replacement batteries for my Powertap, so I'll need to put the Powertap wheel on the Allaban and see how the two compare, directly. The small complication is the need to use multiple Garmin head units, and pair each Powertap with a separate Garmin.