How concentrated ranges, impermanent loss, and rebalancing policies shape LP outcomes in Uniswap‑style AMMs.
Concentrated liquidity boosts fees but introduces negative gamma. We analyze periodic vs trigger‑based rebalancing and how to size ranges.
Narrow ranges behave like leverage when price stays in‑range.
Recenter when impermanent loss exceeds fees earned since last rebalance.
Base wide range for safety + small active range for alpha.
Uniswap v3 lets LPs concentrate liquidity in a price band (often around TWAP), boosting fee density like leverage, but increasing exposure to impermanent loss — a classic negative‑gamma tradeoff.
Center ranges near the current price or TWAP to keep capital “in‑range” and earning. Narrower bands raise fee income per dollar deployed in range‑bound markets.
When price leaves the band, the position converts into the losing asset and stops earning. Concentration amplifies IL — similar to being short gamma.
Tight ranges = high fee density & faster out‑of‑range risk; Wide ranges = lower fee density & more time in‑range. Stable pairs can run tighter than volatile pairs.
Two approaches are evaluated and compared to a passive full‑range baseline. Rebalancing recenters the band and realizes any IL accumulated since the last move.
| Strategy | When it moves | Pros | Cons | Best conditions |
|---|---|---|---|---|
| Periodic (TWAP) | Fixed interval (e.g., weekly), centered on latest TWAP | Keeps liquidity near the market; simple to implement | Realizes IL frequently; can “chase” price; gas churn | Oscillating or range‑bound markets |
| Trigger (IL > Fees) | On‑demand only when IL since last rebalance exceeds fees | Adaptive; fewer, more meaningful resets; reduces prolonged divergence | May rebalance “unnecessarily” if the market mean‑reverts soon after | Trending or regime‑shifting markets |
| Passive full‑range | Never | Minimal management; no realized IL from rebalancing | Lowest fee density; capital spread thin | Often fares better than active LPs in strong trends (though still lags HODL) |
We simulate the two active strategies versus a passive baseline on ETH‑USDC (0.3% tier). Positions start 50/50 by value. Periodic recenters weekly around TWAP; Trigger fires when IL > fees since the last move. Fees accrue as a function of simulated swap volume (scaled with volatility).
| Scenario | Price behavior | Outcome highlights |
|---|---|---|
| Range‑bound | Oscillation within ±10% around start | Concentrated ranges earn higher fees; trigger rarely fires; active > passive |
| Uptrend | Steady +20–30% drift | All LPs < HODL; passive slightly > active (less realized IL) |
| Downtrend | Steady decline | All LPs < holding stables; passive again slightly > active |
| Shock jump | +50% step‑move then new level | Trigger recenters sooner; still trails HODL; periodic can miss the move entirely |
Both active policies outperform full‑range by concentrating where trades happen and minimizing idle time.
LPs underperform HODL as the strategy sells the winner while climbing; passive typically ends highest among LP variants.
LPs accumulate the falling asset; fees only partially offset IL. Passive edges active by avoiding frequent realized losses.
Trigger detects IL > fees and re‑centers faster, reducing time out‑of‑range, but absolute performance still trails simple holding.
Blend a wide “base” position for safety with a small active narrow band for alpha. This moderates drawdowns while keeping upside to fee density.
Scale range width with volatility. Tighter for stable pairs; wider for volatile assets unless you can actively manage.
Selected materials that inform the analysis and framing.