I've read this thread and I still think that your system has the risk of default (and the subsequent collapse).
Since you can't print bitcoins (not even for a while), you can default.
My proposal in the sister thread actually mentions that temporarily printing bitcoins might be an emergency measure added to the protocol:
http://forum.bitcoin.org/index.php?topic=31645.0I don't know what is wrong with the derivatives with bitcoin escrows. You still get what you want, oil-coins.
Bitcoin-denominated derivatives are only slightly useful if bitcoin values are bouncing all over the place. I might be right that oil will rise 5% over the next six months, but my gain is dwarfed by a 50% drop in bitcoin values that I used to escrow my bet. I have to have some kind of bitcoin option strategy to hedge against a drop in bitcoin values, which is getting way too complicated for anyone but an options expert.
If I've understood vector76 properly, you would have an oil-coin versus a derivative short oil and long bitcoin.
The problem I didn't get about fungibility is that any of the parties could redeem the contract at any time.
That problem could be reduced if the system can look for contracts enders on both sides of the bet.
Remember that both sides of the derivative are always solvent (or the contract is automatically redeemed).
I also came to the conclusion that you cannot use bitcoins for the escrow, because maybe you need to transfer bitcoins of it from a user to the other and you don't own the private keys.
You need derivativeCoin.
Everything I have proposed so far requires changes to the existing bitcoin protocol. It might also be possible to engineer something that runs distributed, is backed by bitcoins, but does not need to touch the bitcoin block chain. I would be very interested to hear proposals on how that might work.