In a nutshell, closer ratios where you need em - up high. What with aerodynamic drag being the cyclist's enemy, and that bastard going up by your speed cubed (IIRC), it's a bit of a bummer that we're faced with an inescapable limitation of derailleur gears - that the gaps between ratios get bigger as you go up...

I've long thought the step between 15 & 14 was a bit much... let alone the jumps to make it to 11. So upon recently discovering Shimano had dabbled with 10mm pitch in the 80s (something I've daydreamed about before), I gave it another look, and tried to dig up what I could about it...



...Which is precious little, other than that they did have a track group out for a bit, using smaller diameters for lower weight, but it never caught on. I reckon the main reason it never gained traction was that Shimano failed to take proper advantage of 10-pitch; to my mind its actual benefit isn't allowing smaller chainrings and cogs - there's a good reason to maximise their size.

Which is to provide closer gearing on the road and more choices for the track. I'm thinking if they managed to make a 10mm track chain work in 1982, maybe they can make a 10mm road chain work in 2010... I wonder how the shorter links would affect shifting?

Anyway, I'd say the potential gains are worth more R&D than 10 pitch has received... see what you think of these numbers I crunched for the deal on ratios.

So, hopefully most of us are on the same page at this point re an ideal gear spacing (or debate it, by all means), and you'd say the green cassette below looks pretty good. Problem is, you can't have it cause we're stuck with discrete steps between cog diameters. But if we reduce the size of those steps, just look how much more corncob action we can get happening up high. BTW, that's a 12-cog cassette there just to cover the range from 28 to 11, so you can prolly find your preferred range in there somewhere.



Check out what happens when you subtract one ratio from another to see the steps between them. These cogs are pretty close in size; a 10mm 50t = 39.37t... it's a factor of 1.27.



If you think the ideal curve makes sense, that's what determined the non-integer ideal sprocket sizes in figure 1. The following then, is the difference from that, giving a flat graph of which side of ideal each step is. Maybe it's just me, but I reckon it's infinitely better to err below that line... especially as the ratios get taller.



I can't help thinking guys who cruise in a peloton doing 45-50km/h would be pretty interested in this... I would be too, for the much closer ratios in the middle of the cassette, where I spend most of my time. If the smaller diameters possible were made available, that could also cater to weight weenies and the small wheel crowd. I guess there'd be mixed reactions among the folks who paid big money on eBay for the old Dura-Ace kit...