sysrenan
commented
6 years ago This is interesting and I can definitely see a need for it, wish they had implemented this earlier, might be too late now. Also, would probably have made ICO period more profitable.
Please consider adopting an improved DAA, such as that implemented in the Masari project, in next_difficulty_v3.
What this does is to mitigate a particular problem that has manifested with smaller coins: ostensibly, a person will hire an overwhelming amount of hashing power, even in excess of the entire network hashrate (during this time, it may be possible to find one pool which reports hashrate in excess of that of the entire network, though that is not necessarily true, as the "attacker" could simply set up a private pool), and point it at a given coin, overwhelming the coin's DAA, and cherry-picking hundreds of consecutive blocks.
During this time, blocks will be found at, say, double the target rate, with difficulty gradually adjusting upward. Once the difficulty has adjusted sufficiently, the "attacker" will then (again ostensibly) point his hashing power at some other small coin, leaving the "loyal" miners of the first coin to suffer though a corresponding period of difficulty re-adjustment back to the coin's "normal" rate.
None of this would be a problem, were there only one coin in existence. And likewise, it would present little problem for a coin that has a network hashrate in the hundreds of MH/s, since the up-front investment necessary to overwhelm the network would render the strategy cost-prohibitive. But it is a viable strategy at the current time, given the number of smaller coins in existence, and it is up to each coin to try and protect against it, individually, though adoption of better DAAs than they inherited.
For the discussion that precipitated this change in Masari, please see here. As a miner of Masari, I can testify that its improved DAA has nearly eliminated the issue, even as the coin has remained comparatively small, with a network hashrate typically hovering around only 1MH/s.