LP strategies: picking a range, rebalancing and when not to LP

Every liquidity pool strategy is a handful of decisions: which pool, how wide a range, when to move it, how often to harvest, and when to stay out. This guide walks through each one with worked numbers, including why the most common rebalancing rule quietly loses money.

lp.xyz team12 min read
Strategy

In short

A good LP strategy starts with the pool, not the range: high fee volume relative to TVL, a fee tier that matches the pair's volatility, and a token you would be happy to end up holding.

A narrower range earns a bigger share of fees while price is inside it, and suffers more impermanent loss and more time out of range when price moves. A wider range is the reverse.

Re-centring a concentrated position every time it goes out of range turns a temporary loss into a permanent one; rebalance on a rule with a delay and a cap, or not at all.

Harvest when the claim is worth clearly more than its cost, use one-sided ranges (walls) as limit orders, and don't LP into strong trends or illiquid tokens.

What is an LP strategy?

An LP strategy is the set of rules you follow for providing liquidity to an automated market maker (AMM): where you deposit, over what price range, and what you do as price moves. With concentrated liquidity (Uniswap v3 and v4, Orca and Raydium CLMM, Meteora DLMM, PancakeSwap v3 and Infinity), you choose a price range, and your position only trades and earns fees while price is inside it. Outside the range it sits idle, fully converted into one of the two tokens. That single fact drives every decision below.

  • Pool: which pair, which venue and which fee tier.
  • Range: how wide, and where relative to the price.
  • Maintenance: when to rebalance, harvest and compound.
  • Risk: whether to hedge, and when not to provide at all.

Step 1: how do you pick a liquidity pool?

The best liquidity pool for a strategy is one where trading volume is high relative to the liquidity competing for its fees, the fee tier fits the pair, and the tokens are ones you are willing to hold. Range tuning can't rescue a pool with no volume or a token that is about to go to zero.

Volume / TVL as a fee-yield proxy

The ratio of 24-hour volume to TVL (total value locked) tells you how hard each dollar in the pool is working. Multiply it by the fee tier to get the pool's fees as a share of its TVL. Say a pool did $2,000,000 of volume yesterday on $1,000,000 of TVL at a 0.25% fee: that is $5,000 of fees, or 0.5% of TVL in one day. Annualised naively, 0.5% a day is 182.5% a year, and that number is the reason to be careful:

  • It is backward-looking. One day of launch volume or one large trader can make a pool look far better than it will be next week.
  • It is before impermanent loss, which on a volatile pair can be larger than the fees.
  • A concentrated position earns its share of the liquidity that is active at the current price, not of total TVL. A narrow position can earn more than its share of TVL; an out-of-range one earns nothing.
  • Some of the fee goes to the protocol on many venues. Uniswap's fees page lists the current LP and protocol split per tier.

Fee tier and token risk

Higher fee tiers pay more per trade and suit volatile pairs, where LPs carry more inventory risk; low tiers suit stable pairs, where volume is large and price barely moves. Uniswap v3's standard tiers are 0.01%, 0.05%, 0.30% and 1%, while v4 lets a pool set any fee or a dynamic one. When the same pair exists at several tiers, compare their volume / TVL, not just the tier. Then check the token itself: whether the supply can still be minted or balances frozen, how concentrated the holders are, and whether the volume is many wallets or one. Most money lost around AMMs is lost there, not to impermanent loss. The LP risks page lists the signals lp.xyz checks.

Step 2: how wide should your range be?

Range width is the main trade-off in a concentrated liquidity strategy: a narrow range earns a larger share of fees while price is inside it, and loses more to impermanent loss (IL, the gap between your position's value and simply holding the tokens) and spends more time out of range when price moves. A worked example shows the size of the effect.

Say SOL is at $150 and you deposit $10,000 in a SOL/USDC concentrated pool over a $140 to $160 range. At $150 the position holds 32.2398 SOL and $5,164.02 USDC. Now SOL falls to $130:

$140-$160 range
Price is below the range, so the position is entirely SOL: 67.8750 SOL worth $8,823.75. Holding the original tokens would be worth $9,355.20, so the position is $531.45 (5.68%) behind holding, before fees. It also earns nothing until price comes back above $140.
Full range
The same 13.3% drop costs a full-range position 0.26% against holding its own starting tokens. It keeps earning, but its fee share per dollar is far smaller.

