Zwift TT tests Take 2 - CdA relationship to rider weight - does this look right?

You’re exactly right, changes in speed don’t usually affect Crr or CdA (at least, at the speeds cyclists typically ride) which is why we can estimate CdA in a wind tunnel that runs at a different speed than most of us can usually ride. (The same goes for Crr).

When people used to do real field tests on real roads, they would try to control speed and power and slope and acceleration to make solving the equation of motion easier (usually, they’d try to control slope and acceleration to zero). From a practical point of view, that makes the experiment much harder in return for avoiding having to take into account the acceleration and changes in slope. It also restricted the choice of test venue. I don’t have a completely flat test venue anywhere near by, and by riding on roads that weren’t completely flat my power and speed weren’t constant. So I trade off a little more calculation for the equation of motion when I get home and can sit at my desk with a cup of coffee in return for a lot easier situation when I’m out in the field collecting the data.

3 Likes

This is turning into quite an interesting topic.

Quite apart from the specifics relating to Zwift the fact that it is a simplified environment means it can be used to demonstrate how to estimate CdA in the much more complicated real world.

It feels as if it would be possible to write a “How to estimate your CdA” article where the process is explained in the context of Zwift (which anyone can then replicate for themselves) then compared with an example from the real world.

This would also cover the implications of using differing approaches. I suspect I would not be alone in using flying laps as fixed power without realising its drawbacks.

So I will definitely do some more tests using the “Chung” method if only to check my understanding of it so as to use it IRL

Just to check one point. I last used this several years back. Have there been any updates to either the method or Golden Cheetah since? Anyway best if you point me to the latest version of your document. Thanks

1 Like

I also have a Rouvy account so once I have the Zwift data I’ll do another set of tests there as well.

1 Like

There are a few people interested in updating the Aerolab module in GC but real life events always get in the way. For example, it’s currently not possible to “mirror” an out-and-back run, nor is it possible to “stack” laps, nor cutting or snipping segments where you got passed by a car. I do all those things by hand, in a different analytical environment (I use R) and once I’ve done that I don’t want to re-import to GC. Eventually I’ll get some time and try to implement those things.

I’ve done more work on it, but I haven’t released for public consumption a newer version of that presentation–it’s already kind of long and there’s stuff in the slide notes that many people still haven’t understood (and my day job got more time-consuming) so the incentive to make it longer and clarify old things rather than work on new things was kinda weak.

But I’m taking a “sabbatical” next year so maybe that will be the chance to update it.

1 Like

Yeah, I did that for a comparison of Alpe d’Huez in Fulgaz vs. the Alpe in Zwift. Turns out they’re not exactly the same.