Strava Power Curve as perfromance assesment !

(Ivan Borcard (VO2C)) #1


Sometime I ride with Philippe Gilbert or Chris Froome. Some performance are 16 w/kg for 2’ climb.

Zwift will be a great plateforme for cyclist but you will never get ride of cheater.

As you have connection trough Strava API, It would be great to have some kind of certificate and correlation about your performance in W/kg and waht we call RPP ( Record Power Profil).

Kind Regards


Coach @ VO2 Coaching

(Duane Gran [Vision]) #2

If a person doesn’t have a power meter, what makes the estimated power on Strava a better authority than Zpower from a known trainer?


Probably what Zwift may need is a sort of self calibration routine for new riders where they answer some honest questions (because most of what is called cheating is simply misunderstanding) about how quickly they ride outdoors.  I can envision a ~5 minute protocol where they ramp up to a target power and answer whether it feels easier or harder than the same speed outdoors.  Do a few of these and it starts to become clear if you have under or over tightened the resistance knob on the trainer.


Most people want to do the right thing.  Some form of guiding them to better more predictable results is more valuable than trying to police when the wrong stuff happens.

(Francois Coppex [X][Velosophe]) #3

I kind of like the approach which consists in answering some honest questions. Indeed, from what I’ve seen most people with poorly calibrated trainers are just unaware their performance is incorrect. Therefore it may make sense just to ask a few basic questions in order to determine the user’s fitness like e.g. if you’re a casual or competitive cyclist, how many hours per year you’re training, etc. One idea to exploit that information may then be to correlate it in some way with e.g. Friel’s “Training bible” to get the cycling category equivalence (e.g. the suggested annual training hours table on page 124), which may further be correlated with Coggan’s “Training and racing with a Power Meter” w/kg table (on page 54) in order to define user-specific w/kg alerts (including the adequate adjustments for gender, age, BMI, etc.).

For example, if someone mentions he’s training about 200 hours per year (would typically fall into cat. 5) but one day sustains 5.5w/kg for 5 minutes (top cat. 2, bottom cat. 1) then he may get a Zwift alert asking about adequate trainer calibration. On the other hand, if someone indicates about 700 hours per year (cat. 2) and puts that same power for 5 minutes then the Zwift client wouldn’t raise any alert.

This approach doesn’t solve the issue of cheating or low amplitude miscalibrations, but at least it is an educational approach to all those who are not even aware their trainer is improperly calibrated.

(Ivan Borcard (VO2C)) #4

Ok my message was not to blame or to tell that they are all cheaters.

Zwift is a great application, but it will make sense only if the racers set there data or equipment properly.

You could also allow people to enter their personnal data et their signature would be public and seen realtime in Zwift.In addition to duration training per years, we could add also CP1 CP5 and CP20 if people know about their power profils.

Another topic it would be nice to have W/kg value in the leaderboard for jersey.


(Philip Amos) #5

One of the biggest ways to cheat is not having accurate weight.  How are Zwift suppose to validate peoples weight?

(Ivan Borcard (VO2C)) #6

I have seen KOM on Strava and leaderboard are really a mess ! More than 1000 watts for 1’ !!!

Yesterday I try to get 20 kg lighter et I had 2 jersey (KOM and green) in a row !!!

It’s a pity but this is really key in the success of Zwift I believe. Otherwise It will only be fun and a game.


(Philip Amos) #7

There isn’t really anything Zwift can do if people aren’t honest with their weight.  If you are a fat person its easy to remove a load of weight without looking suspicious.

I do think that people get a little worked up about people cheating but that is virtual life unfortunately.

There have been suggestions in the past that zPower isn’t exported to the FIT file so you can ignore people with poor trainer setup from the Strava leader boards and only compare yourself against people with smart trainers or power meters but i don’t think there is much interest in doing this based on it being easy and not implemented.