The narrow position needs fees worth more than 5.68% of its value, earned while in range, to beat holding. That is why narrow ranges suit pairs that genuinely range (stable pairs, or a token pinned by heavy two-way flow), and wider ranges suit anything that trends or gaps. The impermanent loss guide works through the formulas.

Narrow
Most fees per dollar while in range. Largest IL and fastest exit from range. Needs monitoring or a plan for when price leaves.
Wide
Lower fee share, more tolerance for movement, less maintenance.
Full range
Never out of range and the smallest IL for a given move, but the lowest fee share. The passive default, and the only option on constant-product pools.

Should you LP passively or actively?

Passive full-range LPing suits pairs you want to hold for a long time and don't want to manage: fees accrue, and on constant-product pools they compound into the reserves automatically. Active concentrated LPing earns more per dollar while it is in range, but every adjustment costs transactions and swap fees, and the decisions themselves can lose money. A useful test: if you can't say in advance what you will do when price leaves your range, choose the wider or passive position.

When should you rebalance a liquidity position?

Rebalance a liquidity position only when you have a reason to believe price will not come back, and the fees from the new range will outweigh what the move costs. The common rule, “re-centre the range whenever price leaves it,” fails that test in choppy markets, and the example above shows why.

Take the $140 to $160 position again, now all SOL at $130. A naive rule re-centres it at the same relative width: $121.33 to $138.67. To do that, the position sells 35.0508 SOL at $130 to fund the USDC side. Then SOL recovers to $150:

Left alone
Back inside $140-$160, the original position is worth exactly $10,000.00 again, the same as holding. The loss at $130 was temporary.
Re-centred
Price is now above $138.67, so the new position sold all its SOL on the way up and is entirely USDC: $8,963.69. The $1,036.31 gap is now permanent.

These figures are before fees on both sides, and the re-centre also pays a swap fee and transaction costs, so the real gap is larger. The cause is structural, not bad luck. A concentrated position's IL depends only on where price is relative to its range, so a round trip erases it. Re-centring closes the old position at the bottom, which realises the loss, and opens a new one whose own IL starts from the new price. In a market that swings back and forth, each re-centre sells low and the next one buys high.

Rebalancing rules that hold up better

Wait for a dwell time out of range, so a wick doesn't trigger a move. Cap moves per day. Re-centre in the direction of a move you believe in rather than every move. Compare the cost of the move (swap fee, transactions, realised IL) with the fees the new range is likely to earn before you make it.

How often should you harvest and compound?

Harvest LP fees when the claimable amount is clearly larger than the cost of claiming it; on concentrated pools fees sit apart from your liquidity until claimed, and each claim is a transaction. Where transactions are cheap, as on Solana, a modest threshold works. On Ethereum mainnet, frequent claims on a small position can cost more than they collect. Compounding (claiming and re-adding the fees to the same position) grows the position, but each compound is also a transaction, so it pays off only when the fees are large relative to that cost.

Constant-product pools need neither: fees accrue into the reserves, and you realise them when you remove. Uniswap v4 and PancakeSwap Infinity can claim and re-add in a single transaction. On lp.xyz, Harvest runs at a threshold you choose ($1, $5, $25 or your own) and a cadence (hourly, every 6 hours or daily), and skips a run rather than guess when a fee leg can't be priced. lp.xyz takes 5% of the native side of fees claimed through it, disclosed on the fees page, so factor that into your threshold too.

How do walls work as limit orders?

A wall is single-sided liquidity placed entirely above or below the current price, and it behaves like a limit order that earns fees while it waits. A buy wall holds only the quote token below the price and converts into the token as price falls through it; a sell wall holds only the token above the price and converts into the quote as price rises. Uniswap's concentrated liquidity docs describe the underlying behaviour: as price crosses a range, the position ends up entirely in one asset.

