[EM] Sims - method terminology

[Kevin Venzke]

Hi Jameson,

--- En date de : Sam 12.6.10, Jameson Quinn <jameson.quinn at gmail.com> a écrit :
>>I'd thought of this method, and consider it a very good one if you're 
>>willing to go with a non-?>summable, computationally-complex method. I am 
>>surprised to hear you say that "sincerely voting >this isn't very 
>>realistic". True, it is not LNH, so truncation is possible; but I don't 
>>see truncation as >a dominant strategy here....
> 
>I believe you will have a perfectly typical truncation dilemma when your
>side has two near-clones and you expect a candidate on the other side to
>be the initial leader.
>
>No. If voters are using purely rational strategy, and they know which 
>near-clone is likely to be the first-round frontrunner, the voters for 
>that candidate have no motivation to truncate. The second clone's voters 
>will lower their threshold first. If that's not enough for the first clone 
>to win, because of truncation, then the first clone is out of the running 
>(if they can't expect other-side crossover votes) or a lock-in (if they 
>can). The only thing that first-clone truncation can accomplish is to 
>elect the other-side candidate.

Sorry, I'll use letters to explain. Call the clones B and C, and it's
expected that A will win the first round. Actually maybe it's clearer to
think of B and C as being perceived as sort of like clones but not
necessarily. Then after A wins supporters of B and C have a single
opportunity to say what they will do in response to A leading, which
everyone expected A would do. Voters may feel divided on whether to
give in and support the other side (which is basically equal to offering
no distinct support to their preferred of B/C to begin with) or hold
out and hope for new support for your first choice.

If this isn't a problem in CdlA then it shouldn't be a problem in many
other methods that I also force to be truncated.

>>By the way, it has occurred to me that I can attempt to determine what
>>is a "realistic" set of candidate placements by excluding scenarios that
>>are not competitive. Unfortunately I'm not producing this data yet.
>
>There's plenty of real-world elections which aren't very competitive. Or >am I not understanding what you're saying? Do you mean "useful for 
>discriminating between methods" instead of "realistic"?

No, you have my meaning correct. If one candidate wins all the time in
a given set of positions of candidates then those nominating the other
candidates didn't strategize very well. In reality there is incentive to
try to nominate competitively. It happens occasionally that only one
competitive candidate is nominated, but in simulations it happens quite a
lot.

Kevin