Quote from: eagleflies on January 16, 2015, 09:49:29 PM

Quote from: fairglu on January 16, 2015, 10:44:55 AM

Quote from: ReserviorHunt on January 16, 2015, 10:28:50 AM

Also bear in mind Apex was a dead coin to start with, with little or no stake. It's 51% of the stake that is needed, not 51% of the entire supply. A lot of difference.

I know, but check https://chainz.cryptoid.info/apex/#!extraction and the links, it's more complicated than 51% of the stake.

The attack was done with 0.07% of the coins, but among the apex stakers, there was an average of 15 nodes during the period of the attack, and there were larger holders, #10 & #14 alone held 4% of the coins and were staking.

At the moment the Noble node network is less than 10 nodes on average, which means exchanges, mining pools, explorer, and almost no users.

The problem is that on normal PoS the weight of staking depends not just on coins, but on age, so lots of smaller amounts that have not staked for a while (just keep the wallet offline) that are suddenly allowed to stake in a private sub-net will have field day, same as a "whale" suddenly coming online in a small PoS network. The only missing technical step is to tweak the attack to work around the centralized checkpointing (if there is one), which can be accomplished in a variety of ways.

I think such attack could be mitigated with typical PoS if we used quite low Max Stake Age e.g. 1 week? And Min Stake Age of couple of hours? This way it does not matter if someone keeps staking for months since Weight will be capped.

It's entirely possible that general pos can deal with such problems with some tweaking and adjustements during building phase. It's good that we are aware of these problems now.

Looking forward to the wallets, eagleflies. Thank you for taking up this work