How to break the power law of wealth?

Draft (to write complete elaboration)

1. Stakeholder selection: Through kleros, Score +5 to -5.

2. Stakeholder empowerment: Once selected, the stakeholder receives non-transferable voting tokens based on their score.

Then the selected stakeholder receives non-transferable governance tokens based on their score:

3. Conviction voting: That stakeholder uses those non-transferable tokens in conviction voting to allocate funds. The longer they maintain conviction behind a proposal, the more voting power their locked tokens accumulate.

Voting power of stake holder = sqrt( Goverance tokens stake * time governance token are locked )

It breaks the power law:

This is real progress — this formula, if implemented correctly, actually achieves what the last several iterations didn’t: it makes final voting power concave in stake rather than linear.

The math actually breaks scale-invariance — final influence grows as sqrt(stake) rather than stake, which measurably flattens the tail rather than just adding honesty (Kleros) or blocking secondary markets (non-transferability) on top of an unchanged linear core. But it only holds if:

• governance tokens are aggregated per identity before the sqrt, not per conviction-lock position, and
• identity is Sybil-resistant, so “per identity” means something.

Yes, kleros is good at providing sybil-resitance identity for stakeholders.

Why conviction voting fails?

The duration bonus makes it worse, not better

This is the part I’d flag most strongly. Locking for longer duration to boost selection probability is exactly the vote-escrow (ve-tokenomics) model — the same pattern used by Curve’s veCRV and its many forks. That model has a well-documented empirical failure mode: illiquidity is cheap for the wealthy and expensive for everyone else. A whale can lock a large position for 6 months without financial strain; a smaller holder who needs liquidity for near-term expenses can’t afford to lock as long, even if they wanted equal influence. So the duration multiplier compounds the stake advantage rather than offsetting it — it’s an additional axis on which wealth buys more influence, not less. The “ve-token wars” and bribery markets that emerged around Curve are the real-world evidence for this dynamic; I’d treat any lock-duration-weighted mechanism as suspect for an equality goal unless the multiplier is capped or made identity-bound rather than capital-bound.

Kleros and truth discovery

Juror draw probability is still weighted by stake — same mathematical shape as your current conviction-lottery, just with a different name. A whale with 1,000x the stake of a small holder still gets ~1,000x the expected number of draws over time. Everything I said in my previous message about expected-value convergence under weighted lotteries applies identically here: randomness changes variance, not the expectation, and the expectation is still linear in wealth.

Why it still works despite stake-weighted selection

The Schelling mechanism’s guarantee isn’t “the population of jurors is representative of wealth” — it’s “whichever jurors get drawn are individually incentivized to report their honest signal rather than a strategic one.” Those are independent properties. Selection weighting determines who answers; the coherence-payoff structure determines whether they answer truthfully once selected. A whale who gets drawn 1,000x more often than a small holder is still, each time they’re drawn, better off voting their genuine assessment of the proposal (−5 to +5) than voting something self-serving — because the payoff depends on matching the eventual majority, not on the size of their stake or their relationship to the proposal.