[Impetus]

I ran a 28x17 for a while, and saw no ill effects.
There's some evidence of faster wear, but honestly with a steel front ring and a steel surly cog, both will probably live longer than you will. (half kidding)

There's some math that shows small gears are more likely to throw a chain, due to frame flex, but I did not experience that.
disclaimer: i only did it for a few hundred miles before moving to a standard 32xwhatever, mostly because I occasionally toss on a 11-42 10speed cassette and wanted to minimize the parts I was installing/swapping. 26x42 is IMO, unusable for anything, and 32x42 is a decent 'sit and spin' climbing gear. I tried 28 initially, because that's what my RF cinch crank came with.

Also, there were no 26t or 28t oval rings available when I bought my current oval ring; only 30/32/34/36t. I bought a 'Neutrino Components' ring because they were the only ones to offer a blue anodized one, so I went with a 32. They now offer an oval 26 and 28.

https://neutrinocomponents.com/index.php?id_category=12&controller=category&id_lang=1

(edit: 26 and 28 are direct mount only. 64bcd smallest is 30t). Bummer. I'm super happy with my Neutrino oval and would 100% buy again.

I will say this: I did not care for the aesthetics of 2 small rings. It just looked funny.

 

[Impetus]

aesthetics are personal preference. It was purely observational.
It's probably more to do with the chain stays on my mtb being much longer than a BMX.

I think trials bikes look 'odd', too.

either way, I successfully had 'small gears' on my Jabberwocky and it worked without issue.

 

[Impetus]

Running bigger chainrings and cogs is more efficient...

.... lower friction (more efficient)....

But how much efficiency difference are we really talking about here? I'd be genuinely interested in seeing the numbers on this. I'm not saying you're wrong, but I just can't see it being "enough to matter".

A paper investigating efficiency of derailleurs vs Rolhoffs shows drivetrain drag as well into the upper 90's% with many factors influencing the system drag, including of the determinations of drag being chain angle- with the largest and smallest being the most 'draggy'. reasonably, the best, straightest chainline has very little drag, mostly the RD jockey wheels, and that still allows 97% efficiency.
A single speed bike would ostensibly be above 98% as measured in the study due to no jockey wheels, but obviously not 100%. I'd be skeptical that the difference in drag between a two extremes of 26x16 (47.5 gear inches) and a 36x22 (47.8 gear inches) is more than a few tenths of a percent.

I'm not advocating for one extreme or another. just 'discussing on a discussion forum'.

 

[Impetus]

I wasn't saying that it's a big deal one way or another, only that if chainring clearance isn't an issue why not go bigger?

I do not intend this to be argumentative, ultimately we both agree the differences are very small, I do agree 'wear rate' is a consideration. Some folks want a little more bling than a humble steel Surly cog, so it's a more important consideration.

Regarding the other point you make:
Frame clearance or ground clearance?

I just measured the difference (using calipers) between the 28t and a 34t RaceFace direct mount rings hanging in my garage. it was just over 1/2 an inch radius difference. (0.5355 inches)

I've left my mark on more than a few large rocks, and broken one chain, due to misjudging. I'd happily take 3/4-1" more clearance using a 22x13 vs 36x21 (approx 32x19) if it was all the same.

All in good fun. not looking to start a war, I'm riding 32x20 (29+) right now and likely always will range from 32x... 18 to 21.

 

[Impetus]

That's a good lookin' bike ya got there, Steven. :thumbsup: