"Real world" testing

FLO Product Support, FAQ, Cycling, Triathlon, Etc.

"Real world" testing

Postby brimerican » December 8th, 2016, 11:43 am

I'm hoping this is not a duplication, but I couldn't find any posts or content relating to the comparison of all the cool FLO testing and theory to actual result on the road (or track). Of course it is extremely difficult to isolate test conditions and correct for the many and various variables that exist out on the asphalt; but it would be awesome to see some test results that confirmed the theoretical time savings that might be achieved on timed solo efforts (like 40 km or IM bike legs). Has anyone done this?

One way to run a test would be to get a group of riders with known FTPs to hold a fixed power output over a fixed course using a baseline machine equipped with a training wheel. Have them ride the course (20 km) several times, leaving at close intervals, recording as much information on temperature, wind etc. as possible. Holding as many variables constant as possible: tires, pressures, rider position(!), rider gear (esp. helmet), rider weight, gearing etc., etc. Then repeat using alternate wheel configurations. Again, performing as many repeats as possible, recording ambient conditions, and holding rider/machine variables as constant as possible (and as close to the baseline as possible). Even better, vary the power output with the same group. Assuming reasonable control could be achieved, and enough iterations performed, this type of test could yield some real insight into what racer can expect from their wheels alone.

The risk for FLO, and other aero-wheel companies, is that this type of testing might reveal that the computed savings overstate the advantage of fancy wheels. But the upside is that data showing real riders on real surfaces with actual bikes -- going faster by meaningful amounts would be very compelling. I just bought a FLO90 to replace my old Zipp404 rear but I don't think I want to run 20 time trials tests with both setups to try and find some statistical justification : ) [[I am an AG triathlete, albeit a slow one, and I seem to be overtaking a lot of disc wheel and 80-90 deep section wheel equipped bikes this year; so the marketing is working if nothing else]]
brimerican
 
Posts: 2
Joined: December 8th, 2016, 10:27 am

Re: "Real world" testing

Postby Chris Thornham » December 8th, 2016, 3:45 pm

brimerican wrote:I'm hoping this is not a duplication, but I couldn't find any posts or content relating to the comparison of all the cool FLO testing and theory to actual result on the road (or track). Of course it is extremely difficult to isolate test conditions and correct for the many and various variables that exist out on the asphalt; but it would be awesome to see some test results that confirmed the theoretical time savings that might be achieved on timed solo efforts (like 40 km or IM bike legs). Has anyone done this?

One way to run a test would be to get a group of riders with known FTPs to hold a fixed power output over a fixed course using a baseline machine equipped with a training wheel. Have them ride the course (20 km) several times, leaving at close intervals, recording as much information on temperature, wind etc. as possible. Holding as many variables constant as possible: tires, pressures, rider position(!), rider gear (esp. helmet), rider weight, gearing etc., etc. Then repeat using alternate wheel configurations. Again, performing as many repeats as possible, recording ambient conditions, and holding rider/machine variables as constant as possible (and as close to the baseline as possible). Even better, vary the power output with the same group. Assuming reasonable control could be achieved, and enough iterations performed, this type of test could yield some real insight into what racer can expect from their wheels alone.

The risk for FLO, and other aero-wheel companies, is that this type of testing might reveal that the computed savings overstate the advantage of fancy wheels. But the upside is that data showing real riders on real surfaces with actual bikes -- going faster by meaningful amounts would be very compelling. I just bought a FLO90 to replace my old Zipp404 rear but I don't think I want to run 20 time trials tests with both setups to try and find some statistical justification : ) [[I am an AG triathlete, albeit a slow one, and I seem to be overtaking a lot of disc wheel and 80-90 deep section wheel equipped bikes this year; so the marketing is working if nothing else]]


The closest thing we have to this type of data is all of the work completed by Ryan Cooper from Best Bike Split. Ryan is an optimization mathematician who has developed a model that predicts real world bike times with remarkable accuracy. Here is a link to Ryan's computations for a large number of our wheel combinations on a couple of popular Ironman courses.

http://flocycling.blogspot.com/2016/05/flo-cycling-aero-vs-weight-part-3.html
Chris Thornham
FLO Cycling: http://www.flocycling.com
FLO Blog: http://bit.ly/b3Zlpq | Twitter: http://bit.ly/bOQ08x | Facebook: http://bit.ly/d8KDhI
Chris Thornham
Site Admin
 
Posts: 576
Joined: March 2nd, 2012, 3:35 pm
Location: Las Vegas, Nevada

Re: "Real world" testing

Postby AlathIN » December 17th, 2016, 11:56 am

I'd love to see the results of the test you're proposing here, but as a researcher and all around research nerd, I am seeing some major methodology problems.

Given all the variables you mentioned, plus how difficult it is in real life to hold a precisely consistent power output for a sustained time, you would need a very large test sample to nullify the noise and isolate the contribution of the wheels. I know some people think their power meter demonstrates their machine-like consistent power output, but 99.999% of the time that is an artefact of how their system reports the data, not a true reflection of their robotic consistency.

