Originally Posted by
Spoonrobot
Are you able to see the link and graph in post 100? I mean that's it, that's the demonstration, on smooth and moderate asphalt the rolling resistance goes down as pressure increases to the breakpoint. It falls apart on gravel but there is no "good" testing protocol for loose surface riding, yet IMO. Everyone is extrapolating rough surface riding from what is essentially data from pavement. I suspect the major mountain bike XC tire companies have some very eye opening data but they're not sharing anything other than accidentally or very obliquely.
The biggest thing is that there is no breakpoint for the roller so it requires judgement and supposition to determine how relative higher pressures stack up compared to the suppleness of the tires. I wish BRR would have recorded sag values for all tires as it would make comparisons for off-road much easier. A supple tire is going to have a much higher real world breakpoint that a less supple/stiffer casing tire.
That is one of the many demonstrations, I recommend reading the Silca blog links as it goes into a lot more detail.
This paints a slightly different picture but it still very interesting:
https://www.bikeradar.com/road/gear/...-tested-49101/
What methodology are you using? I have a smooth flat section of pavement and another section of consistent gravel that I perform "rundown" tests on. I ride to 16 miles per hour and then coast to a stop in the same position, record where I stop relative to a known marker and extrapolate from there. I make sure to only do the testing for tires on the same day and about the same time, it doesn't necessarily control for temp/wind but I can usually get it pretty good and then practical experience seems to be consistent with what my testing shows. It's coarse but it seems to work ok.
For example, I wasn't able to determine any difference between Grand Bois Hetres and Compass Baby Shoe Pass tires despite being different weight, size and construction. But it was shockingly obvious that Panaracer Gravel King slicks were much much slower despite having extremely similar characteristics as the BSPs.
I didn't like the
Tour test because they ran the 35mm tire at 87 psi and the 25s/26 at 101 psi. That's way too high for the 35, the tire max is only 90 psi! They should have ran it closer to 45 psi. I think the pressure they used significantly overstates the resistance on the rough asphalt and understates the resistance on smooth asphalt.
Another interesting wrinkle is Tom Anhalt's testing. He's mainly focused on gravel with a triathlon bent but he seems to be finding a lot of interesting data on some of the Compass tires. You can see some of his comments here, they help add another well educated and robust voice to the discussion:
https://janheine.wordpress.com/2018/...-in-the-world/
This is another good thread with Tom A. input - start at post 100:
https://forum.slowtwitch.com/forum/S...63727/?page=-1
Personally I've also found Panaracer tires to vary widely in fit/finish which can have a profound effect on rolling resistance. I would really not be surprised if this ended up being the case for the BRR test, a bad tire, relatively. What's interesting is that Mr. Heine is probably well aware of this as he had a run of contract Grand Bois tires that ended up showing up and testing terrible despite close work with the factory to produce what he wanted. They were cleared out for a pittance and apparently forgotten.
Anyway, this is a very interesting topic for me as more than 1/3 of my yearly riding mileage is either timed events/races or rides 100 miles and longer. Having good fast rolling tires is one of the most important factors in my enjoyment of the sport. I have opinions formed by experience but generally try to keep an open mind and learn without too much ego.
I have read the links you posted to (in the past), except for the Slowtwitch discussion.
But I keep coming back to the same point I made before regarding using drum roller tests to rate rolling resistance in real world conditions.
In regards to the roller drum tests, the chart you link to as well as the study it comes from does indeed show that (up to a point) higher pressure gives lower rolling resistance for a given tire. But that is really it.
However, that is not really what people are most interested in here, or at least not what I was commenting on. What is of interest is comparing different tires and ranking them against each other, so people know which tires are faster than others in real world conditions. Some quantifiable metric would be more helpful, but ultimately we are interested in comparisons.
And with this in mind, that study is NOT demonstrating that a roller drum test will predict how tires will behave on real world surfaces, in a way that can be used for comparisons. The graph shows that even for pressures below the breaking point, real world RR values are different than roller drums, and the rougher the road, the bigger the difference. That in and of itself might not be a problem IF it could be established that
the differential in RR between the roller drum tests and real world conditions were consistent for all tires, but I have yet to see such evidence. Until that is established, I do not think that claim that roller drum tests can be used for comparisons between tires can be defended very well. Perhaps I have missed such a study. If so I would be interested in seeing it.
I have read Tom A's ideas on why the roller drum test SHOULD be translatable to real world conditions, but again, not seen it tested.
I feel that conclusions around the roller drum experiments fall into a common trap that many areas of research do... that is, deciding data is useful based on the fact that it is accurate, precise, reproducible, and easy to work with. Nobody can deny it is wonderful data to work with. You can fairly quickly get wattage numbers that not only lets you compare tires, but you can quantify the differences, and even use the data for calculations..... However, t
hat does not mean that the results answer the question you are interested in..
There is a saying I read once that goes along the lines of "
too often we value what we can measure, rather than measuring what we value". Roller drum measurements are very accurate, precise, and reproducible, so we assign great value to them... however, what we SHOULD value is real world conditions, and give more weight to experience and experiments that measure that directly, even if the results are messy and often inconsistent. In other words, messy and inconsistent data measuring exactly what you want to know is better than clean and tidy data measuring something that may or may not be relevant.
So when you ask me what my methodology is, I don't mind saying that it is simple subjective feel..... but it is subjective feel of the exact thing I am interested in (how fast I roll on the tires). And i will take that over high quality results from an experiment testing something else for which the relationship to what i want to know is unclear.
For all the drawbacks, Roll down tests like you describe are IMO far, far more useful than roller drum tests for determining a rank order for tires RR in various real world conditions.... because it IS real world conditions. And if properly designed roll down tests on various surface conditions confirm the ranking that roller drum tests indicate (including Bon Jons) I would accept that.