Impermanent Loss Calculator

It is the gap between what a liquidity position is worth and what the same two tokens would have been worth if you had simply held them. An automated market maker keeps a fixed share of its value in each token, so every time the market price moves, arbitrage traders rebalance the pool back to that share, selling the token that is rising and buying the one that is falling. Those forced trades are good for the pool's pricing and bad for the depositor: you end up with less of the winner and more of the loser than you started with. In a 50/50 constant-product pool the size of the gap is 2√r / (1 + r) − 1, where r is how far one token's price moved relative to the other.

Because the name describes the one case where it reverses, not the usual case. If the two prices return to exactly the ratio you deposited at, the gap closes completely. That is the impermanence. But the ratio very often does not return, and the loss is realised in full the moment you withdraw, at which point there is nothing impermanent about it. Divergence loss is the accurate name and the one you will see in the research literature. Treat the number on this page as a real cost that happens to be reversible only under a condition you do not control.

In a 50/50 pool, no: only the ratio matters. A token doubling against its pair and a token halving against its pair cost exactly the same 5.72%, because 2√r / (1 + r) is unchanged when you replace r with 1/r. That symmetry is a property of the even split. In a weighted pool it breaks: an 80/20 pool loses 3.27% if the heavy token doubles but 4.28% if it halves, because the pool is far more exposed to the side it holds most of.

Nothing at all: impermanent loss is exactly zero. The pool prices tokens against each other, not against a currency, so if both double, the ratio between them is unchanged, no arbitrage trade is profitable, and the pool never rebalances. You withdraw exactly the balances you deposited and the position is worth precisely what holding would have been worth, plus every fee you collected. This is why impermanent loss depends only on divergence and never on market direction, and it is the single most misunderstood point about it.

Because the two prices are pinned to the same thing, so the ratio barely moves. A USDC/USDT pair trading between 0.998 and 1.002 has a divergence ratio inside half a percent, which works out to well under a thousandth of a percent of loss, invisible next to the fees. That is why stable pools can run with very thin fee tiers and still be worth providing to. The risk that replaces divergence is a depeg: if one side breaks its peg and goes to 0.90, the pool will have bought all the way down and you will be left holding almost entirely the broken token.

Exactly enough to cover the gap the divergence created, and this page solves for it directly. On a 10,000 deposit with a 2× move in the price ratio, the divergence costs about 858, so 90 days in the pool would have needed roughly a 34.8% fee APR to break even. At a 20% fee APR you would have needed around 157 days at that divergence. Both figures are computed rather than estimated, and the fee income is modelled on the value you deposited, which is the basis every quoted pool APR uses.

Because it only has to rebalance a fifth of its value rather than half. The pool keeps 80% of its value in one token, so when that token moves, the trade needed to restore the target weighting is far smaller and far less of the winner gets sold. On a 3× move a 50/50 pool loses 13.40% and an 80/20 loses 7.38%; at 10× it is 42.50% against 23.05%. The trade-off is real, though: a lopsided pool holds less of the second token, sees less trading volume per unit of liquidity, and therefore earns proportionally fewer fees. Lower divergence is bought, not free.

No, and it would be dishonest to pretend otherwise. This page models a full-range, v2-style position, where your liquidity is spread across every price from zero to infinity. Concentrated liquidity narrows that to a band you choose, which multiplies both the fee income and the divergence loss by roughly the same leverage factor. It also adds a behaviour v2 does not have: once the price leaves your range, the position is entirely in one token and stops earning fees altogether until the price comes back. The v2 figures here are a reasonable lower bound on what a wide v3 range does, and they are simply the wrong model for a tight one.

Gas to deposit, to withdraw, and to claim; swap slippage and price impact on the way in and out; the market price of any farm or incentive token paid on top of trading fees, which is worth whatever it is worth on the day you sell it; protocol fee switches that take a cut of the trading fees; and tax, which in many countries treats entering and exiting a pool as disposals. It also cannot model the risks that have actually cost liquidity providers the most money: contract exploits, upgradeable or malicious token contracts, a stablecoin losing its peg, and a token going to zero.

No. There is no price feed, no wallet connection, no address lookup and no account. Every figure comes from a number you typed, the arithmetic runs in your browser, and nothing you enter leaves the page. The example pairs in the panel are illustrative typed prices to save you filling six boxes. They are not quotes, and you should overwrite all of them with the prices from your own position.