Bkool Ghost race issue

[Barbes66]

Has anyone else come across this issue when doing a Bkool Simulator ride using only ghosts for company?.... so tonight I selected a bunch of ghosts within a narrowish time range (about 26-32 minutes) for a 12.5km route and had all 25 ghosts fly past me like I wasn't there on the first uphill section (which was most of the ride). The 'winner' came in with a time of about 10 minutes with everyone else similarly quick. After I trailed in a dismal last in about 28 mins I looked at the result on the Bkool app and found that I was actually only a minute or two behind the winner. WTF?? I've done this type of training many times before on Bkool and never experienced this which leads me to wonder if it's a bug relating to last week's release. Anyone else had this happen?

Thanks
Craig

 

[Brusgaard]

Has anyone else come across this issue when doing a Bkool Simulator ride using only ghosts for company?.... so tonight I selected a bunch of ghosts within a narrowish time range (about 26-32 minutes) for a 12.5km route and had all 25 ghosts fly past me like I wasn't there on the first uphill section (which was most of the ride). The 'winner' came in with a time of about 10 minutes with everyone else similarly quick. After I trailed in a dismal last in about 28 mins I looked at the result on the Bkool app and found that I was actually only a minute or two behind the winner. WTF?? I've done this type of training many times before on Bkool and never experienced this which leads me to wonder if it's a bug relating to last week's release. Anyone else had this happen?

Thanks
Craig

Probably V4 (or earlier) ghosts. Not really useful in BSIM V5.

 

[Barbes66]

Probably V4 (or earlier) ghosts. Not really useful in BSIM V5.

Interesting thought. If so, given the ghosts aren't version controlled, it's going to be hard to know who's who.

 

[Brusgaard]

Interesting thought. If so, given the ghosts aren't version controlled, it's going to be hard to know who's who.

Exactly, only way to tell is to look at the date for the specific ghosts ride, which is almost impossible to find due to private profiles or having to scroll through lots of pages on that users "Recent activity" page to find that ride.

Using your own V4 (or earlier) rides as ghosts will have the same effect. In BSIM V4 your watts were calculated (in the trainer I think) from the speed of the roller, and the simulator then used those watts to calculate the speed, taking into account your weight, on an ascent or a descent etc. Let's say your Bkool trainer reported 300watts in BSIM4, where in reality you only produce like 250w on a known good power meter. These (false) 300w watts were then stored in your old rides.

In BSIM5 the speed is calculated in the computer based on the watts/speed reported by you trainer, and the other parameters, and then some magic calculations to make the watts more real in BSIM (but not in Zwift etc.). This should (in theory) end up closer to reality, ie. the 250w.
A power meter or a good Direct Drive trainer, like a Tacx NEO, reports the correct wattage, which would then be 250w.

That means that your old ghost would be around 20% faster than you, even though you do the same amount of work. The flying ghosts, I think, in the simulator somehow trying to compensate for their supposed higher BSIM5 speed (given the 300w avg.), but ultimately fails. I think it gets confused by their finish time and the avg. watts, which doesn't correlate to the new speed calculation.

 

[Barbes66]

Exactly, only way to tell is to look at the date for the specific ghosts ride, which is almost impossible to find due to private profiles or having to scroll through lots of pages on that users "Recent activity" page to find that ride.

Using your own V4 (or earlier) rides as ghosts will have the same effect. In BSIM V4 your watts were calculated (in the trainer I think) from the speed of the roller, and the simulator then used those watts to calculate the speed, taking into account your weight, on an ascent or a descent etc. Let's say your Bkool trainer reported 300watts in BSIM4, where in reality you only produce like 250w on a known good power meter. These (false) 300w watts were then stored in your old rides.

In BSIM5 the speed is calculated in the computer based on the watts/speed reported by you trainer, and the other parameters, and then some magic calculations to make the watts more real in BSIM (but not in Zwift etc.). This should (in theory) end up closer to reality, ie. the 250w.
A power meter or a good Direct Drive trainer, like a Tacx NEO, reports the correct wattage, which would then be 250w.

That means that your old ghost would be around 20% faster than you, even though you do the same amount of work. The flying ghosts, I think, in the simulator somehow trying to compensate for their supposed higher BSIM5 speed (given the 300w avg.), but ultimately fails. I think it gets confused by their finish time and the avg. watts, which doesn't correlate to the new speed calculation.

Thanks for the explanation - that makes sense then if the v5 algo differs. I use a Drivo so my watts should be consistent across both Bsim versions (unless Bkool's magic stuff extends to adjusting real power outputs?) but as a wee experiment I will do a ride against my own v4 ghost on a similar terrain / distance course. Just need to make sure I eat, drink and sleep the same....actually no that's going too far!