I'm not sure I like the PPLNS method so much, as far as I can see from mineco.in. They estimate your reward by counting your contributions to the last 750,000 shares. With making the window so long, I'm not sure if this pool has much incentive for people to keep mining continuously. Although, currently its total hashing power is about 80 GH/s, which exceeds EMC's. I wonder if they can keep this consistently. And I did hear they had some long blocks recently.
There's no such thing as "incentive for people to keep mining continuously". Either the reward system is hopping-proof or it's not. If it is then anyone can mine whenever they want and get on average the exact fair reward for their contribution. If implemented correctly, PPLNS is hopping-proof. All currently existing PPLNS pools that I know of use a naive implementation that is only approximately hopping-proof, but still good enough.
My point was that variance also comes from the changes in the pool's total hash rate. Especially in the long blocks, people hop-off to leave and mine on other pools. If they can guarantee payment by contributing in the last N shares without any decay, they can do this without loss. If significant amount of people leave a pool then that hurts the chances that the pool will find a block. So it increases the variance for the consistent miner on that pool. The geometric method, in addition to being hopping-proof, also encourages the miners to stay put in the pool. This may not have been the original intention of the method, but I think it still a valid outcome.