Tire pressure did not affect power output over a non-flat 10.5 mile time trial course ridden in near identical fashion for three test pressures: 80/85 (front/rear) psi, 90/95 psi, and 100/105 psi .

Towards the end of the 2000s, bike-people started to embrace wider rims and tires running them at lower pressure. From 19 and 21 mm, 23 mm tires then 25 mm tires became the norm. Narrower tires for more speed had made intuitive sense but truths of reality don't care how you feel.

By that time, I had come across a range of religious tenets for tire pressure and tire/tube choices on road bikes. Many conflicting assertions: 120 psi all the time, go for as high a pressure as possible (over 140 psi), tubulars will always be faster, no, some clinchers are faster, use thinner tires obviously, no, wider tires are better. (And eventually crazytown-banana-pants fatbikes would appear with 4.6"/116 mm and 5–10 psi.)

The main benefits seemed to be that:

1. Wider tires would have wider contact patches and therefore be less deformed in the direction of rotation,
2. Wider tires at lower pressure would conform to road irregularities more smoothly, and be more gripful which would mean better handling.
3. Lower pressure would also deliver an extra degree of suspension and comfort.
4. Less chance of tubes going pop.

Still, wider tires less rolling resistance seemed unlikely. And the tires would be heavier. Cyclists like to fret over things and rotational weight is a solid fret-inducer.

I first used a Hed wheels with 23 mm rims in 2009. I had a Hed Jet 9 C2 on the front, and a Hed C2 Belgium rim with a Powertap and disc cover on the back (thanks to Wheelbuilder). A clear prescription to use lower pressure came with these wider, larger volume tires.

Now this one page of internal book of bike beliefs had been revealed to be without foundation. This was ... unsettling.

I realized that I had to reconsider the 120 psi I had been using for many years by that point, and doing so with no great conviction. I was simply copying what many other people did (who were also themselves copying; it's what we do).

Here's how I demonstrated to myself back in May of 2012 that lower pressure does not slow the bike down, and made handling and comfort better.

• I set up for a loop course around (but not up) Mount Philo in Charlotte, Vermont. I was riding a 2010 Cervelo P3 (my second and still current triathlon bike) with Specialized clinchers with butyl race tubes (this was a pretty good set up but am now using Continental TT and latex tubes). Here's a later version of me taken by Dave Connery at the end of the Greenbush Time Trial (the Festival of Speed). Mount Philo is in the background. Pain is in the foreground.
  
  
  
  
• I rode the first loop with just above 300W average power. The goal was to put out a fairly even effort though it's not a flat course. I had a Garmin 500 for feedback.
• I then saved the route as a course and set up the Garmin so I could ride the same course again shadowing my initial time trial.
• On the second and third loops, I had the Garmin show me only where I was relative to that initial loop. in terms of distance (feet/yards). I didn't look at power or have any other information feeding back. I tried to stand when climbing in the same places, ride aero everywhere else. I simply focused entirely on traversing the loop as close as I could to the first time trial. In this way, I would then be able to see how the power required changed for the same performance. And it would not matter how hard it was for me.
• The plan worked well, and I was able to emulate the first ride two more times, never getting too far or behind at any point. The three overall times were just over 27 minutes and were all within a few seconds of each other. It was a nutritious training session.
• Here's a replay of the first and second time trial that shows I tracked myself well:
  
  
  
  
• Somewhat randomly, I ran 80/85 (front/back), 100/105, and 90/95 psi. It would have been good to try 70/75 psi but I had put in enough work.
• It was windy but there was no obvious change in wind speed or direction over the time I was out there. (I did not wear an anemometer on my helmet.)
• I took about 10 minutes rest between the loops.
• The Strava entries for the three rides:
  1. https://www.strava.com/activities/9875804/
  2. https://www.strava.com/activities/9875822/
  3. https://www.strava.com/activities/9875842/

Very surprisingly to me at the time, there was no real variation in power.

Average power (W): 306, 306, and 309. The 309 makes some sense in that I was beginning to work inefficiently and I was using the wrong gears on some climbs. Well, more wrong than usual.

Normalized power (W) for each piece were indistinguishable. The Garmin 500 didn't record normalized power at the time so here are two measures calculated afterwards:

• Coggan's normalized power (W): 326, 329 327.
• Strava's effective power (W): 313, 314, 314.

See the table below for more statistics.

Note 1 from training record: "I did not follow power at all on the second and third time trials as I used Garmin's course thing I stayed within a few seconds of my first ride, seeing up to a 16 second gap on the second effort. Dropped behind a little towards the end of the third and managed to just come in a few seconds ahead. The course thing worked impressively well."

Note 2 from training record: "The power numbers are solid power and I guess I trust them now. Still, I should be going faster with that kind of input, right?" Right. Various improvements helped in the next few years: Better seat, better tires, latex tubes, and, a big, big deal: leg shaving.

Note 3 from training record: "The chain slipped off the small cog on the third effort during the initial downhill and I managed to get it back on without stopping. I have to get that fixed."