Alright: how about a small side bet for us little fish? I bet you 1 BTC that--should the most-work blockchain include a single block larger than 1 MB--that no minority chain will make it beyond 2100 blocks (~2 weeks and the length of a difficulty adjustment) without either a difficulty reset or a change in the proof-of-work algorithm. If I win, the 2 BTCs in escrow will be sent to me. If you win, the 2 BTCs on both chains will be sent to you. If nothing happens by December 31, 2017, we both get our 1 BTC back.
The terms are rather unfair to me (e.g. if the fork happens just before the deadline, then it's impossible for the minority chain to mine 2100 blocks) but I'll overlook that. As long as I also win if BU makes a failure of its fork (i.e. after some time miners realize the BTC chain is worth more, with lower difficulty -- so they switch back). In that circumstance by the end of the year BTC will be the fork with the most proof-of-work, even though at one time it was a minority.
Anyway, I'm happy with 1 BTC and will go as high as 100 BTC. May I suggest Dooglus as an escrow, but I'll take anyone of the standard escrows here.
Good point, if the fork happens just before the deadline then how about we wait up to 4 months for the minority chain to mine 2100 new blocks. If the minority chain doesn't mine 2100 blocks by then, we'll call it dead and I win the bet.
Regarding your term that if a fork happens, but BU reorgs so that the most-work chain no longer contains a block > 1 MB, I might be OK with that, but it's a bit unfair to me because what if we just get a random 1.1 MB block like we saw with bitcoin.com? I could see that happening again, and I don't think that should count as a fork. How about we define a "fork attempt" as occurring once the large block chain has at least 72 blocks built on top of it?
I am happy with Dooglus as escrow.