= 27d567bc2d48d9f40e9ccc20ea370b15)
SoullessAutomaton comments on Why Support the Underdog? - Less Wrong
You are viewing a comment permalink. View the original post to see all comments and the full post content.
= 783df68a0f980790206b9ea87794c5b6)
Loading…= b715d02e85f7b3b4ae65ca3d3e450868)
Subscribe to RSS Feed
Powered by Reddit
= f037147d6e6c911a85753b9abdedda8d)
You are viewing a comment permalink. View the original post to see all comments and the full post content.
Comments (86)
Your mention of signaling gives me an idea.
What if the mechanism isn't designed to actually support the underdog, but to signal a tendency to support the underdog?
In a world where everyone supports the likely winner, Zug doesn't need to promise anyone anything to keep them on his side. But if one person suddenly develops a tendency to support the underdog, then Zug has to keep him loyal by promising him extra rewards.
The best possible case is one where you end up on Zug's side, but only after vacillating for so long that Zug is terrified you're going to side with Urk and promises everything in his power to win you over. And the only way to terrify Zug that way is to actually side with Urk sometimes.
In other words, if Zug believes you to be the kind of agent who will make the naively rational decision to side with him, he will not reward you. You then side with Zug, because it makes more sense.
However, if Zug believes you to be the kind of agent who will irrationally oppose him unless bribed, he will reward you. You then side with Zug, because it makes more sense.
This seems to be another problem of precommitment.
While my own decision theory has no need of precommitment, it's interesting to consider that genes have no trouble with precommitments; they just make us want to do it that way. The urge to revenge, for example, can be considered as the genes making a sort of believable and true precommitment; you don't reconsider afterward, once you get the benefits, because - thanks to the genes - it's what you want. The genes don't have quite the same calculus as an inconsistent classical decision theorist who knows beforehand that they want to precommit early but will want to reconsider later.