  • Your fill is an average across the wall's range, not one price.
  • A filled wall can unfill. The converted tokens stay in the pool as liquidity, so if price comes back through the range, the wall converts back. To keep the fill, withdraw after price trades through.
  • Walls need a concentrated pool. A constant-product pool always wants both tokens at the pool ratio.

On lp.xyz, a wall can be set to Withdraw on fill, which removes it once price has stayed past it for a confirmation window, or to Follow price, which rebuilds a wall the market has left behind after a dwell time, with a cap on moves per day. The single-sided liquidity guide covers walls venue by venue.

Can you make an LP position delta-neutral?

A delta-neutral LP strategy pairs a liquidity position with a short perpetual future (perp) on the volatile token, so price moves in one roughly cancel price moves in the other. It can work, and it is more involved than it sounds:

  • The hedge has to keep moving. An LP position holds more of the token as price falls and less as it rises, so the short must be resized as price moves. Each resize costs trading fees, and rebalancing the hedge realises the same kind of loss the IL represents.
  • Funding is a running cost or credit. Perps charge or pay a funding rate, which changes over time and can outweigh the LP fees.
  • The short can be liquidated in a sharp rally if it isn't collateralised generously, leaving the LP position unhedged at the worst moment.
  • It hedges direction, not the cost of volatility. What is left is roughly the fees minus IL, funding and trading costs. It is a bet that fees beat volatility, with the directional bet removed.

When should you not provide liquidity?

Providing liquidity loses to holding when price moves far in one direction and stays there, or when there isn't enough fee volume to pay for the risk. Stay out, or keep it to one-sided walls, when:

  • The token is in a strong trend. A two-sided position sells your winners on the way up or buys the loser all the way down, then stops earning.
  • The token is illiquid or untrusted. Thin volume pays little, and a mintable supply, frozen balances or a few wallets holding the float can take the position to zero.
  • Volume / TVL is low. Plenty of liquidity competing for little volume means a small fee share.
  • Costs are larger than fees. A small position on an expensive chain can spend more on claims and repositions than it earns.
  • You may need the asset. A position can end up entirely in the token you didn't want, at the moment you need the other one.

What an LP toolkit needs

A good set of LP tools covers the whole loop above: finding the pool, placing the range, tracking whether it is actually making money, and maintaining it without constant attention. Whatever you use, check it for these:

Pool discovery
Every pool for a token across venues and fee tiers, with TVL, volume and fee-to-TVL, so you choose the pool rather than take the first one listed.
Range visualisation
Your range drawn on the price chart against the depth already in the pool, so you place it against the book rather than guessing.
IL and PnL tracking
Position value and fees shown separately, measured against what went in, so earned fees don't hide a losing range.
Automation
Auto-harvest at a threshold, trailing (follow price) with dwell and move caps, and fill-and-stop for walls, with every automated action recorded and revocable.
Multi-wallet
Acting across several wallets in one flow, if you split positions between accounts.

lp.xyz is built around that list. The hub finds every pool for a token across every supported AMM on a chain, on Solana and EVM chains. The token page draws your range on the chart as a band you can drag, beside a liquidity silhouette of the pool and a pools table with fee-to-TVL. Activity and PnL tracks each position's value against what went in with fees counted separately, a daily calendar, and realised results for closed positions. Automation covers Harvest, Follow price and Withdraw on fill, plus strategies such as trigger walls and limit orders, signed through a delegation you can revoke. Deposits can fan out across several wallets in one flow.

Other tools cover parts of this well. Venue interfaces are often the best place to go deep on one AMM: Meteora's Dynamic Terminal, for example, shows 24h fee / TVL and PnL per position and has a one-click rebalance. When you compare tools, look at which venues they cover, whether you keep custody of your keys, exactly what any automation is allowed to sign, and whether fees are disclosed before you sign.

Next steps

This guide is educational, not financial advice: providing liquidity can lose money, and no range, rebalancing rule or hedge removes that risk.

More guides

Start provisioning.

Connect a wallet and LP across 6 AMMs on Solana and EVM from one place.