The number that gets quoted at me the most is staking APY, and it is close to useless for the thing people actually want to know. Someone tells me a token pays 12 percent to stake it and treats that as free money. But if the total supply is also growing 30 percent a year from emissions and unlocks, staking is not paying you anything real. It is the toll you pay to fall behind slightly slower than everyone who did not stake. The APY is a share of the dilution, not an escape from it.
So the question I care about is dead simple to state and annoying to answer. How fast is the total float actually growing over the next year, once you add up everything that mints new tokens and subtract everything that destroys them. Get that one number and you can hold it up against how fast demand for the token is plausibly growing, and most of the tokenomics argument resolves itself. Everything else is decoration.
What actually adds to the float
The float is not the fully diluted supply and it is not the marketcap-quoted circulating supply either, both of which lie to you in different directions. What you want is the number of tokens that can hit the market over your window, so you have to hunt down every source of new units. In practice there are four buckets, and people almost always forget at least two of them.
- Protocol emissions. The block reward, staking reward, or inflation schedule written into the chain. This is the one everybody knows about and it is usually the one written down most clearly.
- Scheduled unlocks. Tokens already minted but locked in vesting contracts for the team, investors, foundation, and ecosystem funds. These do not show up as new mints on the explorer because they were minted at genesis, so an emissions-only view misses them completely. They are the biggest source of surprise dilution I see.
- Incentive and liquidity programs. Grants, liquidity mining, points-to-token conversions, airdrop seasons paid out of a treasury. Discretionary, easy to miss, and often larger than the base emission rate.
- Burns, going the other way. Fee burns, buyback-and-burn, slashing. These subtract. On most tokens they are small relative to issuance, but on a few they genuinely matter, so you net them out rather than assume they cancel anything.
Building the number
Here is the workflow I use, and it takes maybe an afternoon per token the first time. You are trying to produce annualized net new tokens divided by current float. Start with the current circulating float as your denominator, because dilution is always relative to what is out there now.
- Pull the annual protocol emission from the chain's own docs or the inflation parameter. If it is a percentage of supply, convert it to a token count so everything is in the same unit.
- Open the vesting schedule and sum only the tokens that unlock inside your next twelve months. Ignore the cliffs that land in year three, they are not your problem yet. This is the step everyone skips and it is usually where the ugly surprises live.
- Estimate the incentive spend. Treasuries publish some of this and hide most of it, so I take the trailing rate of what they have been paying out and assume it roughly continues unless there is a stated end date.
- Subtract the trailing burn rate. If burns depend on volume or fees, use a conservative recent average and do not give the token credit for a burn spike that will not repeat.
- Add the first four, subtract the fifth, divide by current float. That is your annualized net issuance.
The output is one percentage. If protocol emissions plus unlocks plus incentives come to roughly 40 percent of current float over the year, and burns claw back a couple of points, you are looking at something like 38 percent real inflation. The 12 percent staking APY does not offset that. It softens it, and only for the fraction of holders who bother to stake.
Where the data actually lives
None of this requires anything you cannot find. The block explorer gives you total supply, mint events, and burn addresses, so you can sanity-check the emission rate against what actually got minted over the trailing period rather than trusting the docs. The vesting contracts are usually readable on-chain or published in a token-unlock tracker, and I always cross-check the tracker against the actual contract because trackers go stale when a project quietly amends a schedule. Incentive programs live in governance forums and treasury dashboards, which is the least fun part but also where the discretionary spend hides.
The failure mode I want you to avoid is the emissions-only estimate. Someone reads the inflation parameter, sees 5 percent, and calls the token low-inflation. Then two quarters later a team-and-investor cliff unlocks 15 percent of supply in a single month and the price behaves exactly as you would expect. The emissions number was honest and also nearly irrelevant, because for a young token the vesting cliffs dwarf the block reward for the first few years. If you only measure the thing that is easy to measure, you will keep getting blindsided by the thing that is easy to ignore.
What the number is for
Once you have a net issuance figure, you hold it against realistic demand growth, and be honest about realistic. New buyers, tokens getting locked into staking or collateral and taken out of circulation, actual protocol usage that requires holding the token. If demand is plausibly growing faster than net issuance, dilution is being absorbed and price can hold or rise on flat sentiment. If issuance is running well ahead of any demand story you can tell with a straight face, you are relying on new speculative money arriving faster than tokens unlock, forever, which is a bet not a thesis.
I do not treat the final percentage as precise. The incentive estimate alone can be off by a lot, and I would rather be roughly right than confidently wrong. So I usually build a low case and a high case, mostly by flexing the discretionary incentive spend, and I look at the range. A token where even the optimistic case shows issuance outrunning demand is telling you something the APY was designed to hide. That is the whole point of doing the arithmetic. The float grows whether you measure it or not, and the only version of the number that helps you is the one that adds up every source instead of the one that is easiest to quote.