> Once the Bitcoin miner has a competing branch with greater total difficulty than the rest of the network, the network must accept that branch as the winner.

> Once the Bitcoin miner has a competing branch with greater total difficulty than the rest of the network, the network must accept that branch as the winner. Could this attack be addressed by changing the "total difficulty" function to value chains published earlier more than competing chains published later? Say a block published today is considered twice as difficult to achieve as a block starting from the same point published in 48 hours. Blocks don't need to be available in real time, but there should be enough communication between the miners to get some consensus on roughly when each block was published. Intuitively, it seems like there is absolute agreement on the fact that a 51% attack happened (and which chain was mined by the attacker), nobody is saying "we can't be sure, this can sometimes just happen by mistake due to the distributed nature of mining". So if that can be trivially detected, why not build the detection into the mining algorithm in the first place?