Westin, how do you think there will be two chains? Even if there is a second 5% chain, no one will take those coins as payment, and no one will want to mine on that chain. Even if there are those coins they will be useless if nobody will use them.
Let me first say that we agree that if the Core-chain drops to only 5% of the mining power, it's likely not to survive.
I do, however, think there are many plausible scenarios in which the Core-chain survives. It only needs to make it through the first two difficulty adjustments after the fork. I suspect this will be easiest if the real fork happens (the first >1MB block is successfully mined) approximately halfway through the adjustment period.
Suppose Gavin's threshold of 75% of miners is met, but only just. With 25% of the miners still mining on the original chain, the Core-chain would drop to approximately one block every 40 minutes and the difficulty adjustment would come within 2 months. That's survivable. Once it's clear it can survive, it's even possible some miners come back. A lot would depend on the Core price vs. XT price, if that can even be measured during those first few days and weeks.
Here's an even more likely scenario. It's easy to imagine the XT miner votes plateauing at, say, 60% for a few months. I've read enough of these threads to feel confident that the reaction to that will not be, "oh well, we can't get the miners on board, so let's stick with the 1MB cap." The reaction will be to lower the threshold. This became very clear to me after seeing that clip of Mike Hearn saying that he would support the fork even with a
minority of the hashing power. It's very clear to me that if the Core-chain can maintain at least 40% of the mining power than it will survive.
Don't get me wrong. I think a fork will be a disaster for the price of both coins. Not to mention colored coins and counterparty. A fork of bitcoin will fork those too.
I'm not a miner, so I don't really need to pick a side here. There are good arguments both for and against raising the block size limit. I get the feeling those in favor of raising the block size limit have convinced themselves they're 95% and there are 5% of holdouts, roughly. If they're right, then probably this hard fork won't be nearly as dramatic as it could've been. But I think they're wrong, and I think there's no way to convince them they're wrong except through the fork itself.
Sorry for the wall of text. This issue's been keeping my mind busy.