It seems quite logical to me that compressing more and more down into a black box, would result in decreasing clo performance per fill volume. I would expect to see a slight curviture to the graph at close to 100% overfill. I would expect this to trend to accilerate with higher stuffing amounts, 200%, 300%, 400% right on up to about a 1000% or 10x. (Theory says you should be able to stuff MUCH more, but given the expansive nature of down, it becomes impractical to consider for anything except a rigid container, hence of no value to a hiker.) More compression should yeild smaller air pockets, hence, more effective insulation up to a point. At the point heat is removed as quickly as the increase in insulating value, the curve will flatten and perhaps turn down. But this will be different than the maximum insulation per weight curve. And, exactly how much overfill should be added, in terms of percentages, to have practical value to a hiker, ie, a formula that will let everyone calculate his needed overfill for brand X bag. What constitutes a saftey factor: 10%, 20%, 30%, 40%?
2-4oz of down can mean more than the weight. It means a larger volume bag, also. Neither person really considered the volume of the stuffed bag. We know that larger volumes tend to multiply outwards, perhaps increasing the size of the pack needed to carry it, potentially increasing a packs weight. Or, if we cram it down into a small compression bag, what would be the long term effects on durability on two month hike? More damaged down due to increased compression and loss of loft more quickly? And the compression bag adds weight. Then why overstuff at all? Is it possible to rely on the bag being an "overstuffed 32F bag" or a "normal 20F bag?"
Both Richard and Jerry did an excellent job on presenting data on a very complicated subject. Thanks, guys! As any scientist will tell you, often there is no answer, just more questions.