Cycling News tire test

[benb]

Originally Posted by echappist

Thanks for posting.

What is particular interesting is that some tires that were comparable to cohorts at one speed are slower at higher speeds (see GP5000 AS TR vs its immediate neighbors on the chart)

Interesting but where are the error bars and other information? It is fine if the journalists keep doing these tests but the problem might be when they extrapolate the results to mean more than they actually do based on the quality of the test or lack thereof.

If the error is +/- 5W a lot of the chart is bogus and the GP5000's medium/high speed resistance compared to the peers in the chart might not be a real thing. Do they even know the error range?

Also how many tires of each model did they test? There could be variance between different tires of the exact same model & size.

They always focus on this particular test... I want to see tests like how long the tire lasts with a rider of a particular weight riding it at a particular average speed, etc.. along with cost information as well.

And maybe some kind of tests about puncture resistance?

I really like the S-Works Turbo tires (non-cotton). Their flat resistance vs durability vs rolling resistance seems to be a real sweet spot for me. I would not care about a tire being 5-10w faster if it wore out 25% faster or gave me a 50% higher chance of a flat. I have had zero flats on the S-Works Turbo tires from 2016-2024. (Knock on Wood)

What happened to tests regarding lateral grip or braking?

[jadmt]

I just bought 2 conti 5000’s in 700x32 and there is 20 grams difference in weight

[Mark McM]

Originally Posted by meyatt

maybe, in my experience it's slower and harsher on gravel and road (I'd say this is purely anecdotal but based on the frequency of particular roads I ride on, I feel like I've some quantitative backup).

The SRAM pressure calculator tends to be more on the money for me.

Plus, increase the pressure, increase risk of punctures — the thing that really slows you down.

As gravelreformist said, the Silca pressure calculator is intended to optimize for minimal rolling resistance (speed), and not comfort and handling. In an episode of the Marginal Gains podcast, Josh Poertner said that the Silca pressure calculator will generally produce a higher pressure than most people would want to use.

Testing has shown that rolling resistance asymptotically decreases as pressure increases until it reaches a "breakpoint" pressure, after which rolling resistance starts increasing quickly. The Silca pressure calculator attempts to provide a pressure just below the breakpoint. In other words, the Silca pressure calculator provides the maximum pressure a user may want to use. So the Silca pressure calculator really just provides a starting point for finding the best pressure, and the user would then incrementally lower the pressure until they find the best compromise between comfort and performance.

[Mark McM]

Originally Posted by benb

Also how many tires of each model did they test? There could be variance between different tires of the exact same model & size.

On the Marginal Gains podcast, Josh Poertner talked about Continental doing rolling resistance tests on their GP5000 tires. For production efficiency, Continental has multiple molds for making these tires, and they found that the tires that came out of a particular mold were noticeably faster than the same model of tire that came out of the other molds. However, they chose to keep the identity of this mold secret, lest customers try to get only tires that came from this mold and reject tires from the other molds. (They did however supply tires to sponsored teams from the "fast" mold.)

Originally Posted by benb

What happened to tests regarding lateral grip or braking?

What happened to the tests is that they are almost non-existent. If you think rolling resistance tests are difficult to do precisely, traction testing is even more difficult - and they are particularly hard to do if you try to replicate real world conditions. Far more organizations have done rolling resistance testing than have even attempted traction testing.

That being said, some organizations have done some traction testing. For example, Bicycle Rolling Restance does a semi-static lateral traction test. Avocet also built a machine to do traction testiong. In this particular case, Avocet was attempting to demonstrate that their smooth treaded tires had more grip than tires with patterned treads. Their test setup was similar to many rolling resistance tests, in which a tire was pressed against a rotating drum with a fixed vertical load on the tire. But for the traction test, the drum was able to be tilted, in order to create a lateral load on the tire (as when leaned over in a corner). As the tire was rolling on the drum, the drum was slowly tilted. A load sensor measured the vertical force on the tire, and the the tire's lateral traction limit (ean angle) was reached there would be s sudden drop in the force on the tire. (And yes, Avocet's testing showed that their smooth treaded tires had more lateral grip than other companies patterned tread tires.) Avocet then verified their test results with road testing, using a bicycle instrumented with a tilt sensor. The lab tests showed that their tires had a maximum lean angle of about 45 degrees, and their test rider was able reliably corner with a lean angle of about 40 degrees without losing traction.

