[EM] Results, for Jameson; ER-DAC; ER-IRV

[Kevin Venzke]

Hi Jameson,

--- En date de : Lun 21.3.11, Jameson Quinn <jameson.quinn at gmail.com> a écrit :




2011/3/19 Kevin Venzke <stepjak at yahoo.fr>

Hi,

Jameson, I did most of what I looked into.


Wow, thanks.
No problem, I'm glad to look into areas of interest.

 
I didn't complete the Asset
methods though. I did come up with a good universal way to estimate the
candidates' utilities for each other though: Have the candidates read the
minds of the voters (the base quantities of each bloc), and multiply each
opinion by their opinion of the candidate asking the question. (I'm
assuming utilities range from 0 to 1.) So, a candidate will like the same
candidates you like, if you like that candidate. This way, B will like
C not because of the spectrum, but because most of his "goodwill" comes
from the C voters.



I think my trick is somewhat better: only have them look into the minds of voters which top-rate them specifically. My rationale is that with your method, I believe it would be impossible to have a Condorcet cycle among the candidates, since A feels about B exactly as B feels about A. With my method, a Condorcet cycle would be possible, which is good - a real-life robust Condorcet cycle would probably be some variant of three candidates, three single-issue voting blocs, and appropriate strength for each bloc. In that circumstance, I'd expect candidates to transfer to the other one who is middling where they are strong, not to the one who is strong where they are middling.
Hmm, I think you're right about the cycles. I'm not sure it's less accurate to do it
my way though. We're sort of finding a correlation between candidates' high rankings,
so I'm not sure why we would expect to find cycles. Not sure there is more information
there to be found.
 
If we do it your way, B will do the bidding of the middle bloc, which prefers A to C.
If that's how B "feels" it's odd that C voters like him so much.


Anyway, the Asset methods stumped me somewhat because I couldn't come
up with a deterministic way to solve the method that doesn't seem to
be contrived. For instance, it's possible that two of the three candidates
are able to transfer. Who has initiative? How do they even know if they
would like to have initiative? Maybe they'd rather do nothing. So, I
didn't attempt to write a method that might not be faithful to the idea.



All "transfers" are simultaneous and represent "copies" rather than bowing out. Since the "can I transfer to you" criterion is the same as the "will you beat me without transfers" criterion, at least in the 3-candidate case there are no issues of initiative or transfer strategy. The pre-transfer 2nd place has no motivation whatsoever to transfer to the pre-transfer 1st place, and no ability to transfer to the 3rd place. So, if transfers are happening at all, it's just that 3rd place is acting as a kingmaker (pseudo-IRV style); that's simple.
 
Aha. I'll think about it again.
 

"MCARA"         96      87
BBB-MMM 99      100     86
TTTT--- 75      100     90
TTTTM-- 46      100     86
TTTTMM- 44      100     85
BBT-MMM 41      100     86
TBB-MMM 35      100     88
TTTT--M 31      100     89
TTTTMMM 29      100     84
BTB-MMM 28      100     85
BBB-CCC 28      93      94
***MCCC 26      94      93
TTTT-MM 26      100     87 ...

You were right that this would do pretty well by SCWE, but your guess
for strategy was ---TTTT (which is actually a quite rare outcome under
any method).

"MCARP"         97      87
TTTTMM- 71      100     84
BBB-MMM 70      100     87
***MCCC 62      94      95 (*=M+P+B, i.e. "A=C>B")
TTTTM-M 48      100     89
TTTTMMM 32      100     88
TTB-MMM 32      100     88
BBT-MMM 30      100     87
TTTT-MM 29      100     85
BTB-MMM 28      100     87
TBB-MMM 24      100     87 ...

This was a little higher by SCWE but your MMMTTTT guess never occurred
once in any (recorded) result for any method!



After some thought, I understand why my guesses were wrong. I was thinking that truncation was a strategy for near-clones, emphasis on near, whereas actually it's a strategy for competitive near-clones, emphasis on competitive. So even though A and B are farther apart than B and C, the voters are treating A and B as the "competitive near clones" because C is an also-ran with the methods I care about.
 
Interesting description.
 
I also ran some sets of non-spectrum-based scenarios with different numbers of blocs.
MCARA and MCARP rank consistently very notably with high sincere Condorcet 
efficiency (both were always top 4), high rate of compression (usually top two, but
Approval grabbed #2 with 8 blocs), and relatively very high rate of "pushover."
 
On the latter point, nothing can touch Antiplurality (13-17%). MCARA and MCARP
were between 4-6% of voters. Several methods hit 1-2% at least once, including
ER-IRV(whole). And there is no contest between ER-IRV(whole) and ER-IRV(fractional).
So maybe there is something to Chris' pushover warning.
 


So, ER didn't make IRV any better but it did reduce the amount of
compromise. Chris had an idea, I believe, that ER-IRV(whole) was
susceptible to a straightforward push-over strategy, but I'm not sure
I see that here unless some Ms or Cs are playing that role. Since the
C voters are very frequently using that strategy (in many methods) I'm
doubtful...

Also, you can note here that the IRV methods did better wrt utility
maximizers than the other methods in this post. You can see (by scanning
for high values) that this is expected when C is losing his win odds
due to the voting patterns, and B is winning when sincere Condorcet says
it should be C.

 
In other words, what's happening is that voters are pre-emptively compromising for "insurance" against an unreliable system, and that ends up electing a centrist, utility-maximizing candidate even when it turns out that the compromisers' sincere favorite could have been the CW.
 
That description is accurate for a lot of situations. But I want to note that I had to
take back this observation for this instance due to my (subsequent) post on utility
normalization. I think it was counter-intuitive and not informative to normalize sincere
utilities when the scenario is based on a spectrum. I'm not actually sure what the
utility situation is for each candidate, but it doesn't particularly favor or disfavor the
IRV outcome, relative to MCA for instance.
 
Kevin


      
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20110321/d5b82b9b/attachment-0004.htm>