Quoting from Josh through Lennard Zinn (through Velonews):
In general, a wider tire of (the) same construction will have lower rolling resistance for exactly the reasons (you stated). Ironically, the best description and data on this comes from studies done in Britain in the 1800’s looking to optimize the width and diameter of wheels for locomotives. There is also a lot of great info related to this in “Bicycling Science” from MIT press, as well as Paul Van Valkenberg’s writing on racecar tires.
Generally, though, the decrease in rolling resistance becomes smaller as the tires get bigger. So for example, going from a 19mm to a 20mm may save 1 watt, from a 20mm to a 21mm may save 0.8 watt and from a 23mm to a 25mm may save 0.3 watt. There is great data on this in “Bicycling Science,” using old Avocet Fasgrip tires, which were available from 18-32mm. The 28mm and 32mm were nearly identical, but moving from 18mm to 25mm saved a few watts.
What they are missing is the aerodynamic piece. We have data from the Zipp 303 launch showing the 303 with different width tires (see graph). The figure tells the story of how you can really optimize for tires below a certain (width) number, but eventually the tire really dominates the airflow and ruins everything. In general, our wheels are optimized around 23mm tires, which means that 21mm tires usually run about equal, maybe a fraction of a watt faster, but don’t change the behavior of the wheel. Moving to a 25mm adds drag, but can also change the stall behavior of the wheel. And by the time you are at 27mm, you have something that behaves quite differently.
The question really needs to be in regards to the balance of lower Crr (coefficient of rolling resistance) from the wider tire against the aero penalty. The 303 was designed to be as good as possible with 23mm tires, and as a result, its rim is 28.5mm wide. To behave similarly with the 25mm, it would likely have to be at least 2mm wider. In the graph you see how the 25mm tire has the same curve shape as the 23mm tire on the X45 (code for 303FC clincher). The 27mm tire is on the 285FC (code for 303FC tubular), and you notice that not only is the drag higher, but the curve shape is completely different. In fact, the curve shape looks more like the Easton or Mavic. This is indicative of the rim not being able to clean up the dirty air behind the tire. Ultimately, the offset should be Crr watts vs. Aero watts. In this case you have grams of drag on the left; every nine grams is one watt, so from 23mm to 25mm, you have nearly no penalty up to 10 degrees, and then three-to-six watts at the higher yaw. With the 27mm, you have something like no penalty to five degrees, and then a five-to-eight watt penalty after that.
Ultimately for the Specialized I would say that the 0.2 watt (0.3 to 0.8 watt) of rolling resistance does not overcome the zero-to-six-watt aero penalty.
Last interesting note: we have been working with Jordan Rapp on this since he noticed that his ‘training Firecrest’ wheels with 25’s were ‘twitchy’ compared to his race wheels with 23’s… we thought this might be largely aerodynamic, but the shorter contact patch (you) discuss is actually the culprit; the longer contact patch serves to resist steering input and adds a slight damping effect to steering inputs. By lowering tire pressure to increase contact patch, the effect could be eliminated, even though the aero properties of the wheel remain the same.
In the last paragraph, I also did feel that my front wheel (Enve Smart 6 clincher) was a tad twitchier than my 2009 C2 Hed Jet 6 on 23c.
So, right now, I think I will stick with 23c front, 25c rear for races. And probably 25c front/rear for training.
