Financial engineering, six years later

The original idea

Back in early 2020, The Defiant wrote about a DeFiZap recipe I was obsessed with: deposit ETH in one transaction and walk away as a liquidity provider who still kept full upside on ETH. The zap split your deposit ~66% into an ETH/DAI Uniswap pool and ~34% into a 2× leveraged ETH long on Fulcrum. An LP only gives ~half the directional exposure of holding, so it underperforms when ETH rips — that gap is impermanent loss. The leveraged long bolted the missing exposure back on: net ~1.0× ETH, full upside, and you collected swap fees the whole time. Then Fulcrum/bZx got exploited and the recipe became folklore.

What changed

Hedger is the same idea, rebuilt. Fulcrum → Index Coop ETH2x, minted atomically from a single asset via a verified FlashMint. Ethereum L1 → Base. One DEX → Uniswap v4 or Aerodrome, your pick. And a genuine one-transaction zap in and out, returning a single position NFT.

The mechanism

A factory deploys instruments. The IL Eliminator is the orchestrator: it computes the split, mints the leverage leg through an adapter, adds liquidity through a venue adapter, mints your position NFT, and refunds dust — all in one transaction, with slippage, deadline, reentrancy, and pause guards. Adapters are pluggable, so new venues or leverage sources drop in without redeploying the core.

The math (and the honest caveat)

A 50/50 LP has ~0.5 delta to the volatile asset. To restore a target net delta t with a leverage token of ratio L and pool delta d: allocate c_LP = (L − t)/(L − d) to the LP and c_lev = (t − d)/(L − d) to the leverage leg. For L = 2, d = 0.5, t = 1.0 → ⅔ LP, ⅓ leverage. Impermanent loss is IL(r) = 1 − 2√r/(1 + r).

Here is the part the name oversells, and I won't hide it: an LP is short gamma.Restoring delta to ~1× does not remove impermanent loss — IL is a convexity (gamma) cost, and a leveraged long is, to first order, linear. What you've actually built is staying ~1× long while selling volatility for LP fees — a covered-call-shaped payoff, not a neutral hedge. It hedges the delta of an LP's underperformance, not the gamma.

Verified results

This isn't a whiteboard. The contracts are fork-tested against live Base state: a full one-tx zapIn → zapOut of 1 WETH recovered 0.988 WETH in a flat market (98.8%)— the ~1.2% being Index Coop's 0.10% mint + 0.10% redeem fees plus the UniV3 and Aerodrome swap round-trips. The ETH2x leg round-trips at 99.5%, the Aerodrome LP leg at 99.7%, and net delta is restored to ~1.0 at entry.

But those are just the round-trip frictions. The real question is what happens when ETH actually moves — so here's the projected payoff vs simply holding ETH, or plain LPing. At +30% Hedger tracks close to a full ETH position (~+29.5%) while a plain LP captures only ~half (+15.4%) — that gap is the impermanent loss it hedges. At +50%it's ~+48.4% vs +24% for the LP. The flip side: on the way down the leverage amplifies losses and the plain LP cushions more, so this is a directional, short-volatility instrument — drag the sliders to see any move.

An illustrative model (single monotonic move; assumptions listed under the chart) — not a guarantee, and distinct from the fork-verified frictions above.

Security

One thing Hedger does not have is an admin backdoor. The instrument tracks every leg token custodied for open positions, and the emergency rescue() is provably unable to move more than balance − custodied — so an owner, even a compromised one, cannot drain user positions. Ownership is intended to sit behind a TimelockController; the contracts are Ownable2Step, Pausable, and guarded against reentrancy. It is still pre-audit software.

Risks

• It hedges delta, not gamma → short volatility; choppy markets bleed.
• Net long, downside amplified, not delta-neutral.
• Volatility drag from the constant-2× token, stacked on a short-gamma LP.
• Leverage rebalance/liquidation, correlated with LP losses in the tail.
• ~3.65%/yr streaming + 0.10% mint/redeem + swap slippage; win only if fees + appreciation beat carry + drag.
• Counterparty dependency on Index Coop; pre-audit contracts.

Contract addresses (Base)