[FRIAM] 2019 - The end of Trumpism

[Eric Smith]

Steve, 

I wonder if there is a game theory problem to be worked on here.

Referring to your statement: 

>> Arrow's Impossibility is real but no more significant IMO than the real-world ambiguities and paradoxes introduced by practical realities such as voter suppression and fraud, system hacking and mechanical errors (e.g. hanging chads)…   

The Impossibility Theorem has the character of a case-existence proof: for any algorithm, there is a case of voter preferences for which that algorithm produces an unwanted outcome.  In the sense of only counting cases, it reminds me of no-free-lunch theorems: for any algorithm that is fast to solve one problem of combinatorial search, there is some other problem for which it is slow.  However, the NFL _threorem_ — that no algorithm is any better than any other — depends on an appropriately symmetric search space and a suitable associated uniform measure over problems on that space.  If search and optimization are embedded in a larger dynamic where correlation between algorithms is allowed, there can be global better or worse approaches.  I don’t (as in every other area) have details and references ready in memory, but David Wolpert wrote some of his later papers on NFL addressing the ways it ceases to apply under changed assumptions.

I wonder if anyone has done an analysis of Arrow Impossibility in a context of a kind of ecosystem of adversaries.  To game any algorithm, crucially with the outcome that not only _some_ voter is handled poorly, but that _a sufficiently large pool_ of voters is handled poorly that the algorithm is not best, requires arranging the preference case that violates the algorithm for suitably many voters.  Is this coordination problem harder for some preference-orders than for others?  Is there something akin to “canalization” in evolutionary biology, where some algorithms live further from the boundary of being collectively tipped into producing the wrong outcome than others?  Thus, are there measures of robustness for statistical violation of algorithms based on what happens in most cases rather than what happens in the worst case, as there are for spin-glass phase transition problems?

Another thing it seems unlikely I will ever put time into being serious about.  Or maybe there is already a large standing literature that claims to have addressed this.

Eric




> On Dec 28, 2018, at 7:04 PM, Steven A Smith <sasmyth at swcp.com> wrote:
> 
> oops... originally sent only to Marcus by mistake...  
> 
> On 12/28/18 6:59 PM, Marcus Daniels wrote:
>> https://www.cnn.com/2018/12/28/politics/maine-governor-certifies-congressional-election/index.html
>> From: Steven A Smith <sasmyth at swcp.com>
>> Sent: Thursday, December 27, 2018 9:50:02 AM
>> To: Marcus Daniels
>> Subject: Re: [FRIAM] 2019 - The end of Trumpism
>>  
>> Marcus writes:
>>> Steve writes:
>>>  
>>> "Democracy is the tyranny of the majority over the minority"
>>>  
>>> The majority elected Hillary Clinton.
>>>  
>>> Marcus
>> The Electoral College is archaic and ambiguous: 
>>     https://www.archives.gov/federal-register/electoral-college/faq.html#changes. 
>> Perhaps our current orange-tinted clusterf*ck will continue to degenerate to the point of motivating the necessary will to mount the necessary constitutional amendment.
>> Republicans are acutely good at gaming vulnerable systems to their benefit (gerrymandering, voter suppression, etc.) but the DNC and Hillary proved to be their equal during the primary with Superdelegates.   
>>     https://www.theguardian.com/commentisfree/2018/jun/11/democrat-primary-elections-need-reform
>> Ranked Choice voting seems ultimately yet more promising to *improve* the selection of our representatives.  I believe that Maine is running that experiment for us now at the State level. Arrow's Impossibility is real but no more significant IMO than the real-world ambiguities and paradoxes introduced by practical realities such as voter suppression and fraud, system hacking and mechanical errors (e.g. hanging chads)...   Technology (can a direct democracy be facilitated by something like block-chain technology?) might resolve some of these questions, but very likely it will miss the more fundamental philosophical questions.
>> We are a Federal Republic with a Representative Democracy for good reasons... some of the context of those "good reasons" surely has evolved over the 250ish years it has been in place while the mechanisms maybe have not evolved as quickly.   Individual and small groups of Opportunistic, Brash, Narcissists can usually outmanouvre such a slow moving leviathan.   I'm not sure what to do about that.
>> How does Direct Democracy distinguish itself from Populism and Mob Rule?   What constitutes (guarantees/assures?) an engaged and informed electorate?
>> But the question remains:  Is there a better way to meet the goals of governance than the democracies we have tried and/or imagined?  How do we balance (or align?) the needs of the group and of the individual?  Is "Democracy the worst form of government except for all of the others we have tried" (Churchill paraphrase)?
>> - Steve
>> 
>> 
>> 
> ============================================================
> FRIAM Applied Complexity Group listserv
> Meets Fridays 9a-11:30 at cafe at St. John's College
> to unsubscribe http://redfish.com/mailman/listinfo/friam_redfish.com
> archives back to 2003: http://friam.471366.n2.nabble.com/
> FRIAM-COMIC http://friam-comic.blogspot.com/ by Dr. Strangelove