Quote from: Daedelus on February 24, 2015, 11:27:30 PM

You weren't frothing at the mouth in your post so there is no reason why anyone else should be Grin

and you seem genuinely interested so the following applies to Nxt POS

Longest chain rule with highest difficulty applies in Nxt too. Nxt is pretty much the same as bitcoin if you imagine each Nxt as a mini mining rig. The current solution to joining the network after a break or initially is the yet to be implemented Economic Clustering feature. The idea being, to complete a transaction you both have to be on the same chain. So everyone joining would look to an entity (or two or ten) that makes a lot of activity I.e. a store, exchange etc. This is where the economy clusters around. Everyone has the incentive to be on the same chain, as otherwise their transactions are void so number of forks should tend towards 0. I believe Vitalik refers to this concept as Weak Subjectivity whrn he wrote about it in his blog.

In theory, if 51% the network nodes and passing around garbage chains then as long as people stick to the economic cluster (which I believe will be automated but don't quote me) then the attack will be resisted (a lot of garbage floating around though). I believe this is what led BCNext to believe Nxt was 90%+ resistant to attack. And the garbage chain would still have to have a higher block height and difficulty to stand a slim chance of a shred of success, which I think Kushti and Anduiman have shown is very very unlikely.

Kushti will correct me if I am wrong Grin

ok. The highest difficulty being the 'Chain with the most Stake involved' is nice for sorting out chain forks on the valid chain. But I don't think that helps with the 'bootstrap' scenario, as an attacker can fake as much stake as he wants on his 'fake' chain. 'Fake as much Stake'.. lol..

Also - the Economic Clustering, which as I understand it involves putting the hash of a previous block that must be on any chain this txn is added to, does indeed help reach consensus on the valid chain. But again an attacker could fake as much or as little activity as he wanted on his fake chain. Thus in a bootstrap scenario, I'm not sure that helps either.

?