J Mole, good points. Personally I'm interested in non theoretical real world 2 cup usage for various canister burners, by model and type, outside, in air that is moving.
I was going to mention simmering too, but that gets very unfair to canister stoves, I did a quick test after finishing the ion stove, which is easy to make, relatively, and created a simple can top with 1" hole in it as a simmer ring, and was stunned by the results, 10ml, 8 grams, simmered two cups of water for about 32 minutes, insanely efficient. I had to keep lifting the pot to convince myself the thing was still burning. Anyone who simmers should take a very very close look at a well done alcohol stove setup, if I actually cooked meals, not rehydrated, the alcohol would be a slamdunk no competition winner, since 25 ml/20 grams would cook a non dehydrated dinner, give or take.
I think a lot of the views of alcohol efficiency come from the early emphasis on raw speed over efficiency, and losing sight of the specific gravity of an ounce of alcohol, ie, comparing re weights apples to oranges, ml to grams that is.
Another issue is that almost nobody does real efficiency testing on the various burner/screen combinations out there for canister stoves, based on what I found with alcohol, it's quite likely that a remote canister with a high quality, custom fit wind screen like you use on a good alcohol stove setup, is going to significantly improve your efficiency, but by how much I can't say. So when someone points to a 1.5 ounce canister burner, no mention is made of the actual efficiency of that set of burner sets/valves, in a sense it would be interesting to do that type of research but that would involve buying a lot of expensive gear I'd never use, and a huge amount of canister cartridges. Easier if people who own their setups just report the information accurately, 2 cup boil, how much fuel required with burner x, screen, no screen, etc.
Since the 1/2 ounce liquid alcohol efficiency has been around some 8 years now at least, and the stoves that do that are easy to build and light and fill, I don't really see any particular reason to not use them as a standard, that will also show how various other alcohol setups perform, using SLX, which is also a standard, for methanol, figure some 10% more consumption, for ethanol, some 10% less.
By my math, using a 100gm canister, you can never reach the weight efficiency of alcohol, it's simply not possible because the canister body weighs almost as much as the contents if the weights listed here are correct, that eliminates the higher btu boost. So that's a non starter, no trip would ever be lighter using that setup, I don't see how it can be since 5 oz of alcohol fuel always weighs less when you take the container/burner into account, so I don't believe there's anything subjective there going on.
One good thing about thinking/researching this is that if I ever get a canister, it will be a remote, because then I can leverage all the same methods and efficiency techniques that alcohol stoves have taught me, once it's debugged I would imagine 8 gram or less boils should be quite doable, at which point the fuel / container gets competitive again, though I don't know the weight of a light remote setup.
I think there's a tendency to ignore these fairly empirical facts when selecting for the ease and convenience of the canister setup, it's fine to prefer one method over the other, but the numbers should be real and based on comparing apples to apples, not comparing a quick and dirty cat can setup with zero efficiency testing and no work on screen at all to a commercially produced machine, a canister burner unit, that is highly tested and optimized.
What I was, however, very pleased to discover, is that the cone is absolutely not required to reach high efficiencies with alcohol, well, rediscover, since this has been known for getting close to 10 years, it just seems to get forgotten because it's too simple and non glamorous so I guess people think there's some catch to it.