But back to the choice of 95% for future Bitcoin soft forks, it isn’t a bad choice. A good system needs to be tolerant to minorities, but not too tolerant, and studies show that intransigent minority effects often require 3-5%. Thus I personally might lean toward 90% or 92%.

Christopher Allen

@ChristopherA

In recent bitcoin-dev list discussions, @TheBlueMatt recaps some soft-fork approaches & recommends 95% approval. Why not 99%? Why not some other super-majority like 2/3rds? A data point on the % consensus choices is this article on intransigent minorities: https://medium.com/incerto/the-most-intolerant-wins-the-dictatorship-of-the-small-minority-3f1f83ce4e15

1/11/2020, 12:03:11 AM

Favs: 28

Retweets: 5

link

Christopher Allen

@ChristopherA

Remember, the problem with “majority” voting is that 50% is equal to 100% lack of consensus. A super-majority is better for the long term buy-in for decisions by a community, but what form of super-majority? Math (Paxos) shows us that 2/3rds+1 is one sweet spot. Are there others?

1/11/2020, 12:11:44 AM

Favs: 4

Retweets: 1

link

Christopher Allen

@ChristopherA

I have reservations against community decisions by unanimous consent. It can work but requires a special culture. I find consensus -1 (i.e. 2 dissenters is required to block a proposal) is good for small groups (<~25-80 people). See my Spectrum of Consent: http://www.lifewithalacrity.com/2015/09/a-spectrum-of-consent.html

1/11/2020, 12:18:41 AM

Favs: 5

Retweets: 1

link

Christopher Allen

@ChristopherA

For larger groups, the easier it is for intractable minorities and other adversarial opportunities to cause problems. I’m not quite sure 95% is quite the optimal solution — personally I think low 90s % or high 80s % is sufficient. There may be some math possible to identify it.

1/11/2020, 12:22:12 AM

Favs: 2

Retweets: 1

link

Christopher Allen

@ChristopherA

Collective choice has other weird non-intuitive spots in it — there are a number of interesting unsolved problems & paradoxes. For instance see “Arrow’s Impossibility Theorem” where you can’t create a “fair” ranked or cardinal voting system. https://en.wikipedia.org/wiki/Arrow's_impossibility_theorem

1/11/2020, 12:34:01 AM

Favs: 2

Retweets: 1

link

Christopher Allen

@ChristopherA

Resistance to various adversaries is also a challenge for collective choice. Everything from forms of coercion, censorship, misinformation, collusive vote trading, gerrymandering, etc. All of these get worse the larger the number of parties are involved in making the choice.

1/11/2020, 12:53:11 AM

Favs: 2

Retweets: 1

link

Christopher Allen

@ChristopherA

I have a number of older posts on collective choice in my blog, but I feel they only scratch the surface of a large multi-disciplinary topic. http://www.lifewithalacrity.com/tags/collective-choice/ My recent book @MeeplesTogether also touches on some related topics in Chapter 10. But more exploration is needed.

1/11/2020, 12:59:03 AM

Favs: 6

Retweets: 1

link

Christopher Allen

@ChristopherA

But back to the choice of 95% for future Bitcoin soft forks, it isn’t a bad choice. A good system needs to be tolerant to minorities, but not too tolerant, and studies show that intransigent minority effects often require 3-5%. Thus I personally might lean toward 90% or 92%.

1/11/2020, 1:14:11 AM

Favs: 5

Retweets: 1

link