Naive strategy

[Clay Shentrup]

So, there are four candidates that I know of running in my district in Berkeley. I prefer my favorite to my second favorite by a wide margin, and my second favorite is the clear favorite to win, by a large margin.

Well, I recently found myself feeling strongly intuitively compelled to push my #2 down to the bottom to exaggerate the effect! When this idea first washed over me a few days ago, it happened so fast and unconsciously that my "voting theory brain" didn't even register it at first.

This is fascinating to me, even though it's just an anecdote. Even know that I've thought about it and I know it won't help me to do that, I continue to feel a lingering emotional compulsion to bury #2. I cannot imagine this feeling is rare.

[Clay Shentrup]

> Even know that I've thought about it 

Even now. Wow.

[Nevin Brackett-Rozinsky]

Would you say that your intuitive rankings can be approximated as follows?

• You have a utility uᵢ for each candidate, which is “how much you like them”.

• You estimate the probability pᵢ for each candidate to win.

• So your average (expected value) utility for the election is m = ∑(pᵢ * uᵢ).

• You ascribe to each candidate a “utility leverage” xᵢ = (uᵢ - m) * pᵢ, which is the expected amount of extra utility that candidate is worth to you above (or below if negative) the average.

• You feel intuitively compelled to rank the candidates in order of decreasing utility leverage xᵢ.

Is that about right?

It’s what I devised for my simulation, to generalize the idea of exaggerating front-runners. Candidates who are more likely to win end up farther from the mean (on the same side they already are), while those less likely to win get pulled in toward the middle.

So if you like your favorite enough that your second choice is “below average”, then the fact they are the most likely to win makes their leverage strongly negative.

I also have a related version for strategic *rated* ballots, which uses the concept of utility leverage in a somewhat different way. I haven’t thought about allocated ballot much, but there’s probably a way to make the same idea work for those as well.

Nevin

ps. If the subscript letter ‘i’ doesn’t show up in what I wrote, it means some part of the web didn’t handle Unicode properly (or it’s being displayed in an outdated font).

[Toby Pereira]

On Wednesday, 19 October 2016 07:23:42 UTC+1, Clay Shentrup wrote:

> Even know that I've thought about it 

Even now. Wow.

I think the only conclusion we can draw is that even Clay Shentrup has a human side and not all of his thoughts are rooted in game theoretic principles.

[Brian Kelly]

On Wednesday, October 19, 2016 at 12:19:51 AM UTC-6, Clay Shentrup wrote:

Well, I recently found myself feeling strongly intuitively compelled to push my #2 down to the bottom to exaggerate the effect! When this idea first washed over me a few days ago, it happened so fast and unconsciously that my "voting theory brain" didn't even register it at first.

Sounds like someone is experiencing the appeal of the Later No Harm criterion.