As long as we have people hitting us and leaving us it is going to keep bouncing around with the difficulty changes.
I was waiting for this answer, from where does this additional hahsrate come???You all realize this can happen with ZERO change in network hashrate right? You all realize that finding a block is brute forcing some numbers together and that sometimes you get lucky early, and find it, and sometimes you don't, right?
The jumping difficulty makes sense from a botnet/pool-hopping prevention standpoint, but from a "predicting profitability" standpoint, is really unfriendly. I've watched the difficulty swing by a factor of 6.5x today. I saw 1.55k as a low and 10.2k as a high (not to say I didn't miss something more extreme than that). It may still be too early to tell but I think you may have missed the mark on re-targeting slightly and maybe it should also utilize a daily (i.e. past 1440 blocks) cap. For example, difficulty can raise up to 8% per block, but no more than 15% over the difficulty from 1440 blocks ago. That might have wound up being the limiting factor, and would likely need some hard math to determine what best to use, but something that gets you there in a more gradual fashion. This is just something I'd had rolling around in my head, and a day may be too short of a choice too, but some sort of delta-diff-limiter. As a miner, I hate wild swings in difficulty - most of which have only been necessary on coins that have large multi-pools that can hit them, of which, UTC is not currently one, or likely to be one in the future due to the lack of coins using the same algorithm on the same N factor.