One reason that real-world bicycle tire traction testing is so difficult is that bicycle tires (or at least narrow, high pressure, thin treaded road tires) exhibit very little slip angle. This means there is little transition between solid cornering traction and near total loss of traction. Anyone who has had their tires slip out while cornering on hard pavement will tell you that at one moment traction was fine, and then in an instant they were on the ground. So pushing the limits of cornering to find the ultimate limit of traction can often lead to crashing. Given that caveat, a few organizations have attempted to do these tests.

Tour magazine used to perform traction testing and publish the results. Their test measured wet traction. For their test, they set up a constant radius corner on an asphalt surface, and used a sprinkler to keep the pavement constantly wet. A test rider on a special "bike" would roll down a ramp and then execute the corner at successively higher speeds until the tires slipped out. They then published the maximum cornering speed for the tires under test for comparison purposes. (The "bike" was actually a purpose made vehicle that resembled a push scooter with rider standing on a platform, but it had the same wheelbase and steering geometry of a standard bicycle and used standard bicycle wheels. The test rider wore a full face helmet and lots of padding for protection, as the vehicle would usually crash when the tires slipped out. I believe the push scooter design was used so the test rider could attempt to jump off and land on their feet when the tires slid out.)

Michelin also attempted to build a bicycle tire traction vehicle. This vehicle resembled a standard bicycle, except on one side it had an outrigger, almost like sidecar, except that the wheel on the outrigger was a few feet off the ground. The wheel was on a caster, so it could roll even if the bike was sliding sideways. Normally, the outrigger wheel would not touch the ground, even when leaned over in a corner, but if the tires slipped out the bike would "land" on the outrigger wheel instead of falling over flat to the ground. The outrigger also had a net to catch the rider and keep them from hitting the ground if they fell off the saddle. I've seen pictures of this test bike, but not heard of any results from its use.

[jadmt]

Originally Posted by Mark McM

On the Marginal Gains podcast, Josh Poertner talked about Continental doing rolling resistance tests on their GP5000 tires. For production efficiency, Continental has multiple molds for making these tires, and they found that the tires that came out of a particular mold were noticeably faster than the same model of tire that came out of the other molds. However, they chose to keep the identity of this mold secret, lest customers try to get only tires that came from this mold and reject tires from the other molds. (They did however supply tires to sponsored teams from the "fast" mold.)



.

funny because I just posted above that I bought two conti 5000's and one was 20grams lighter but it also appears to have a thinner tread. My guess is this tire would indeed be a little faster. both have very recent date codes.

[benb]

If they have molds that are fast and others that aren't you'd probably want to see these tire test websites test multiple examples.

But it seems like it might be tricky to actually find one from the "fast" mold.

I mentioned the other types of tests that we don't see partly because I feel like over the years it has not been hard for me to feel a difference in stopping, cornering, or "feel" as you increased cornering angle between different tires. There have always been some I felt were noticeably better than others. Even if it was just feel it effected my confidence and would make me not want to buy a tire again if it was one of the ones that felt strange.

If one of these tires tests "fast" but it's marginal and yet it feels odd under hard braking or something there's no way I'd buy that tire.

[gravelreformist]

Originally Posted by benb

If they have molds that are fast and others that aren't you'd probably want to see these tire test websites test multiple examples.

But it seems like it might be tricky to actually find one from the "fast" mold.

I mentioned the other types of tests that we don't see partly because I feel like over the years it has not been hard for me to feel a difference in stopping, cornering, or "feel" as you increased cornering angle between different tires. There have always been some I felt were noticeably better than others. Even if it was just feel it effected my confidence and would make me not want to buy a tire again if it was one of the ones that felt strange.

If one of these tires tests "fast" but it's marginal and yet it feels odd under hard braking or something there's no way I'd buy that tire.

