The Option the Pool Sold: Concentrated Liquidity as Implicit Strangles

Every Uniswap v3 position is a short options position in disguise. When you deposit liquidity in range [Pa, Pb], you are selling the market a strangle — a short put struck at Pa and a short call struck at Pb — and collecting fee income as the premium. The market “exercises” by driving price outside your range, converting your mixed position into a single asset at a loss relative to holding. What the LP calls impermanent loss, the options desk calls assignment.

This is not an analogy; it is an identity. The payoff curves are equivalent. The Greeks follow the same formulas. And the same Black-Scholes machinery that prices a vanilla strangle prices the implicit premium an LP should demand to break even. Most LPs never run that calculation — and some pools systematically underpay.

The Geometry of a Bounded Range

Uniswap v3 LPs choose a price range [Pa, Pb] and deposit a mix of two tokens. While spot price S sits inside the range, both tokens earn fee income proportional to the volume passing through. As S drifts toward Pa, the position rebalances continuously toward 100% of the base asset (ETH in an ETH/USDC pool). When S crosses Pa, the LP holds only ETH — fully exposed to further downside — and earns nothing. The mirror holds at Pb: cross it, and the LP holds only USDC, missing any further upside.

The value function for a normalized position (V = 1 at initial price S₀) is:

V(S) =
  L × (1/√Pa − 1/√Pb) × S          if S < Pa   (100% ETH)
  L × (2√S − √Pa − S/√Pb)           if Pa ≤ S ≤ Pb
  L × (√Pb − √Pa)                   if S > Pb   (100% USDC)

where the liquidity constant L = 1 / (2√S₀ − √Pa − S₀/√Pb) ensures V(S₀) = 1.

Plot this against a straight HODL portfolio and the gap is impermanent loss — always non-positive, worst at the range boundaries, and steepest in the tails. That concave, tent-shaped payoff is the signature of a sold strangle.

Short Put at Pa, Short Call at Pb

An options strangle sells both a put and a call on the same underlying with different strikes. The seller collects upfront premium; the buyer profits when the underlying moves enough to cover the premium. The seller’s maximum gain is exactly the premium collected; the loss is unbounded on the call leg and substantial on the put.

Map this onto the CLMM position:

The analogy is not loose: the CLMM value function is the exact payoff of a short strangle (plus a residual HODL component that cancels out in the IL calculation). This was formalized by the Panoptic team, who built a protocol on the equivalence.

Narrow range = tight strangle. A ±5% range sells an almost at-the-money strangle: high premium per dollar at risk, but any meaningful move assigns it. A ±50% range sells a wide-wing strangle: low premium, very unlikely to be assigned, but far from the capital-efficient core of the position. The width paradox explored in an earlier article — where tighter ranges earn higher fee APY but rebalance more frequently — is really the same phenomenon described in options language: a tighter strangle sells for more premium but gets exercised more often.

The Greeks of a CLMM Position

Options Greeks quantify price sensitivities. A CLMM position has them too.

Delta (dV/dS): the derivative of LP value with respect to price equals the amount of ETH in the position. For the in-range case:

Δ = dV/dS = L × (1/√S − 1/√Pb)

This is non-negative and bounded by [0, L × (1/√Pa − 1/√Pb)]. An LP never holds a short ETH position — delta always lies between 0 and 1 (the full ETH amount). At midrange, the position is approximately 50% ETH and 50% USDC: delta ≈ 0.5.

For a ±15% range on ETH ($1,661 spot, Pa = $1,411, Pb = $1,911), the initial delta is:

L = 1 / (2√1661 − √1411 − 1661/√1911) ≈ 0.168
Δ = 0.168 × (1/√1661 − 1/√1911) ≈ 0.168 × (0.0245 − 0.0229) ≈ 0.00269

In dollar terms, each $1 increase in ETH price increases LP value by $0.00269 on a $1 notional — equivalent to holding roughly 46% ETH and 54% USDC, consistent with a position that starts near the center of its range.

Gamma (d²V/dS²): the second derivative of LP value is always negative in the in-range region:

Γ = d²V/dS² = −L / (2S^(3/2)) < 0

Negative gamma is the defining characteristic of a short options position. It means the LP position is concave: large moves in either direction hurt more than a linear position would. The LP loses relative to HODL regardless of direction — the only question is how large the loss is. At $1,661 spot:

Γ = −0.168 / (2 × 1661^1.5) ≈ −0.168 / 135,400 ≈ −1.24 × 10⁻⁶

A $100 ETH move (from $1,661 to $1,761) changes the position’s delta by approximately Γ × 100 ≈ −1.24 × 10⁻⁴. This is small in isolation, but cumulative over many moves it produces the characteristic concave payoff.

Negative gamma also implies short vega: when realized volatility rises, the LP’s expected loss relative to HODL increases. This is why impermanent loss is sometimes described as “volatility tax” — the LP is short volatility.

Pricing the Strangle: What Black-Scholes Says

If the LP is writing a strangle, what is the fair premium? The Black-Scholes formula for a European strangle at strikes [Pa, Pb] with spot S₀, implied volatility σ, and time T:

Fair premium = BS_Put(S₀, Pa, T, σ) + BS_Call(S₀, Pb, T, σ)

At current ETH parameters — spot $1,661, ±15% range (Pa = $1,411, Pb = $1,911), σ = 54% realized vol from a 50-hour window, T = 30 days = 0.082 years, r = 4%:

For the put at Pa = $1,411 (−15% OTM):

d₁ = (ln(1661/1411) + (0.04 + 0.54²/2) × 0.082) / (0.54 × √0.082) ≈ 1.152
d₂ = 1.152 − 0.155 ≈ 0.997
BS_Put ≈ 1411 × e^(−0.003) × N(−0.997) − 1661 × N(−1.152) ≈ $17.10

For the call at Pb = $1,911 (+15% OTM):

d₁ = (ln(1661/1911) + (0.04 + 0.54²/2) × 0.082) / (0.54 × √0.082) ≈ −0.808
d₂ = −0.808 − 0.155 ≈ −0.963
BS_Call ≈ 1661 × N(−0.808) − 1911 × e^(−0.003) × N(−0.963) ≈ $28.50

Total fair strangle premium: $17.10 + $28.50 = $45.60, or 2.75% of the $1,661 notional for one 30-day period. Annualized: 33.4% APY.