Then you'd have to account for different combinations of tires, bikes, rider body shapes, positions, etc.

You could test in an indoor velodrome or similar controlled environment, but then you're right back to the question of do these results translate to real life riding.

Again as a data nerd, one of the things that totally sold me on Chris and Jon's wheels is how they go about asking and answering questions and testing their ideas. Based on my experience as a researcher, I think they're making the best possible use of in vitro and in vivo data. These guys are smart.
AlathIN
 
Posts: 11
Joined: November 9th, 2016, 7:51 am

Re: "Real world" testing

Postby Chris Thornham » December 17th, 2016, 12:40 pm

AlathIN wrote:I'd love to see the results of the test you're proposing here, but as a researcher and all around research nerd, I am seeing some major methodology problems.

Given all the variables you mentioned, plus how difficult it is in real life to hold a precisely consistent power output for a sustained time, you would need a very large test sample to nullify the noise and isolate the contribution of the wheels. I know some people think their power meter demonstrates their machine-like consistent power output, but 99.999% of the time that is an artefact of how their system reports the data, not a true reflection of their robotic consistency.

Then you'd have to account for different combinations of tires, bikes, rider body shapes, positions, etc.

You could test in an indoor velodrome or similar controlled environment, but then you're right back to the question of do these results translate to real life riding.

Again as a data nerd, one of the things that totally sold me on Chris and Jon's wheels is how they go about asking and answering questions and testing their ideas. Based on my experience as a researcher, I think they're making the best possible use of in vitro and in vivo data. These guys are smart.


Thank you. We appreciate the kind words.
Chris Thornham
FLO Cycling: http://www.flocycling.com
FLO Blog: http://bit.ly/b3Zlpq | Twitter: http://bit.ly/bOQ08x | Facebook: http://bit.ly/d8KDhI
Chris Thornham
Site Admin
 
Posts: 576
Joined: March 2nd, 2012, 3:35 pm
Location: Las Vegas, Nevada

Re: "Real world" testing

Postby brimerican » January 11th, 2017, 2:07 pm

Chris, thanks for the reply (and sorry for my tardy acknowledgement). The Best Bike Split tools are very interesting and I spent quite a bit of time playing with the drag coefficient and other variables on some of the case study data that Ryan makes freely available. Agree with 'AlathIN' on the many and various challenges associated with actually producing a defensible data-set from a 'reasonable to achieve' series of tests.

How cool would it be to have Tony Martin do 10 x 1 hour TT reps with Mavic training wheels and 10 with FLO90/DISC at the same power output -- and have the data show a 50 second advantage for aero-wheels? I can see the advertisement -- 9th place with FLO; 15th without ((TdF stage 13 timing as an example)). If you alternated the reps; std wheel/aero... Okay, I'll stop.

Huge fan of the work that you are both doing at FLO btw; and it is fun to see the company/products develop and evolve. When is the FLO TT frame coming out? ; )

PS I do a lot of my rides on TrainerRoad with my Kickr so I've been obsessing with Golden Cheetah, BBS's ERG export feature and ERG files in general. Not much aero advantage to be gained in the garage...
brimerican
 
Posts: 2
Joined: December 8th, 2016, 10:27 am

Re: "Real world" testing

Postby Chris Thornham » January 12th, 2017, 2:41 pm

brimerican wrote:Chris, thanks for the reply (and sorry for my tardy acknowledgement). The Best Bike Split tools are very interesting and I spent quite a bit of time playing with the drag coefficient and other variables on some of the case study data that Ryan makes freely available. Agree with 'AlathIN' on the many and various challenges associated with actually producing a defensible data-set from a 'reasonable to achieve' series of tests.

How cool would it be to have Tony Martin do 10 x 1 hour TT reps with Mavic training wheels and 10 with FLO90/DISC at the same power output -- and have the data show a 50 second advantage for aero-wheels? I can see the advertisement -- 9th place with FLO; 15th without ((TdF stage 13 timing as an example)). If you alternated the reps; std wheel/aero... Okay, I'll stop.

Huge fan of the work that you are both doing at FLO btw; and it is fun to see the company/products develop and evolve. When is the FLO TT frame coming out? ; )

PS I do a lot of my rides on TrainerRoad with my Kickr so I've been obsessing with Golden Cheetah, BBS's ERG export feature and ERG files in general. Not much aero advantage to be gained in the garage...


Thanks again for the kind words. We appreciate them. The TrainerRoad/Kickr combo is the best isn't it! I use it all of the time! No plans for a TT frame right now.
Chris Thornham
FLO Cycling: http://www.flocycling.com
FLO Blog: http://bit.ly/b3Zlpq | Twitter: http://bit.ly/bOQ08x | Facebook: http://bit.ly/d8KDhI
Chris Thornham
Site Admin
 
Posts: 576
Joined: March 2nd, 2012, 3:35 pm
Location: Las Vegas, Nevada


Return to General Discussion

Who is online

Users browsing this forum: No registered users and 1 guest