Bicycle Rolling Resistance has tested so many versions of the GP5000 that we can be pretty confident in their numbers. The results are repeatable enough that even if there are manufacturing differences, it's probably not more than a watt or two maximum.

[Mark McM]

Originally Posted by benb

I mentioned the other types of tests that we don't see partly because I feel like over the years it has not been hard for me to feel a difference in stopping, cornering, or "feel" as you increased cornering angle between different tires. There have always been some I felt were noticeably better than others. Even if it was just feel it effected my confidence and would make me not want to buy a tire again if it was one of the ones that felt strange..

All too often, our feelings can be fooled. A tire that may "feel" fast might actually be slower than one that "feels" slow. It took a long time for racers to move to wider tires and lower pressures, because the narrower tires and higher pressures "felt" faster to them. likewise, a tire that "feels" like it has more traction may or may not actually have more traction. Keith Bontrager tells a story from his motorcycle tuning days where he would make a tuning adjustment on a motorcycle and have the rider test it out. Sometimes the rider would report that they went slower after the adjustment, even though the actual lap times were shorter.

On the other hand, confidence plays a large role in how far we are willing to push something. A tire which "feels" better when corner may give a rider more confidence to corner harder than they might on another tire that doesn't feel as good when cornering.

[EB]

Originally Posted by Mark McM

As gravelreformist said, the Silca pressure calculator is intended to optimize for minimal rolling resistance (speed), and not comfort and handling. In an episode of the Marginal Gains podcast, Josh Poertner said that the Silca pressure calculator will generally produce a higher pressure than most people would want to use.

Testing has shown that rolling resistance asymptotically decreases as pressure increases until it reaches a "breakpoint" pressure, after which rolling resistance starts increasing quickly. The Silca pressure calculator attempts to provide a pressure just below the breakpoint. In other words, the Silca pressure calculator provides the maximum pressure a user may want to use. So the Silca pressure calculator really just provides a starting point for finding the best pressure, and the user would then incrementally lower the pressure until they find the best compromise between comfort and performance.

There is something very wrong with the Silca tire pressure calculator when it comes to mountain bike tires. It gives laughably wrong answers, like 12 psi for a rear 2.4" tire for a 208 lb human. I have no idea what it does for road tires, but this has never given me any trust in that thing for any use case.

The Zipp calculator at least gives answers in plausible ranges.

[MikeD]

Originally Posted by EB

There is something very wrong with the Silca tire pressure calculator when it comes to mountain bike tires. It gives laughably wrong answers, like 12 psi for a rear 2.4" tire for a 208 lb human. I have no idea what it does for road tires, but this has never given me any trust in that thing for any use case.

The Zipp calculator at least gives answers in plausible ranges.

It works very well for road and gravel, at least in my experience. I've not used it for mountain biking, preferring experimenting on my own. I plugged in the values for my mountain bike and yes, pressures are way lower than what I'm using. But, at least Silca has some science behind it, and is adjustable for road/trail conditions, unlike the other calculators out there.

[Mark McM]

Originally Posted by EB

There is something very wrong with the Silca tire pressure calculator when it comes to mountain bike tires. It gives laughably wrong answers, like 12 psi for a rear 2.4" tire for a 208 lb human. I have no idea what it does for road tires, but this has never given me any trust in that thing for any use case.

What parameters are using to get 12 psi? I tried a bunch of parameters in the Silca calculator, and I couldn't get anything that low for a 208 lb. rider with 60mm (2.4") MTB tires. For most parameters I get something in the range of 14 - 16 psi, which sounds about right.

I'm lighter (about 155 lb), and I usually run 15 - 20 psi in 53mm (2.1") tires. When I enter the corresponding parameters in the Silca calculator it calculates 19 psi rear and 18 psi front, which is in the right ballpark.

[EB]

Originally Posted by Mark McM

What parameters are using to get 12 psi? I tried a bunch of parameters in the Silca calculator, and I couldn't get anything that low for a 208 lb. rider with 60mm (2.4") MTB tires. For most parameters I get something in the range of 14 - 16 psi, which sounds about right.

I'm lighter (about 155 lb), and I usually run 15 - 20 psi in 53mm (2.1") tires. When I enter the corresponding parameters in the Silca calculator it calculates 19 psi rear and 18 psi front, which is in the right ballpark.

