Skip to main content

Synthesis & Recommendations

Updated: 2026-01-20 (Arbitrum era).

This is the “single doc” overview that ties together:

  • What we measured on-chain (delegation flows, churn, withdrawals, post-withdraw transfers)
  • What we learned from solution proposals (Lisar, IDOL/Arrakis, Tenderize)
  • What tends to work across protocols for growing small participants
  • What we recommend Livepeer should do next (and how to measure it)
Deep dive: outflows + retention

The core evidence pack behind churn and “cashout” claims.

Design space: incentives + tokenomics

Practical mechanism options and sybil constraints.

TL;DR (what we believe is true from evidence)​

  1. New delegator inflows are trending down on Arbitrum (first-time bonders drop year-over-year).
  2. Bonded stake is extremely concentrated: small wallets dominate count but hold negligible stake; security-relevant growth is bracket growth (1k–10k, 10k+).
  3. Recent growth is mostly in larger brackets (e.g. 10k+ stake grew materially in 2025 while mid-retail bands shrank).
  4. Retention is not “instantly everyone farms and exits”, but early churn exists and increases over months.
  5. The “mass sybil cashout wave” hypothesis is not supported by daily unique-unbonder counts (peaks are ~O(100), not thousands), though small trickle sybil behavior is still possible.
  6. Any “give smaller addresses higher APR” mechanic is sybilable unless tied to a uniqueness primitive (identity / proof) or made economically expensive to split; prefer utility + liquidity + retention gating.

1) What we measured (hard evidence)​

Chain / contract / method​

  • Chain: Arbitrum One
  • Livepeer staking contract: BondingManager proxy 0x35Bcf3c30594191d53231E4FF333E8A770453e40
  • Method: RPC eth_getLogs scans (no explorer API keys)
  • Window scanned: 2022-02-11 → 2026-01-17 (see research notes)

Primary sources in this repo:

  • Delegation flows + sybil/churn research: research/livepeer-delegator-outflows-research.md
  • Tokenomics/program ideas grounded in the data: research/livepeer-delegator-incentives.md
  • Board view (brackets + inflow/outflow + delegate gain/bleed): research/delegation-board.md
  • Bracket time series + concentration: research/delegator-band-timeseries.md

Primary artifacts (raw-ish outputs) referenced by those notes:

  • artifacts/livepeer-bm-scan-arbitrum-v2/retention_report.md
  • artifacts/livepeer-bm-scan-arbitrum-v2/unbond_report.embedded.md
  • artifacts/livepeer-bm-scan-arbitrum-v2/earnings_report.md
  • artifacts/livepeer-bm-scan-arbitrum-v2/outflow_destination_classification_top50.md

Key network-level results (Arbitrum)​

From research/livepeer-delegator-incentives.md and research/livepeer-delegator-outflows-research.md:

  • First-time bonders (“new delegators”) are trending down:
    • 2022: 2,213
    • 2023: 1,492
    • 2024: 685
    • 2025: 482
  • Churn / retention (eligible-only):
    • Unbond within 30d: 10.78%
    • Unbond within 90d: 17.24%
    • Withdraw within 30d: 6.32%
    • Withdraw within 90d: 12.56%
  • Sybil-cashout wave not supported by daily unique exits:
    • Max unique unbonders/day: 122
  • Rewards are highly concentrated:
    • Total rewards claimed: 17.495M LPT
    • Top 10 earned: 33.69%
    • Top 100 earned: 75.51%
  • A large fraction of rewards appears to be withdrawn (“stake leaves BondingManager”):
    • Proxy “rewards withdrawn”: 7.198M LPT (upper bound 11.292M LPT)
  • Post-withdraw destinations (top cashout cohort) skew heavily toward leaving:
    • Bridge/burn-to-zero and EOA transfers dominate the observed flows.

Interpretation: the core growth issue is more plausibly liquidity + UX + weak retention hooks, not a sudden “sybil apocalypse”.

2) What “delegation problem” are we solving?​

There are two overlapping but distinct goals that often get conflated:

  1. Protocol security / decentralization
    • Primarily driven by how much stake is bonded and how concentrated it is.
  2. Number of delegator addresses
    • A proxy for decentralization and community participation, but easy to game if you reward “# of accounts”.

We should treat the canonical success metrics as:

  • Net-new bonded stake (bond − unbond − withdraw)
  • New participants (first-time bonders) and/or new stake participants (e.g., LST depositors) with clear definitions
  • Retention (still bonded/participating after 30/90/180d)
  • Concentration / decentralization impact

This aligns with docs/problem.md + docs/rubric.md.

3) What we learned from specific solutions​

See docs/scoreboard.md for the roll-up.

3.1 Lisar — “Fiat delegation” (treasury-funded)​

Where: solutions/lisar/

Finding (strictly vs the proposal KPI): adoption appears far below the stated target so far.

  • Proposal KPI: 500–1,000 active delegators in ~4 months.
  • Observed (as of mid-Jan 2026): ~13 active delegators, ~14 ever bonded.

Implication: anything claiming “we can onboard delegators” needs:

  • A clear funnel (users → funded → bonded → retained)
  • Clear on-chain definitions (what counts as “active”)
  • A plan for retention and preventing “funded but never bonded” drop-off

3.2 IDOL / Arrakis — Improve DEX liquidity (treasury-funded)​

Where: solutions/ydol/

Finding: the liquidity UX problem is real, but a large LP deployment has economics + incentive-alignment risks.

