This market resolves according to Dynomight's proposed experiment:

IMPORTANT INFO: I have tweaked the setup a bit. In addition to the changes in red, each of the ten bits had a 59% chance of being a 1 and 41% chance of being a 0. These bits have already been generated, so the 'weight' of coin B is locked in, but the actual 'flipping' of the coin will be done publicly at time of resolution.

-The coin with the highest predicted chance of landing heads resolves according to what is detailed above

-The other coin, the coin that has the lower predicted chance of landing heads, resolves NA

MD5 hash of the ten bits which determine Coin B's chance of landing on heads: 4aedaed1791d8b56f0f714abad5c81ba

All random chance will be publicly implemented via 

I have bet the chance of coin A resolving heads to 59%, but from this point on I will not bet on either coin.