Measure divergence loss against the assets you would have held
The term impermanent loss describes a comparison between a liquidity position and holding the originally deposited tokens. The comparison can be measured while the position remains open. Withdrawing crystallizes the inventory then available but does not create the underlying difference.
Set up a valid comparison
Choose a common valuation time and currency. Write down the exact token quantities deposited and treat those as the holding benchmark. If the provider later adds or removes capital, update the benchmark with equivalent external flows at the same times. Keep fees, incentive rewards and transaction expenses visible. Comparing a pool position including fees with a benchmark that uses a different starting mix obscures which part of the result came from the market maker.
A textbook full-range, equal-value, constant-product pool keeps the product of its two reserves constant during a fee-free swap. When external prices move, trading tends to change the reserve ratio toward the new relative price. The pool sells some of the appreciating asset and acquires more of the other relative to holding the original quantities. Trading can occur in either direction. The comparison results from changing relative prices, not just a fall in the market as a whole.
Identify the limits of the model
For a full-range constant-product position with no fees, the end-price ratio determines the theoretical difference versus holding under the usual equilibrium assumption. Real pools can charge fees, and path and volume matter for how much fee income accrues. Concentrated ranges and other market-making designs have different inventory behavior. Applying a textbook full-range formula to a concentrated or stable-swap pool can produce an attractive but incorrect estimate.
If the relative price later returns to its starting level, the modeled fee-free full-range difference can disappear. That outcome is conditional. A project can lose demand permanently, an asset intended to track another can depeg, or a provider may need to withdraw earlier. The word “impermanent” offers no timetable or assurance that the original relationship will return.
Put any fee offset in the same unit
Measure how much the benchmark is ahead in currency terms, then compare actual fees net of costs at the same time. An annualized fee percentage observed during one day cannot close a dollar gap that already exists. If fees were distributed in a volatile reward token, value the units received and disclose whether they have actually been sold.
Worked example
Both investments gain, one trails by $1,000
Start with one token priced at $1,000 and $1,000 of a dollar asset. Ignore fees and assume the latter holds its value. When the token reaches $4,000, a full-range constant-product position with starting reserve product 1,000 rebalances to 0.5 token and $2,000, worth $4,000. Holding the originals is worth $5,000. The pool rose from $2,000, yet trails holding by $1,000, or 20% of the $5,000 benchmark.
What to remember
- Define the holding benchmark using the deposited quantities.
- Use a model that matches the actual pool's design and range.
- Compare realized fee value with the benchmark gap in the same currency.
Sources and further reading
- Uniswap · Version 2 protocol whitepaper
- Uniswap developer documentation · Concentrated liquidity
- Uniswap developer documentation · Fees
Sources checked 21 November 2026 · De Leon Coin