At 4:24 PM -0800 11/4/08, Isaac Schell wrote:
>What do you use to test a tires speed? Is this a machine test or a
>human test? Why do pro's ride sewups if their slower? Am I
>misunderstanding this test or are you really busting a long time
>cycling myth?
It's not Bicycle Quarterly who busted this myth. Tests on steel drums have shown for a long time that tubular tires are slower than clinchers. In fact, they predicted that they were much slower. If we busted a myth, it's that tubulars are way slower. We found that they are a bit slower...
Our test was a rolldown test on a real road with a real rider. We used a carefully selected hill - very steep at first to get up to speed quickly, then decreasing gradient, so that the speed remains largely constant. During this "runout" we had a timed section. The speeds were moderate - too fast, and you are testing aerodynamics, rather than rolling resistance.
The advantage of this test is that it uses real-life conditions. In fact, the rider is crucial, because much of the energy of vibration appears to be absorbed in the rider. So a riderless bike will not be a good model for real-life conditions. A steel drum is worse, because it does not consider the ability of the tires to absorb vibrations. A bike that floats over surface irregularities is faster than one that vibrates a lot.
The disadvantage our test is that you need to control the conditions very carefully. Temperature and wind both affect the speed of our rider. (Tires roll much faster when they are hot.) Changes in position would be terrible. We carefully selected the day when we tested, and we tested the same tires several times to make sure we had repeatable results. Also, each tire/pressure combination was repeated at least three times. When we went to the wind tunnel for another test, we found that replicating the same position on the hoods was no problem.
Some have asserted that it would be impossible to control the extraneous variables, especially wind, and that our results therefore were of little use.
To evaluate this concern, we did a statistical analysis of all our results. A statistical analysis shows whether the results are just random chance, or whether they actually show differences between tires. For example, if you flip a coin and you get tails 5 times in a row, that can be chance. If you get tails 50 times in a row, you are likely to have a fake coin with tails on both sides.
For our test, you want the scatter among repeats of the same configuration to be much smaller than the differences between tires. Fortunately, my co-author has a Ph.D. with a minor in statistics - see
http://www.vintagebicyclepress.com/
The statistical analysis found that there was a less than 1 in 10,000 chance that we were just seeing random coincidences. So we are quite confident that our results are good. We could not detect small differences between similar tires (such as between the Deda 24 mm clincher and the Clement 28 mm tubular) - the follow-up article in BQ Vol. 5, No. 3 includes a chart that shows which tires are different, and which are too close to tell.
Anyhow, that is how we did our real-road tests of tires. Among many interesting results, we found that different tires resulted in 20% speed differences, between fastest and slowest. However, among modern racing tires, the speed differences were smaller, but still significant. Wider touring tires showed the biggest differences. Grand Bois' tires scored in the middle of the pack. The maker then used our results to redesign their tires, and when we tested the newer models, they were among the faster tires. Only the handmade Dedas and the Clement del Mundos were significantly faster.
The original article in Bicycle Quarterly Vol. 5, No. 1 was eight pages long, and there have been some follow-ups, so this can only be a very brief summary - especially since it is only marginally on-topic.
Disclosure: Vintage Bicycle Press sells Grand Bois tires.
Jan Heine
Editor
Bicycle Quarterly
140 Lakeside Ave #C
Seattle WA 98122
http://www.bikequarterly.com