Cover Value Explained: Are Your LP Fees Actually Winning?

Ask an LP how a position is doing and you usually get an APR. APR cannot answer the question that actually matters, because it only describes one side of the ledger. You can earn a perfectly real 40% APR in fees and still finish behind someone who did nothing at all.

Cover Value is one number that puts both sides on the same line: how many times over have my fees covered the loss I took by providing liquidity instead of just holding? Below 1, the position is not paying for itself. Above 1, it is.

The formula

Cover Value is a ratio of two dollar figures:

Both halves are dollars, so the ratio is unitless and comparable across pairs, chains and position sizes. A 3.2× on a stablecoin pair and a 3.2× on a volatile pair mean the same thing: fees covered the divergence a little over three times over.

Why the basis is token amounts, not entry prices

This is the detail most home-made spreadsheets get wrong, and it changes the answer completely.

V_hold is not “what I paid”. It is the quantity of each token you deposited, marked at current prices. You are not measuring profit against your cost basis — you are measuring the position against the alternative of having simply held those same coins.

The consequence is clean: price appreciation cancels. If the pair doubles, both sides of the subtraction rise together, and what is left is purely the cost of the rebalancing the AMM did on your behalf. That is exactly what impermanent loss is, and it is why a position can be up in dollars and still show a real IL figure. Those are two different questions and Cover Value only answers the second one.

The math, worked end to end

For a classic 50/50 position, the divergence has a closed form. With r as the ratio of the price change between the two assets:

Put numbers through it. You deposit 1 ETH + 2,000 USDC with ETH at $2,000 — a $4,000 position. ETH then doubles to $4,000, so r = 2.

• V_hold = 1 ETH × $4,000 + 2,000 USDC = $6,000
• |IL| fraction = 1 − 2√2 ÷ 3 = 5.72%
• V_LP = $6,000 × (1 − 0.0572) = $5,657
• |IL_r| = $6,000 − $5,657 = $343

Now the fees decide the verdict. Earn $500 and CV = 500 ÷ 343 = 1.46× — the position beat holding. Earn $200 and CV = 0.58, which reads as −1.72×: the loss was 1.72 times the fees. Same pool, same price move, opposite conclusion — and an APR would have looked healthy in both cases.

Note how fast this scales. At r = 4 the closed form gives exactly 20% divergence. Concentrated ranges amplify it further, because a narrow band converts entirely to the losing side sooner.

How to read the number

• 4.30× — fees covered the loss 4.3 times over.
• −1.80× — underwater: the loss is 1.8 times the fees. Losses are shown inverted and signed on purpose. Every losing position would otherwise compress into the range 0 to 1, where mildly behind and catastrophically behind both round to “0”. Inverting gives losses the same resolution as gains and answers the question you actually have: by how much?
• “no IL” — fees earned with no divergence to cover. The best case. It is never printed as ∞, because a ratio with a zero denominator is undefined, not infinite.
• “pending” — the original deposit is not known yet. Not zero. A missing basis must never be shown as a confident number.

Why a portfolio total is not an average of ratios

Across several positions, Cover Value is ΣF ÷ Σ|IL_r| — total fees over total loss. It is not the mean of the individual ratios, and the difference is not academic.

Suppose one position earned $200 against $100 of loss (2.00×) and another earned nothing against $100 of loss (0.00×). The average of the ratios is 1.00×, which sounds like breaking even. The truth is $200 of fees against $200 of loss — which is also 1.00×, but only by coincidence. Change the sizes and the two answers diverge hard: a tiny position with a spectacular ratio will drag a mean upward while contributing almost nothing in dollars. Summing first weights every position by the money actually at stake.

Three things that make Cover Value lie

The formula is simple. Getting honest inputs is the hard part, and this is where most LP dashboards quietly mislead.

1. A rebalance destroys your entry basis

An actively managed position is not one NFT. Every rebalance burns the old position and mints a new one. Read the entry tick off the current NFT and you are measuring from the last rebalance — often hours ago — not from your original deposit. The position reports close to 0% impermanent loss no matter how far it has actually drifted.

The fix is not a better formula, it is better data: the true original deposit has to be reconstructed by walking the position’s lineage backwards through every re-mint until you reach the one that was funded rather than rolled. Any tool that shows you an IL figure without doing this is measuring the wrong interval.

2. Reinvested fees hide inside the current balance

If your fees are compounded back into the position, they are sitting inside V_LP. Compare that inflated balance straight against V_hold and the earnings quietly net against the loss. A position can then report no impermanent loss at all while having given up a large share of its principal — the fees are papering over the hole they were supposed to be measured against.

Reinvested fees have to be stripped back out before the comparison, so that IL is measured on principal alone and the fees stay on their own side of the ratio where they can be judged.

3. Missing is not zero

If the entry basis cannot be established, the honest output is pending. Treating an absent value as zero turns “we do not know” into “you have no loss”, which is the most expensive kind of wrong a dashboard can be.

Using it to manage positions

Cover Value is most useful as a trend and as a comparison, not as a single snapshot.

• Below 1× and falling — the range is not paying for the divergence it is taking. Usually the pair is trending rather than oscillating, which is the regime concentrated liquidity is worst at. Widen, or step aside until it ranges again.
• Around 1× to 2× — marginal. You are being paid, but not much, for real risk. Check whether gas and rebalance friction have already eaten the margin.
• Above 3× and stable — the range is earning its keep. This is the regime where tightening is worth considering, because you are being well compensated for the divergence you are already absorbing.
• “no IL” with meaningful fees — price has come back to where you started while you collected. Nothing to fix.
• Rising CV on a flat price — fees accruing against a fixed loss. The position is healing; time is on your side.
• A sharp CV drop with no fee change — the denominator moved. Price left your band. Check range position before you touch anything else.

Across a book, rank by Cover Value and the allocation question answers itself: capital sitting in persistently sub-1× positions is subsidising the ones above 3×.

What Cover Value does not tell you

• Friction. Gas, swap tolls at each re-range and any performance fee sit outside the ratio. A 1.2× position that re-ranges daily may be a losing one.
• Token risk. A pair can post a beautiful Cover Value while one side is heading to zero. The ratio measures divergence, not solvency.
• The future. It is a measurement of what has happened, not a forecast. Fees earned are real; the next month is not.
• Taxes. Nothing here accounts for them, and the treatment of LP fees varies by jurisdiction.

Bottom line

APR tells you what you earned. Cover Value tells you whether it was enough. One ratio, both sides of the ledger, comparable across every position you hold — measured against the original deposit, with reinvested fees kept honestly on their own side, and with “we do not know yet” said out loud rather than rounded to zero.

That last part is most of the work. The arithmetic is a division; the difficulty is earning the right to do it.