OK. I think we can work this out. Before I make more arguments, I'm going to try to explain the disagreements as I see them, and ask you more about what you're saying.<div><br></div><div>A. MAV vs. ER-Bucklin (ERB, though we should probably find a better name at some point). That is, completion using above-median, or completion using median-or-above.</div>
<div><br></div><div>I don't see a huge difference here. I think MAV is slightly better because of the chicken dilemma, but it's possible that that regular voters would see ERB as simpler by a big enough margin to make it worth supporting instead. I'm probably not a good judge of that because my mathy brain tends to see them as exactly symmetrical.</div>
<div><br></div><div>I think you're beginning to understand my point about symmetry. Your view is that ERB is better because of apparent simplicity (empirical question about voters; we could find agreement here with more evidence) and also because of some Deep Principle of counting all the votes which I don't understand (because in my view MAV and ERB are <b><i>exactly</i></b> as likely to "count a given vote" or not, that is, to make a given gap in ratings between two candidates significant or not). I'm not sure that explaining your Principle further would help me understand it, because it's probably not going to fit with my logic. I would, however, like to understand more about whether you see the simplicity or the Principle as more important here.</div>
<div><br></div><div>B. MAV vs. EMAV</div><div><br></div><div>On this question, our principal disagreement is around the strategic impact of the voting system, both for general exaggeration and for specific chicken dilemma scenarios. Here's my logic:</div>
<div><br></div><div>...Start Jameson's logic...</div><div><br></div><div>1. In different systems, different strategies are effective. I'll give one "5 candidate linear" (5CL) scenario and one chicken dilemma (CD) scenario to illustrate my point.</div>
<div><br></div><div>Honest utilities</div><div>5CL:</div><div>23: L100, CL75, CR25, R00, RR00</div><div>25: L50, CL100, CR50, R00, RR00</div><div>24: L00, CL50, CR100, R50, RR25</div><div>22: L00, CL25, CR75, R100, RR50</div>
<div>06: L00, CL00, CR25, R50, RR100</div><div><br></div><div>(To make this comparison clearer, I'm going to use 4+1 ratings for both systems, though it would actually work the same if under EMAV, "25" represented a coinflip between the closest available ratings, 50 and 0)</div>
<div><br></div><div>CL is the honest winner under both MAV and EMAV, with a median of 50. Under both systems, the CR voters could win by dropping CL's rating to 25, and the CL voters could defend against this by dropping CR to 25. Under both, the RR voters could elect CR by rating them at 75 or above. But EMAV presents a number of further possible strategies: for instance, the R and L voters could both help their preferred frontrunner by rating CR and CL at 100 and 0 or vice versa, whereas with MAV these strategies would have no impact.</div>
<div><br></div><div>CD:</div><div>40: L100, RA0, RB0</div><div>31: L0, RA100, RB75</div><div>29: L0, RA75, RB100</div><div><br></div><div>RA is the honest winner under both systems. Under both, RB could win if his voters drop RA's rating to 25 (a risk-free strategy), and RA voters could defend by dropping RB to 25. Under both, RB voters could still win by taking the risky step of dropping RA to 0, but if RA voters "defensively" or "in retaliation" took a similar step, then L would win. </div>
<div><br></div><div>But in EMAV, only 7 RB voters would have to drop RA from 25 to 0 in order to win, while in MAV it would take at least 20 of them. Since the 7 it would take in EMAV do not in themselves create a risk of L winning, it would behoove 13 RA voters to defensively downrate RB so that it would take a risky 20 extreme-strategy RB voters to win. But this slippery slope continues, and pretty easily, through miscoordination or over-risky strategies, L could end up winning. In MAV, such problems are less likely, as any extreme strategy that might win is also inherently risky; instead of a long and slippery slope, MAV gives a flat buffer and then a sharper cliff.</div>
<div><br></div><div>2. I believe that candidates, pundits, and/or voters would realize the different strategic potential of MAV versus EMAV, leading to significantly more-exaggerated votes under EMAV.</div><div><br></div>
<div>3. I believe that the BR downside of that additional strategy would be greater than the BR upside of EMAV's greater sensitivity to honest votes.</div><div><br></div><div>....End Jameson's logic...</div><div>
<br>
</div><div>From what I understand, you (Abd) disagree with point 2 above on both empirical and philosophical grounds. That is, you think empirically that people will not in practice exaggerate more in EMAV vs MAV; and philosophically, that if people did exaggerate, that doing so would be a moral choice which shed light on their truer underlying utilities, so that whatever result EMAV gave in that case would actually not be a worse BR.</div>
<div><br></div><div>I can, to a certain extent, understand both of these points. I think that the philosophical point is, to a certain degree, moving the goalposts; if your model doesn't allow honest and voted utilities to differ, then of course Score-like systems will come out better. But from another perspective, it's just the converse of your empirical point; that is, if your empirical point is "people won't exaggerate", then your philosophical one is just adding "unless they actually mean it".</div>
<div><br></div><div>Am I characterizing your arguments fairly? If so, how do they apply to the two specific scenarios above? </div><div><br></div><div>I'll stop here, and won't argue further until I'm confident we understand each other.</div>
<div><br></div><div>Jameson</div>