Belts are still fairly new and uncommon so frame compatibility and cog/cassette availability will improve in the future. They're not likely to ever over take chains, however. My belt drive commuter bike is quieter and smoother than any chain driven bike I've ever owned. Belts are cleaner than chains which can be an issue for commuters who bring their bikes onto trains or into offices. Probably the biggest issue for belts will be cost, not just of the belt itself but of the required multispeed hubs to get sufficient gear range. Those hubs, it should be noted, are more than heavy enough to wipe out the weight advantage a belt has over a chain.

I have a Redline Mono Belt, hard tail mountain bike, steel frame... I love it! The bike and the Gates Belt Drive, quiet & smooth... not to mention pretty much care free. I could see myself with a around town "utility" bike with a IGH or even kind of an All a Rounder. If speed & gear ratios are not all that important to the rider...

Not really surprised to see chains are more efficient than belts, but 34%? Been loving the Gates belt drive on my Globe Live 3, but my Globe Live 2 (which is identical but with a chain) is noticeably faster and easier going up hills. And yep, belts can turn into cartoonishly squeaky noisemakers when wet. I have to wrap mine in toilet paper to get the squeak to stop. Not specifically mentioned above, but if you change your gearing, you're probably going to have to change your belt too, which ain't cheap.

I expect "34% greater efficiency" translates to something like 99% versus 99.34%. Math doesn't lie, but it can be presented in disingenuous manners.

This is the source of the 34% efficiency difference, a VERY well done, well thought out, scientific test with very real results - and, yes, that's a straight up, literal 34% difference, not 34% of 1% or some kind of misleading statement... https://www.friction-facts.com/media/wysiwyg/Gates_Carbon_Belt_Drive_rev.pdf [why they didn't cite this, I have no idea] The problems comes from the extraordinarily high preload(something like a 70 or 80lb difference vs the typical chain) required by Gates for their Carbon belt drive. Timing belts and chains are both about 98% efficient systems('all things equal', well maintained, lubed, *proper* setups - an equivalent straight cut gear drive is 99% efficient, but that drops ~1% with each gear added and is obviously not realistic on a bike), but, in this case the problem is the implementation. Should enough interest ever build up, expect to see truly high performance belt drives that see very little if any efficiency loss to a typical chain drive.

Ry_Trapp0 this is 3 Watt vs 4 Watt when a rider put 250Watt, so this 34% is really less than 0,5% of legs power.