They require you to enter total system weight (rider + bike + gear), so for my enduro bike, I used 243 lbs. Surface conditions they don't really have anything for singletrack, so I used cat 4 gravel to approximate it. Measured tire width is 61mm (2.4"). Tires are 700c / puncture resistant, though I suspect they aren't really modeling mountain bike casings. Average speed single track, mountain bike weight distribution.

This yields 13 psi rear, 12.5 psi front. That is literally a dangerous tire pressure for me that will result in tires folding over in corners and burping, at a minimum, not to mention snakebite flats and rim strikes. My typical pressure for dry conditions would be 22 to 24 psi in the rear depending on casing, which is massively more air volume than their recommendation considering the size of these tires.

[MikeD]

Originally Posted by EB

They require you to enter total system weight (rider + bike + gear), so for my enduro bike, I used 243 lbs. Surface conditions they don't really have anything for singletrack, so I used cat 4 gravel to approximate it. Measured tire width is 61mm (2.4"). Tires are 700c / puncture resistant, though I suspect they aren't really modeling mountain bike casings. Average speed single track, mountain bike weight distribution.

This yields 13 psi rear, 12.5 psi front. That is literally a dangerous tire pressure for me that will result in tires folding over in corners and burping, at a minimum, not to mention snakebite flats and rim strikes. My typical pressure for dry conditions would be 22 to 24 psi in the rear depending on casing, which is massively more air volume than their recommendation considering the size of these tires.

I weigh 195# with 2.2" tires. I'm using 25/28 psi on my bike. Silca says about 10 psi less. I won't ride a mushy tire like that. EB is right.

[marciero]

Originally Posted by Mark McM

Testing has shown that rolling resistance asymptotically decreases as pressure increases until it reaches a "breakpoint" pressure, after which rolling resistance starts increasing quickly.

I am dubious of this "breakpoint" phenomenon. It came up in a Leonard Zinn Velonews article in 2022 where he referenced the Poertner/Tom Anhalt tests. The mathematical/graphical model they presented to explain it was fundamentally flawed. It relied on the sum of two smooth (ie "differentiable") functions- representing crr losses and impedance losses- being non differentiable at this breakpoint. But this cant happen. I actually had an email back and forth with Josh, through Leonard, on this. I am a fan of Josh but his response was essentially you are right but we wanted to simplify the explanation. But its not a simplification. In fairness he may not have fully understood my criticism.

Of course, that the model is flawed doesnt mean the phenomenon isnt real, but it made me question the test results. Additionally, non-smooth physical phenomena - in other words phenomena with jointly varying quantities that vary abruptly in relation to one another- really beg for an explanation/model. Absent that it would take a lot of empirical evidence to convince me. Of course there are tons of counter examples- speaking of grip, the friction force when transitioning from static to dynamic-sliding- situation, or if you had a tire with knobs on the side and smooth in the center, or different composition in the center, then as you leaned into a corner you could have an abrupt change in rolling resistance or grip as a function of lean angle when the knobs or section with different composition made contact. In these examples you have a coefficient that changes values discontinuously or at least very suddenly.

[Mark McM]

Originally Posted by marciero

Of course, that the model is flawed doesnt mean the phenomenon isnt real, but it made me question the test results.

The causes (model) of the breakpoint phenomenon may not be well understood, but it has been observed in multiple independent tests. However, the Silca pressure calculator does not rely on using a model of this phenomenon, instead it is based on experimental data. Josh Poertner has worked with several pro teams to to empirically determine the effect of tire pressure under different scenarios to maximize performance (speed), and from this experience he has lots of test data. What the Silca pressure calculator actually does is interpolates/extrapolates its pressure recommendation from the empirical test data, So it should probably actually be called the "Silca Pressure Interpolator".

As far as whether the Silca pressure calculator works for MTB, Poertner himself has admitted he isn't an expert on MTB riding/racing, and probably has little MTB tire pressure test data. It's entirely possible that when the parameters stray far away from the test data (mostly gathered under road conditions), the Silca pressure calculator's extrapolations break down and become invalid.