Fred -- How about a little consistency. It's like you have multiple personalities or something.
You suggested two alternative options -- both of them involved blockading in Antartica. Now you are up in arms about "totally conced[ing] Antarctica." That's your plan dude!
Your other points are variations on the same theme "conceding both choke points into Africa" and "letting him keep East Africa and take South Africa". This all boils down to attacking Brazil -- that's it. Attacking Brazil doesn't really solve those things, but it's your principal plan in all three of your proposed options.
I understand that attacking Brazil with a stack is very likely to result in taking it. You might even get a net gain on the attack (assuming Imp doesn't slide over and give the spot up).
The down side is you give up +5 income from greenland next turn and Imp starts outearning us.
We are 100% assured of the +5 income if we take Greenland (zero risk). We could have no net gain taking Brazil.
There are other possible bad outcomes with the attack on Brazil beyond foregoing the +5. Imp put's all his income there and we bounce, potentially missing the card piece if the 3x2 attacks fail in greenland. Imp slides to Columbia and does a delayed attack into Argentina -- then next turn we're stuck blockading on the East side of Antartica and Imp is able to bring his stack in Antartica over to South America.
As with last turn - you may guess right and Imp leaves just the right number in Brazil for you to max the net gain and you win your 3x2s and you pick up Columbia with a 2x1. But I prefer not to rely on guessing right unless I have no choice.
There's zero guess work in my proposed #7. Whatever Imp does it yields an improved position next turn. Is it better than if I accurately guess his moves and play a counter -- of course not. But I don't think that's the best way to play WL.
Even if you guess right this turn -- it woudl not shake my view that the right move is to not rely on guessing right, unless you have no choice.