Skip to main content

General Directions (What to Do Next)

This section is intentionally “directional”: it’s meant to help stakeholders align on what kinds of interventions are likely to move the important brackets (1k–10k, 10k+) and how to measure progress.

Direction 1 — Increase LPT utility (so larger participants can enter/exit sanely)

Why: the data implies that security-relevant growth is bracket growth, and big actors need deep liquidity + usable rails.

What it could include:

  • A conservative liquid staking path (stLPT / wstLPT), with integrations (borrowing, LP, vaults).
  • Targeted DEX liquidity improvements with explicit KPIs and stop conditions (see IDOL/Arrakis dossier).

How to measure:

  • Growth in 1k–10k and 10k+ wallet counts and bonded LPT
  • Delegate/orchestrator concentration (top-10 share, Nakamoto 33/50)

Direction 2 — Make incentives retention-gated (not “pay per address”)

Why: paying “small balances per address” is sybilable; retention gating is harder to game and aligns with long-term stake.

What it could include:

  • Treasury bonuses that vest over 90–180 days and are forfeited on early unbond/withdraw.
  • “Re-delegate and stay” campaigns (time-boxed) targeting the 1k–10k bracket.

How to measure:

  • 30/90/180d retention of new entrants in 1k–10k and 10k+
  • Net bonded stake change (not just churn)

Direction 3 — Treat “small delegators” as a UX/onboarding problem, not a security lever

Why: small wallets are important for community participation, but they don’t move bonded stake unless you also solve UX and liquidity.

What it could include:

  • Gas sponsorship + “one-click delegation” funnels (Lisar-like), but with explicit stake/retention KPIs.
  • Programs that nudge users to graduate from “single-digit LPT” toward meaningful stake bands over time.

How to measure:

  • Funnel metrics: funded → bonded → still bonded at 90d
  • Graduation: share of cohort that crosses 100, 1k, 10k thresholds

Direction 4 — Increase independent high-stake delegates (competition + decentralization)

Why: decentralization is primarily a function of stake distribution across delegates, not just delegator accounts.

What it could include:

  • Programs that help new/independent orchestrators reach meaningful stake (without concentrating to a single incumbent).
  • Transparent “delegate quality” dashboards so larger delegators can allocate rationally.

How to measure:

  • of delegates ≥ 100k and ≥ 1m bonded stake

  • Nakamoto coefficients over time (33% / 50%)

Direction 5 — Reduce inflation extractability (so yield doesn’t become structural sell pressure)

Why: if large actors can run delta-neutral staking, inflation rewards can get routinely withdrawn and routed off-chain, creating reflexive price pressure.

What it could include:

  • Protocol-level: separate principal vs rewards and make the reward component time-gated (vesting) and/or penalized on early exit (avoid flat “principal exit taxes”).
  • Program-level (if protocol change is too heavy): bonuses in an escrow contract with vest/forfeit; this helps program incentives but does not fix base-inflation extractability by itself.

How to measure:

  • Reward-only exits over time (requires principal-vs-reward separation in staking accounting)
  • Bridge-out → exchange routing changes (see /research/l1-bridge-recipient-second-hop)
  • Retention curves for 1k–10k and 10k+ cohorts

Where the evidence lives