Here’s how I wasted my morning: I did a rollout test of what I guessed to be the best clinchers (lowest rolling resistance). I followed the example of a slowtwitch poster who rolled down a slight downhill that flattened at the bottom. I made a great effort to keep the same position in all the trials. The shortest rollout was about 250 feet, all others were measured from that point. So for example, the first Michelin Pro Light trial went 35 and 1/2 feet beyond the shortest trial.
Who wants to hazard a guess at what these results show?
Wow, that’s a lot of variability (even adding 250ft to numbers) within each tire. Interesting that the tire that generated the shortest run also had the longest.
I’d say the results are without statistical significance. You’d need to do hundreds of runs per tire to be able to make any conclusions.
Who wants to hazard a guess at what these results show?
Aztec, you’re about right. The answer to my original question is “absolutely nothing” – except, perhaps, that this test protocol (at least as I executed it) is not worth pursuing.
Get a couple of very long bungie cords anchored to the ground (maybe 20ft) to start your roll down and have someone hold you up (from falling sideways). That way your starting speed might be more consistant.
This might sound funny but put a training wheel on one side of the bike so you can start your roll down from a dead stop
Try to steer as straight as possible, maybe along a line on the side of the road.
One problem I see is that all the tires you are measuring have low rolling resistance, so you are trying to measure very fine differences with a protocol that is crude enough to likely only identify large differences. Just the slightest touch of wind is going to throw your results to hell.
I suspect the other tester appeared to show more repeatability for a couple of reasons, one of which being that the tires tested had much larger differences in rolling resistance, and perhaps the other tester more successfully found a test route that was sheltered from the wind and did it on a windless day. Than again, the other tests could be full of crap too, even the most minute change in wind between tire/wheel changes will introduce error.
Rolldown tests are valid and can work – but require excruciating control over conditions and data collection. There is a paper somewhere that validates the approach, but they did it indoors, and used very precise speed traps for capturing and normalizing velocity data. And they didn’t roll to a stop – they rolled through a trap.
Hello, At the least it points out that people who claim tire x is better than tire y because I cut off 2 minutes on the bike from last year using tire x are deluding themselves ;-). Your test was many magnitudes more controlled and showed, well, nothing.
I think you need a fully sheltered HPV to make it work. My top speed was about 18mph, which is high enough that fairly subtle position changes could affect outcomes (e.g. head up or down, shoulders shrugged, arms straight or bent). The problem with lower speeds is that subtle variations in pavement quality will also affect the results.
Have you considered using a set of rollers to do your test? If I were going to do such a test, I would use rollers, ride as close to e.g. 25mph as possible, stop pedalling, and time how long it takes to get to e.g. 10mph. Even with this approach there would be so much testing error that you would have to run it many times to figure anything out.
Have you considered using a set of rollers to do your test? If I were going to do such a test, I would use rollers, ride as close to e.g. 25mph as possible, stop pedalling, and time how long it takes to get to e.g. 10mph. Even with this approach there would be so much testing error that you would have to run it many times to figure anything out.
I think that may well be the way to do it. Instead of doing a rolldown, I’d just use the SRM to see how much power it takes to maintain a certain speed. Even if rollers exaggerate the differences between tires, I think we can assume that a tire that performs better on the rollers will probably also perform better on the road – though we won’t know how by how much. That’s OK though, I don’t want to quantify the RR; I just want to find the tire that has the least RR.