> So is a 3 oz weight savings with the Caffin stove worth the compromise in priming/propane conservation performance?
I haven't noticed any compromise. The only time I experienced flaring was when the Caffin stove and canister were at -25F and I immediately inverted the canister after lighting, before placing the pot on the stove. (Producing and evaluating flaring was the point of that test, documented earlier in this thread.) By the time I reached out and grabbed the pot and brushed the snow off, there wasn't enough flaring to stop me from putting it on the stove with my bare hand, and it settled down quickly after that. The other stoves (Coleman Exponent Xtreme, MSR Whisperlite Universal) that I tested under the same conditions also flared until their pre-heat loops warmed. All three stoves seemed to take a 'reasonable' amount of time to heat up, considering the test conditions.
My normal practice with the Caffin and MSR stoves is to take several seconds to set down the lighter, put the pot on the stove, and adjust the flame before inverting the canister. This ensures the rotation of the canister doesn't flip the stove. The process takes the same amount of time for either stove. I'm in no hurry here because I'm cooking with fire near my tarp and quilt and wearing gloves--no flaring allowed. It takes the Caffin stove about as long to warm up as the MSR, so I don't find there to be a priming penalty. I have enough experience now with the Caffin stove that I don't add a priming delay before inverting the canister; it always seems to be 'warm enough' by the time I get to that point.
Since I also use best practices to maintain a 'warm' canister before I start (unlike in the above test), I always have a reasonable gas mix pressurizing the canister. Propane conservation just isn't an issue.
Added: a friend of mine has an Optimus Vega. That's a very nice little stove, way lighter and cheaper than the MSR, and it performs well. I've been recommending it.