I've been practicing riding hands free (when I've got a safe traffic-free opportunity of course). This is partly because I'm trying to learn to ride a unicycle and hands free balance is a skill I can work on while commuting. It's also good o be able to have a stretch on long rides.
Having tried several different bikes, it's clear that not all are equally easy to ride without hands. In approximate order, starting with the easiest:
- An old BSO, 26"x2ish wheels, at least 1 size too small for me (I'm fixing it up for an errand bike to do about 3km/day and store outside work). 71° head angle
- The rental bikes near work - heavy step-through frames with low gearing (Nexus 3-speed hub gears), 26x1.75ish", either Marathon Plus or a few with solid tyres.
- My hybrid (GT Traffic 3.0 from about 2010), 700C, currently with 35mm studded tyres but normally 28-32mm Marathon Plus. This is fitted with a single pannier. Sometimes my D lock is clipped to one fork leg, which makes it slightly harder. 72° head angle
- My 29er hardtail (Saracen Tufftrax Comp Disc, 2.3" tyre at the front, 2.1" at the back). This is about equal with the GT. 67° head angle
- Hardest by far my tourer, Genesis Tour de Fer, 700x32mm tyres. 71° head angle
I've added the head angles, I could find specs for the Genesis and the Saracen, and used those to check my measurement technique with an angle gauge and spirit level. I confirmed by eye that compared to the Genesis the GT is steeper and the BSO matches.
It's clearly not about quality, unless in an inverse sense. There does seem to be a little getting used to a specific bike after switching from another, but that can be done in a minute or two. On the first few I can do a few hundred metres including gentle bends and speed bumps, often limited by running out of clear enough space, a change in gradient, or too many bumps causing me to slide forwards on the saddle.
So what factors affect the ease? I'm interested in fundamental things like geometry, but also setup/maintenance factors like tyre choice and loading.