A yield percentage is incomplete until you know what it measures. A headline might describe a simple annualized rate, a compounded scenario, a recent fee observation, or rewards valued at a particular token price. These are different quantities. A DeFi meter should expose the assumptions rather than presenting the largest number as the most attractive result.
This guide explains APR, APY, and cost accounting through fictional examples. None of the rates represents a current product or a promised return. The calculations deliberately hold several variables constant so that the arithmetic is easy to follow; real outcomes can differ when prices, rates, liquidity, and operational conditions change.
Distinguish APR from APY
In a simple educational calculation, an annual percentage rate describes a yearly rate without assuming that earnings are reinvested during the year. An annual percentage yield includes a specified compounding assumption. Always inspect how the provider uses these labels; the letters alone do not reveal all the inputs.
Suppose an illustrative rate is 12% per year and the principal is $1,000. Under simple annual accrual, with the rate and principal value held constant, the annual earnings are $120. If the same nominal rate compounds monthly without costs, the calculation is 1.01 raised to the twelfth power, minus one, giving about 12.68%.
The difference comes from reinvesting earlier earnings under the model. It does not appear merely because an interface displays APY. The reinvestment schedule must be feasible, the reinvested amount must actually earn the assumed rate, and any costs must be included in the scenario.
Identify the source of the payment
Write a plain-language explanation of why the yield exists. For a hypothetical position, it might be borrowing charges, exchange fees, incentive distributions, or a combination. These sources have different conditions. An incentive payment should not be silently relabeled as a fee generated by ongoing user demand.
The Uniswap documentation on fees explains that fee collection and distribution differ by version and that active liquidity matters for fee accrual in concentrated-liquidity positions. This illustrates why a general pool-level fee number is not automatically the realized return of every position in that pool.
For your research record, identify the payment asset, the observation period, the position conditions, and the allocation rule. Keep a separate field for whether the displayed rate is historical, annualized from a shorter window, or an explicitly modeled scenario.
Convert a rate into a holding-period scenario
Thirty days, not a year
Assume a fictional $1,000 principal earns a constant simple annual rate of 12% for thirty days, using a 365-day convention. The modeled gross earnings are $1,000 multiplied by 0.12 multiplied by 30 divided by 365, or approximately $9.86. This calculation does not imply that a real product would maintain that rate.
Now assume $3 of entry costs, $3 of exit costs, and $2 of claim costs. The total modeled cost is $8, leaving about $1.86 before any other effects. The headline annual rate did not change, but the short-period result is very different from simply applying 12% to the principal.
Keep the convention explicit. A rate displayed using another time basis or an already-compounded method needs a corresponding calculation. Do not mix an APY input with a simple APR formula merely because both are percentages.
Calculate an illustrative break-even point
Under the same simplified constant-rate model, daily gross accrual is about $0.329 on $1,000. Dividing the assumed $8 cost by that daily amount gives roughly 24.3 days to offset those costs. This is an arithmetic threshold within the model, not a recommendation for how long to hold a position.
The threshold changes when the principal or cost changes. With a hypothetical $100 principal and the same fixed $8 cost, the modeled daily accrual would be about $0.0329, requiring roughly 243.3 days. The comparison shows why a fixed cost has a different percentage effect on different position sizes.
Do not confuse this cost break-even calculation with overall investment break-even. It excludes principal price changes, reward-token repricing, position losses, unexpected operational events, and any other omitted components. Label the exclusions directly beside the result.
Keep reward units separate from their valuation
Suppose a fictional position receives ten reward tokens. At an assumed price of $2 each, they are valued at $20. At an assumed price of $1.20, the same ten tokens are valued at $12. The quantity did not change, but the reported dollar reward fell by 40%.
This example shows why a reward-rate estimate should preserve both the number of tokens and the valuation time. A projected dollar yield based on today's assumed token price includes a price assumption, even when the display does not call attention to it.
Distinguish claimable rewards, claimed rewards, and converted proceeds in your own records. These are different stages of a hypothetical cash-flow process. Do not describe a displayed reward balance as realized cash unless a completed conversion actually occurred and its costs are known.
Separate yield from total performance
A position can generate payments while the value of its principal declines. Consider an illustrative starting value of $1,000, an ending principal value of $930, and $40 of received rewards, with no additional deposits or withdrawals. Before costs, the combined ending value is $970, which is $30 below the starting amount.
The $40 reward is real within the example, but calling the whole experience a 4% gain would omit the principal movement. A complete statement separates rewards from changes in the underlying position and then explains how they combine under the chosen accounting method.
A benchmark can answer another question: how the position compares with a specified alternative under matching assumptions. Define the benchmark before examining results. A comparison to simply holding the starting assets is different from a comparison to an unchanged dollar balance, and neither automatically captures every relevant risk.
Use scenarios instead of a single forecast
A useful research worksheet can hold three explicitly hypothetical cases: the assumed rate persists, the rate falls, or rewards have a lower conversion value. Calculate each using the same principal, observation period, and cost definitions. These cases explore sensitivity; they are not assigned probabilities or predictions.
Change one variable at a time before combining changes. This helps you identify which assumption drives the result. If small changes in the reward valuation or exit cost reverse the simplified outcome, the headline yield alone is a poor summary of the scenario.
Keep unavailable facts marked as unavailable. The DeFiMeter methodology separates measured data from assumptions so that an attractive output does not conceal an invented input.
Ask the questions the rate cannot answer
Yield arithmetic does not establish contract reliability, asset backing, withdrawal eligibility, or execution availability. Those questions belong in a separate dependency review. Do not let a precisely calculated percentage imply that every component of the product has been assessed.
Use the Web3 risk checklist for those questions, and the crypto DeFi Meter for wider market context. Keeping these records separate makes the final summary more transparent.
Is more frequent compounding always better?
Only within a model that accounts for the actual conditions. More frequent reinvestment increases the mathematical effect of compounding when the rate is positive and constant and there are no additional costs. Once each reinvestment has a cost, the result depends on the size of the accrued amount and that cost.
For an illustrative $1 of accrued earnings, spending $2 to reinvest it would reduce the position's immediate net value under those assumptions. That example does not establish a universal schedule. It shows why compounding frequency should be an explicit scenario input rather than an attractive label.
Conclusion: account for the whole scenario
APR and APY are useful only when their rate, period, compounding, reward valuation, and cost assumptions are visible. Translate the headline into a holding-period calculation, keep principal changes separate from yield, and show what the model excludes. A careful scenario does not promise a return; it shows which assumptions would have to hold for the arithmetic to describe one.
