Event/Race categorization - FTP in watts vs w/kg

(Lisa Stone) #1

Forgive me if this has already been discussed and dismissed…

It would seem more logical to categorize based on FTP in watts rather than w/kg.  Many of the races are held on predominantly flat terrain.  While w/kg is going to determine speed on climbs, absolute wattage is what is going to determine speed on flatter terrain.  With the current categorization based on w/kg, lighter riders are penalized.  

In real world cycling, if a 65 kg rider has an FTP of 3.0 that would be 195 watts, vs. a 75 kg rider of 225.  On the flat, the 75 kg rider will be faster and have an easier time remaining in the draft compared to the 65 kg rider.  That seems to be how the Zwift algorithm works of the flat, as well.  Conversely, on a hill, they will be equivalent where the draft is going to be less effective anyway.

Ideally, an automatic categorization would be make based not he Zwift estimate of FTP which it already has data for.  W/kg would still be available to identify flyers.  It would just be the initial categorization.


(Steve Ellis) #2


I see riders unable to keep up with the pace of moderate w/kg group rides because they just aren’t putting out enough watts. They’re doing ~2.1 w/kg but that won’t keep up with the blob of riders where the riders in the draft are doing ~2.0 on the flats. They see a sub 2.0 ride advertised and they can do 2.0 but their 2.0 is not enough watts to keep up.

They need a ride advertised at 100 watts, or whatever.

(Andy Warhol) #3

Lisa, your broader point, I think, is a good one, that FTP, NOT FTP/kg, might be the better way to categorize riders for courses that are flat. I believe that it is not Zwift but rather the organizers of these events (users like you and me) that decide how to categorize riders. Then the Zwift programmer simply does as requested by the organizer. I organize two races, and I simply request that the Zwift programmer categorize the event in a particular way, and he always does.

On a related point, it is not a 0% grade that results in parity between two riders with the same wattage and different mass. It is something less than 0%. I am not sure of the exact value, but it might be -2 or -3% (a downhill). This is because a rider’s mass affects not only the pull of gravity down a hill but also the pull of gravity against the bearings (mechanical drag) and against the tires (tire drag), constantly squashing them against the road and generating more rolling resistance. Additionally, heavier riders tend to have a larger body profile and therefore more aerodynamic drag, although the height metric we enter into our Zwift settings might account for this.

All this is to say that on a dead flat surface, when two riders are putting out the exact same power, the heavier one (i.e., the one with the lower watts/kg) will tend to move slower due to the effect of body mass on mechanical drag, on tire drag, and on aerodynamic drag. Thus it requires some grade that is less than 0% (e.g., -2 or -3%?) for the effect of mass on the pull downhill to overcome the opposing effects of mass on mechanical drag, rolling resistance, and aerodynamic drag.

Given all of this, it MIGHT be the case that FTP/kg is still a better categorizer than FTP, but I am not sure. Maybe some experiments are in order. :wink:

(Lisa Stone) #4


Thank you for your thoughtful response. I definitely agree with your observation on a downhill grade regarding a heavier rider and a lighter rider.  I also agree with your observation about larger than average rider and the effect of aerodynamic drag. However, it seems like there is an still a relative advantage for riders of an average weight/size compared to a lighter/smaller rider in the pack. It seems like a higher w/kg is required to maintain the same position in the pack compared to the average weight/size riders in the vicinity.  I don’t know if the effect of the draft has a different effect relative to weight/size?  Maybe because some of the differential effect of aerodynamic drag is reduced.

If organizers try a few experiments with categorization, it might be interesting.

Thanks again