One way to get around the infinite-future problem posed in my last post would be to view consequences as like ripples in a pond: very strong around the point of origin, but getting progressively weaker as time goes on. If I shoot a man in Reno, just to watch him die, the immediate effects are huge - a man dies, his family are adversely affected, I might be arrested and jailed - but the secondary effects are a little smaller.
Now, if it's true that a 'butterfly effect' occurs, then the secondary effects aren't practically smaller. In a hundred years' time, the world might be substantially different than if I had not shot a man - but we can work around that as well. As time passes, even if the ripples don't diffuse, the responsibility does. Other people make decisions based off my shooting a man; people make decisions off their decisions; and it goes on and on like that.
A consequentialist might say that I am not responsible for a third-tier decision made by somebody, or at least that they bear the responsibility for the consequences of their actions as much as I do. On this view, after a hundred years, the number of people who made decisions based off my decision has ballooned to millions, who all bear the responsibility equally. My responsibility for the state of the world then has broken up into a million fragments - and when near-infinite time passes, my responsibility dwindles to an infinitesimally small value.