  • Slippage is severe at $25k–$50k sizes in current Arbitrum LPT/WETH v3 pools (sell-side cliff exists).
  • Current DEX flow is modest (fees may not offset IL/LVR), so the treasury takes most downside.
  • Recommendation in the dossier: treat as a pilot with tranche funding + explicit KPIs/stop conditions.

Implication: DEX liquidity can reduce friction for larger actors, but it’s not automatically a small-delegator growth engine unless paired with onboarding + liquidity rails (e.g., an LST).

3.3 Tenderize — tLPT liquid staking (historical)​

Where: solutions/tenderize/

Finding: Tenderize shows what LST adoption “looked like” historically on Livepeer:

  • 2,518 unique depositors / 2,786 deposits / ~92.19k LPT deposited
  • Deposit sizes were mostly tiny (majority under 10 LPT total deposited)

Critical nuance:

  • Tenderize did not create 2,518 new protocol-level delegators; the tenderizer contract is the delegator.
  • It did create thousands of stake participants with LPT exposure.

Implication: “liquid staking” can attract many small participants, but if Livepeer’s goal is “more delegator accounts on BondingManager”, LSTs do that only indirectly (unless you add additional mechanics).

4) Cross-protocol patterns that tend to grow small stake participants​

This section is “pattern-level” (not yet a full on-chain replication study across protocols). It’s included because these patterns repeatedly show up when “retail participation” increases in staking systems.

For additional precedent notes, see: /research/cross-protocol-tokenomics-experiments.

Pattern A: Liquid staking + DeFi composability​

Across major ecosystems (e.g., Ethereum, Solana, Cosmos), liquid staking protocols (Lido, Rocket Pool, Marinade, Stride, etc.) grew by:

  • Pooling stake (removes minimums / operational complexity)
  • Issuing a liquid token (used for trading, LPing, borrowing)
  • Building integrations (money markets, DEX liquidity, vaults)

Why it matters for Livepeer:

  • It’s sybil-neutral by design (rewards accrue pro-rata, not per account).
  • It directly attacks the biggest retail blocker: “I want yield, but also liquidity.”

Pattern B: Retention hooks (time + locks)​

Incentives that vest over time (or increase with continuous participation) reduce “farm-and-dump” behavior.

Practical version (without a core protocol overhaul):

  • Treasury bonus vests over N rounds; early unbond forfeits.

Pattern C: Onboarding subsidies (gas, UX, defaults)​

For small participants, friction costs dominate.

  • Subsidize the first delegation (or gasless onboarding) but only pay out after retention.

Pattern D: Avoid per-address progressive rewards without a uniqueness primitive​

If you give better APR to “smaller balances per address”, whales can split.

If you want a “small holder boost”, you typically need one of:

  • Identity attestation (with UX/privacy tradeoffs)
  • Work/usage-based gating (makes sybils costly)
  • Strong retention + minimums (reduces profitability of splitting)

5) Recommendations (Livepeer-specific, prioritized)​

5.0 Define canonical KPIs and publish them​

Before shipping more programs, standardize:

  • “New delegator” definition (first Bond)
  • Retention definition (unbond/withdraw triggers; censoring window)
  • “Net new stake” (net bond/unbond/withdraw)
  • Concentration metrics (top-N share, Gini proxy)

5.1 Ship an audited stLPT / wstLPT (liquid staking) with conservative security posture​

Design goals:

  • Minimal trust surface (avoid unnecessary upgradeability).
  • Handle Livepeer unbonding delay cleanly (withdraw queue / lock IDs).
  • Provide wstLPT (non-rebasing wrapper) for DeFi integrations.

Tenderize is a useful blueprint for the core staking mechanics (burn-before-unstake), but also a cautionary tale on proxy upgrade/admin risk.

5.2 Bootstrap LST liquidity (time-boxed) and measure​

If you want LST adoption, liquidity is not optional.

  • Incentivize stLPT/ETH (and/or stLPT/LPT) LP for a limited period.
  • Use explicit stop/go criteria (volume, depth, spread, LP retention).

5.3 Run a “new delegator” acquisition program that is retention-gated​

Use a treasury-funded program that:

  • Targets first-time bonders (or first-time LST depositors) with a minimum stake.
  • Vests over time; early unbond forfeits.
  • Uses non-linear reward curves (e.g., sqrt(stake)) + caps to reduce whale dominance.

5.4 If you want a “small holder boost”, treat identity as optional/high-tier​

There is no “0 friction, perfectly unique human” primitive.

If Livepeer wants aggressive per-person boosts, make it a higher-tier bonus:

  • baseline incentives: sybil-neutral (stake-proportional + retention)
  • additional boost: requires identity proof (Gitcoin Passport / World ID / equivalent), or credible “proof of work/usage” that makes sybil costly

One candidate design for “proof-of-work/usage gating” is a delegator tiers ladder: /solutions/delegator-tiers.

5.5 Treat DEX liquidity programs (Arrakis-style) as pilots, not forever programs​

If the ecosystem pursues DEX liquidity improvement, use the guardrails from solutions/ydol/:

  • tranche funding, explicit KPIs, stop conditions
  • minimize incentive misalignment (avoid paying on gross fees if possible)

6) What we still need to research next​

  1. Cross-protocol replication study: pick 2–3 protocols and quantify “small participant growth” around LST launches (to the extent feasible).
  2. Livepeer LST design space: audited building blocks, governance model, upgrade posture, and DeFi integration plan on Arbitrum.
  3. Sybil-resistant uniqueness options for any “small holder boost” (UX and legal constraints included).