Skip to main content

Livepeer Delegation Research

This is a working research repo: we prioritize verifiable on-chain evidence and a consistent rubric for judging solutions.

Evidence Packs

Outflows + retention research

Who exits, when they exit, and what happens after withdraw (including “cashout” bounds).

Cashout routing (bridge-outs → exchanges)

Follows Arbitrum bridge-outs to L1 recipients and one hop further to labeled exchange endpoints (best-effort).

Timing traces (withdraw → bridge → L1 → exchange)

Harder-to-explain fingerprint: WithdrawStake → bridge-out → L1 escrow receipt → tight-window exchange routing (best-effort).

Buy-side proxies (exchange outflows → bonders)

On-chain buy-pressure proxies: labeled CEX outflows on L1, and whether recipients bridge + bond on Arbitrum.

Incentives + tokenomics notes

Design space for small-delegator growth (and what is sybilable vs sybil-neutral).

Rewards vs withdrawals (time series)

Monthly/yearly LPT rewards claimed vs WithdrawStake (sell-pressure proxy; includes principal).

Reflexivity + yield extraction (mitigations)

Delta-neutral extraction thesis + mitigation primitives (escrow/vesting, reward-only exit locks/penalties).

Extraction fingerprints (on-chain proxies)

Top-50 “cashout-heavy” wallets: rewards claimed vs withdrawn, post-withdraw routing, and which wallets remain bonded.

Solutions: Lisar

Treasury program outcomes + KPI alignment vs the stated goals.

Solutions: IDOL / Arrakis

DEX liquidity-focused proposal analysis (mechanism, risks, and measurement).

Solutions: Tenderize (tLPT)

Historical Livepeer liquid staking: adoption evidence, mechanics, and failure modes.

Solutions: Delegator tiers (stake + contribution)

A tier ladder that targets small/mid delegators using verifiable usage and contributions (sybil-costly).

Cross-protocol experiments

What worked (and what backfired) in crypto incentives, with takeaways for Livepeer.

DePIN benchmark: Filecoin vs Livepeer

Filecoin’s lock/burn primitives (on-chain) contrasted with Livepeer’s extraction proxies.