[EM] Approval strategy

[Forest Simmons]

As Mike pointed out, it's good to keep in mind that optimal Cumulative
Repeated Approval Balloting strategy may not be the same as optimal 1-step
Approval strategy, especially when you are committed to one strategy
throughout the entire CRAB race.

If not so committed, then strategy considerations for the last step in the
CRAB race would be virtually identical to one step Approval strategy
considerations with near perfect information.

With regards to Mike's question about Weber-Hoffman in DSV, I don't know
the answer, but it seems to me that the non-cumulative version of Repeated
Approval Declared Strategy Voting (DSV) would require more use of
utilities than the cumulative version for optimum results. Here's why:

In CRAB the true strength of each candidate eventually comes out, so the
race becomes a simple grab for what we can get away with, i.e. how far up
our preference order can we go?

In the non-cumulative version the lack of cumulative commitment obscures
the true strengths of the candidates somewhat, so expected utility might
figure more heavily in the cost/benefit analysis.

This could also depend on the stopping rule. Do we stop the repeated
ballots after a predetermined number of repetitions, after one repetition
of no approval change, after ten repetitions in which no approval record
is broken, etc.?

I wish someone could shed more light on this, but at the same time I
prefer the cumulative version so much that (for me) the question is mostly
academic.

Forest


On Thu, 9 May 2002, MIKE OSSIPOFF wrote in part:

<snip>

> Yes, it's something that hadn't occurred to me, that maybe
> the elaborateness of 1-balloting Approval's expectation-maximizing
> strategy isn't needed in Repeated Approval. Would that also mean
> that Approval DSV doesn't need Weber-Hoffman, etc.?

<snip>


----
For more information about this list (subscribe, unsubscribe, FAQ, etc), 
please see http://www.eskimo.com/~robla/em