Has the following reply to Pascal's Mugging been discussed on LessWrong?

1. Almost any ordinary good thing you could do has some positive expected downstream effects.
2. These positive expected downstream effects include lots of things like, "Humanity has slightly higher probability of doing awesome thing X in the far future." Possible values of X include: create 3^^^^3 great lives or create infinite value through some presently unknown method, and stuff like, in a scenario where the future would have been really awesome, it's one part in 10^30 better.
3. Given all the possible values of X whose probability is raised by doing ordinary good things, the expected value of doing any ordinary good thing is higher than the expected value of paying the mugger.
4. Therefore, almost any ordinary good thing you could do is better than paying the mugger. [I take it this is the conclusion we want.]

The most obvious complaint I can think of for this response is that it doesn't solve selfish versions of Pascal's Mugging very well, and may need to be combined with other tools in that case. But I don't remember people talking about this and I don't currently see what's wrong with this as a response to the altruistic version of Pascal's Mugging. (I don't mean to suggest I would be very surprised if someone quickly and convincingly shoots this down.)