Comparison - Bkool Classic (unpaired) with power meter/speedo on bike

[Asa Post]

Yesterday I planned to ride the Bkool as usual, but after the 3.11 update the trainer wouldn't pair with the software. I didn't have time to find a solution, so as the power meter on the bike had paired I decided to do the ride as an unpaired trainer. I wouldn't feel the changes in slope, but Bkool would read my power and adjust my speed to reflect the gradient. I had my Garmin on the bike, which was reading the same power meter and also giving me the real speed that I was travelling at.
Without any perceptible gradient, I was able to choose a gear and cadence and keep with it all through the ride. I couldn't generate lots of power, because there wasn't enough resistance to work against.
Afterwards, I uploaded the Bkool file and the Garmin file to my ride software, and compared them. As expected, the power readings are identical. It's the speed that varies.

The Garmin data is in red, and shows a fairly consistent speed, slowly dropping as I got tired. The Bkool data is in blue and shows the speed varying as it was adjusted for the slope. Most of the ride was uphill, so the calculated speed is lower than the real one.
The other two (smooth) lines show the distance travelled. The first 5 minutes are the same, as that part of the course was downhill/flat. Then the climb takes effect for the rest of the ride. I finished up pedalling an extra 17 miles, and though the power figures are below average for me I was knackered when I finished. I put it down to the unvarying pace of the ride, which doesn't suit me. It felt like a permanent head wind.
The course was Villardompardo - Escañuela - Jaen
I thought some of the mathematicians on here (hi @gbrown ) might find the comparison interesting.

 

[theboxers]

It would be cool if, when, you get yourself sorted out, you did the same course and compared the time and power graphs of the trainer and power meter.

 

[Asa Post]

It would be cool if, when, you get yourself sorted out, you did the same course and compared the time and power graphs of the trainer and power meter.

Your wish is my command

I did the route again today, with a fully-working Bkool Classic trainer . I tried to stay with my ghost from the previous ride as much as possible, to make the comparison more valid. It was ridiculously easy to do so - at some points I had to almost stop pedalling because I was getting too far ahead. I managed to stay within 30 seconds nearly all the time, and "won" by a few seconds on the line.

Here is the comparison of the two power traces:

For some reason, the Bkool trace (blue) starts before the Powertap (red), and I can't find a way of correcting it. As a result, the red line is a bit too far to the right, so you'll have to make allowances for the offset.
Although there are many places where the two power readings don't agree, the general shape of the lines is similar, and the average power for the Bkool is only 1 watt different - 114 to Powertap's 113.

What I find really interesting is the difference between the two Bkool rides. On my software, the times are only 8 seconds different, over the same course, but the unpaired ride has an average power of 136 watts, and the paired ride has only 114 watts. It just confirms that, while Bkool's average power estimate when the trainer is correctly paired to the software is pretty good, there are some other calculations that aren't as accurate.

Here's Bkool's own comparison - number 1 is today's paired ride, and number 2 was the unpaired ride:

I've lost 22 watts of average power, and I'm 4 seconds faster .