This number has a direct interpretation: a rational LP in a ±15% ETH range should expect to earn at least 33.4% APY in fee income just to break even against the expected impermanent loss over time — before gas costs.

The call leg is larger than the put leg ($28.50 vs $17.10) because ETH’s lognormal distribution assigns more probability to large upward moves. An LP in a symmetric percentage range is slightly more exposed to upward blowouts than downward ones — the price-jump probability in log-space is asymmetric.

The Fee Income Gap

Do Uniswap pools actually pay the fair strangle premium?

Using the fee-APY formula from the width paradox analysis, a ±15% range carries a capital-efficiency multiplier of approximately 13.8×. Applied to each fee tier:

The fair BS premium at 54% realized vol is ~33% APY. The 0.05% pool at ±15% earns ~37%, a modest +4% excess — a thin positive edge. The 0.3% pool earns ~30%, meaning LPs are implicitly underpaid by ~3% relative to the fair volatility premium. The 1.0% pool is further behind.

Several caveats apply. Volume/TVL ratios vary by market regime; in high-vol periods, volume surges and fee APY can spike well above the fair premium. The Black-Scholes model assumes log-normal prices and constant vol — ETH has fat tails and vol clustering, so the true risk likely exceeds the BS estimate. And the analysis ignores LVR (loss-versus-rebalancing), the diffuse cost studied by Milionis et al. that an LP pays continuously to arbitrageurs regardless of range selection.

The practical takeaway: at current vol, the 0.05% pool at ±15% range is broadly fairly priced; the 0.3% pool is a mild negative expected value for the LP — unless the LP has a vol forecasting edge that justifies the position.

Panoptic: Making the Equivalence Explicit

The Panoptic protocol operationalizes the LP-as-options identity. Rather than depositing into a Uniswap pool directly, a Panoptic user can:

• Sell an option by providing a Uniswap LP position as collateral. The option seller receives streaming premium from the buyer.
• Buy an option by borrowing (renting) an LP position from a seller. The buyer pays streaming premium instead of an upfront lump sum.

The option price Panoptic charges is derived from the accumulated fees of the underlying Uniswap position — which is itself a proxy for realized volatility. When the pool is active and vol is high, fees accumulate fast and options are expensive; when the pool is quiet, fees trickle in and options are cheap. The fee mechanism is thus a form of oracle-free, continuously-updating implied volatility surface.

Panoptic’s contribution is eliminating the gap. Standard LP positions pay IL while simultaneously earning fees; the net is the implicit options payoff. Panoptic separates the two legs: options buyers explicitly pay for the IL exposure; options sellers explicitly receive premium for bearing it. The same risk transfer, priced transparently.

Agent Opportunities: Automated Premium Scanning

The options-pricing framing opens a precise quantitative edge for AI agents:

IV surface construction. For every Uniswap v3 pool, compute the implied vol surface from the pool’s fee income at each active tick range. A pool’s fee APY at range width W, combined with the Black-Scholes formula, implies an IV. If this implied vol diverges from realized vol, the pool is mispriced. Agents can rank pools by premium gap continuously, long/short the discrepancy via Panoptic, and harvest the spread.

Vol regime switching. The optimal LP strategy depends on the realized/implied vol relationship. When realized vol is high and fee APY is lagging (underpaying the fair premium), agents should LP in Panoptic as option sellers — receiving explicit premium rather than collecting fees. When realized vol is low and fee income is generous relative to the fair strangle price, direct LP positions deliver positive expected value.

Cross-pool arbitrage. The same ETH/USDC strangle is priced differently in the 0.05% and 0.3% pools because they attract different volumes. An agent holding the complete vol-surface picture across all active pools can route LP capital to whichever pool is most overpaying the strangle premium at the current tick range.

None of this requires AI in the sense of learned behavior — it is systematic quantitative analysis. But it does require running constantly, updating every time a new pool observation is available, and executing atomically. That is exactly the profile of an on-chain agent.

⬢ loading artifact…

Takeaways

• A CLMM LP position in range [Pa, Pb] is mathematically equivalent to a short strangle — short put at Pa, short call at Pb. Impermanent loss is the options-exercise payoff.
• Delta ranges from 0 (above Pb) to 1 (below Pa); gamma is always negative in-range. Both signatures match a sold options position.
• At ETH spot $1,661 and 54% realized vol, a ±15% 30-day strangle carries a Black-Scholes fair premium of ~2.75% of notional (33.4% APY).
• The ETH/USDC 0.05% pool at ±15% earns approximately 37% fee APY — a thin positive edge. The 0.3% pool earns roughly 30%, implying LPs are structurally underpaid at current vol.
• Panoptic formalizes the equivalence by separating the fee leg (premium) from the IL leg (assignment), creating explicit oracle-free options on any Uniswap v3 pool.
• AI agents can exploit the premium gap by building implied vol surfaces from pool fee data, identifying mispriced positions, and routing capital to the most favorable strangle seller.