"Do you think the reason for this trend is simply the efficiency of the burn, or just the flames whisking around the edges of the pot? In other words, should I be worried about greater exposure to CO while waiting for my water to boil on low heat? We did assume no wind ;)"
The efficiency of the burn has little to do with it at low settings. Generally, the low settings would require a reduced air inlet at the jet. . .like a Bunsen burner. This would provide the optimal mix of cooler air into the gas hence burning. Since all backpacking stoves have a fixed orifice, usually set to high flow rates, we get too much air in the flame anyway. This means more oxygen than is really needed at low settings. This leads to high efficiency burning for all but overloaded settings. It *is* possible to overdrive a canister and reduce heat production, increasing CO production. But, generally, you do not need to worry at low settings. CO2 is a worry in a confined area. Reburning that will also produce CO. You always get some, anyway. Generally, you will get slightly less CO with plenty of air, but a cooler flame. One of those things that I was explaining about in the plethoria of minor variables. There is a curve to CO production, related to the air input and the actual gas/air mix you are burning. But, I do not worry about it. If I were to use it for 3-4 days in a tent, I *might* detect a slight increase in CO in your blood, but certainly not enough to make you sick. It dispurses slowly, however. For a week, I would use more caution since even small amounts can build up in your blood stream.
I used the alcohol blend because it approximated the SLX I normally use for camping. It burns pretty clean, depending on the size of the vapour outlet. However, the additional air will likely cool the flame somewhat. It doesn't really matter because the heat still needs to extracted from the flame/hot air rising from around the flame. Any air warmed by the flame carries heat to be extracted from the system, so it doesn't really matter. I am not looking to adjust flame temp, I am adjusting the entire heat producing system. Most of that stuff evens out with alcohol, WG, Kerosene, and canisters, too, provided you have enough oxygen for burning them.
Turbulence, vortex generators, heat exchangers (HX's,) nsulated sides are all cheats to get to the infinite plane you are talking about in a finite, small area.
David is quite correct. Painting the fins with a IR absorber will help. As will induced turbulence (through vortex generators) and insulated sides and lid. I disagree that medium temp is the best for raw efficiency, however. It may be the best for time vs efficiency...never checked that with canisters except roughly. The most fuel efficient, ignoring time to boil, is low in a good system.