BTCJP ResearchResearch &
archive edition

DeFi accounting

An APY Dashboard Needs a Denominator and a Clock

Reconstruct a displayed rate from its observation window, reward unit, reinvestment assumption and withdrawal conditions before comparing percentages.

1,166 words6 min readBTCJP Research Editorial Desk

New editorial guide, published 16 October 2026. The historical sources linked below are preserved separately and are not represented as current product or service information.

BTCJP's archive includes a directory of yield dashboards and several research notes comparing rewards, transaction costs and locked tokens. The old percentages preserve what the author reported at the time, not what a participant could receive today. Reading them usefully requires reconstructing the measurement behind each figure. This new guide uses hypothetical arithmetic to separate an observed reward from an annualised rate and a claimable token from an unrestricted asset. It does not forecast returns or recommend a position. The question is narrower and reproducible: what inputs and assumptions would produce the number displayed on a dashboard?

Write the numerator and denominator in their own units

A rate compares quantities. Before interpreting it, identify what is counted as the reward and what is counted as the base. A hypothetical dashboard might compare a reward token's marked value with the marked value of a deposited position. Another might compare units of one asset with units of that same asset. These are different measurements. If the numerator and denominator use different assets, a valuation assumption connects them. That assumption needs an observation time, because a change in the quoted price can change the displayed rate without changing the reward quantity earned.

Keep a simple worksheet with the reward units, deposit units, valuation basis and observation window. Do not start by copying the final percentage. Suppose a hypothetical position of one thousand accounting units receives four units during a week, measured on a consistent basis. The observed ratio is 0.4 percent for that week. This calculation says nothing yet about the following week or an entire year. It identifies one measured period. The useful first result is that the displayed percentage can be traced back to quantities and a clock rather than accepted as an unexplained property of the product.

Annualisation is a model, not another observed year

Using the hypothetical 0.4-percent week, multiplying by fifty-two produces a simple annualised figure of 20.8 percent. Assuming the same weekly rate and full weekly reinvestment instead gives approximately 23.07 percent from 1.004 raised to the fifty-second power, minus one. Neither number is a measured annual result. Each projects a particular assumption beyond the observation window. The calculation should therefore state the period count and reinvestment rule rather than present the larger result as something that has already happened to a participant's position.

The CFPB's deposit-yield calculation rules are a useful primary reference for understanding why compounding conventions matter, but a crypto dashboard is not automatically governed by or equivalent to that deposit framework. A research note should describe the dashboard's actual convention. Some rates may use a short recent window, a schedule of incentives or a different compounding period. The point is to make the convention explicit, not to infer a universal definition from the label alone. Two annual-looking percentages can disagree because their assumptions differ even when their source observations are both accurately recorded.

Compounding requires a practical reinvestment path

A compound-rate model assumes more than arithmetic. The reward must become available in a form that can be added to the relevant base under the assumed schedule. If the reward uses a different token, reinvestment may require a conversion. If a claim or deposit has a cost, the model needs an account of that cost. If the reward is locked, reinvesting it immediately may not be possible. A dashboard can display a mathematically coherent compound number while the participant's actual route differs from the simplified model that produced it.

For the hypothetical weekly example, write down what would have to happen every week for the compound assumption to hold. Identify the claim, any conversion, the reinvestment action and the resulting base. Then state which steps the dashboard includes and which it omits. This is not a recommendation to perform the sequence. It is a way to evaluate the meaning of the displayed calculation. A reader can then distinguish automatic compounding from a model that assumes a person repeatedly completes several separate actions at no additional cost or delay.

A credited reward can still be restricted

The historical BTCJP staking notes discuss rewards subject to an escrow period. That makes availability a separate question from the quantity credited. A participant can see a reward record while lacking the ability to transfer or redeem it immediately. Vesting implementations illustrate why schedule, allocation and release are distinct concepts; OpenZeppelin's finance documentation describes contract components addressing such arrangements. The actual rules of a particular position must be inspected independently. An old note's lock duration should not be imported as a current parameter for a different version or protocol.

Add availability fields to the worksheet. Record what is credited, what is claimable now, what remains restricted and which condition changes that status. If voting power or another benefit depends on a lock, keep that benefit separate from unrestricted units of the underlying asset. This makes comparisons more intelligible. Two positions with similar displayed rates may offer very different timing and control over the resulting assets. A single headline percentage cannot express those differences unless the reader reconstructs the schedule and the associated rights from the underlying rules.

Gross rewards are not the complete position outcome

A reward rate can describe one source of inflow without describing the value of the whole position. Transaction costs, conversion costs, changing asset prices and changes in the underlying claim can all matter to a separate outcome calculation. The historical BTCJP directory itself warns that a high reward does not eliminate contract or pool-related loss. Preserve that context rather than detach a large percentage from the rest of the source. A measurement note should identify what the rate includes before deciding which additional observations are necessary for a broader comparison.

In a hypothetical accounting exercise, list rewards and costs separately instead of subtracting numbers with incompatible units. Convert only under an explicit valuation basis and retain the unconverted quantities too. Then define the comparison: the marked value at a later time, the number of units withdrawable or a baseline holding scenario. Each answers a different question. This prevents a positive reward entry from being treated as proof that the whole position increased in value or outperformed an alternative whose quantities and valuation basis were never recorded.

Save the assumptions with the screenshot

A useful dashboard record includes the observation time, measurement window, reward asset, base, valuation source, annualisation convention, fees included and restrictions on availability. The screenshot can remain part of the evidence, but these fields make its number interpretable after the interface changes. If the dashboard does not disclose an input, record it as unknown rather than solve backward for a reassuring assumption. Repeated observations can then show whether the rate changed because of rewards, prices, deposits or the model used to summarise them.

The archived BTCJP notes provide a practical starting point for that discipline: they place rates beside fees, rewards and locks, even when their conclusions are informal. A new explanation can make the arithmetic and boundaries explicit. Begin with quantities and a period, distinguish observation from annualisation and test whether compounding is operationally possible. Keep restricted rewards and whole-position outcomes separate. An APY dashboard becomes useful evidence when its denominator, clock and assumptions are visible; without them, a precise-looking percentage can conceal several different measurements behind the same small label.

Technical references

From the recovered archive

Keep reading.