Thread by @jbrukh

Hi all, today I am open sourcing some thinking I have been doing around valuing #DecentralizedGovernance. I am releasing a formal spec and Python code to describe the framework and to get community feedback.

github.com/coinfund/gover…

Demo below.👇

The central question I am trying to answer is the following. Given a governance token distribution, how do we quantify the influence that a particular token holder’s stake exerts over the governance process?

To answer this, I define a measure called “decisiveness”.

Intuitively, the decisiveness of a stake relative to a token distribution is the expected % of the time that this stake’s vote (or lack thereof) can overturn the outcome of a proposal.

The properties of decisiveness are based on discrete partitioning & are somewhat surprising.

For example, imagine a system with the following token allocations: {100, 100, 1}. Surprisingly, the 1-unit stake is actually pretty powerful in voting scenarios where the two 100-units token holders are at odds and acts as a tie breaker.

In fact, in this governance system having 1 unit of voting power is just as good as having 100 units. In terms of influence, *every* token holder can overturn the outcome in 50% of possible scenarios.

But in this allocation of tokens, {100, 100, 100, 1}, the 1-unit token holder will never influence a single vote. The decisiveness, and therefore value, of his stake is 0.

Decisiveness has many other interesting properties, including enabling analysis on what governance stakes are advantageous and being able to measure the “dictatorship-ness” of a system.

Today, @MolochDAO, has 24 members — 22 have 100 voting shares, 1 member has 24 shares, and the summoner (@ameensol) has an account with 1 share.

Decisiveness analysis produces the following results for @MolochDAO.

1. The summoner share of 1 unit, by itself, has no power and is worth 0.

2. Today, the 24-share member has just as much influence as a 100-share member despite having a lower-valuation stake.

3. If a new, 100-share member joins the DAO, the 24-share stake will be rendered worthless and will never decide a vote because the number of 100-share tokenholders will be odd.

4. The 24-share member should ragequit to get value or increase to at least 100 shares in order to conserve their decisiveness.

5. An existing member looking to increase their stake should not increase by less than 75 shares, or they will not get any benefit in influence.

6. A @MolochDAO member who increases their shares to 200 will exert elevated influence over other members. Their decisiveness will in fact be more than 2x the decisiveness of 100-share stakes.

Overall, valuing governance tokens by relative influence over votes opens up a new class of valuation conundrums:

1000 users @ 1 unit of voting power each are less influential than 1 user @ 1000 shares. So 1 unit of the currency is worth 0, but together their value increases...

If markets begin to value governance tokens based on their influence, then this has some interesting and negative implications for liquidity of governance tokens (perhaps mitigated by lending.)

Overall, this analysis may help inform design of voting systems using discrete distributions of tokens. I am looking forward to community thoughts and feedback. /end

@coinfund_io @coinfund_al @devinawalsh @HeyEdgeware @dillchen @AragonProject @lkngtn @cburniske @decredproject @daostack @koeppelmann @MatanField @JacobArluck @VladZamfir @MakerDAO