Convert a quoted APR into the APY it really pays (or an APY back into the APR underneath it), then project what a stake earns over a lock period once the validator has taken its commission, and see exactly what restaking is worth against leaving the rewards alone.
A net 3.78% APR compounds to 3.85% APY. Over 1 year, 32 ETH earns 1.23268719 ETH, ending at 33.23268719 ETH. Restaking adds 0.02308719 ETH over leaving rewards to sit.
32 ETH becomes 33.23268719 ETH at a net 3.85% APY, with every reward restaked daily.
On the starting balance, at the net rate
On the starting balance, at the net rate
On the starting balance, at the net rate
Same stake, same net rate, same length. The only difference is whether each reward goes back to work or waits in the wallet.
Restaking is worth 0.02308719 ETH more over 1 year.
That gap is the whole difference between 3.78% APR and 3.85% APY: 0.07% a year of rewards earning rewards. It widens with time and with how often you restake, and it is the reason a longer lock can beat a higher headline rate.
Added back to the stake 365 times over the period
Paid out and held aside, so the staked balance never grows
Restaking is not free everywhere. On chains where you claim and re-delegate manually, each restake costs a transaction fee, and below a certain balance those fees eat the gain shown above. Auto-compounding validators and liquid-staking tokens do it for you, usually in exchange for the commission in the rail.
Read the row for your frequency. One column takes your number as an APR and compounds it; the other takes it as an APY and works out the APR underneath.
| Compounding | Times a year | As an APR, it pays | As an APY, it came from |
|---|---|---|---|
| Per block (12 s) | 2,628,000 | 4.289% APY | 4.114% APR |
| Dailyyours | 365 | 4.289% APY | 4.114% APR |
| Weekly | 52 | 4.288% APY | 4.116% APR |
| Monthly | 12 | 4.282% APY | 4.121% APR |
| Quarterly | 4 | 4.267% APY | 4.135% APR |
| Annually | 1 | 4.200% APY | 4.200% APR |
| Continuous | ∞ | 4.289% APY | 4.114% APR |
Annual compounding is the fixed point: at one restake a year the two columns agree, because there is nothing for the rewards to compound into. Continuous compounding is the ceiling, and the distance between daily and continuous is far smaller than most yield pages imply.
Every balance is calculated from the start of the stake, so a coarse table and a fine one agree to the last decimal. Capped at 240 rows.
| Week | Day | Opening stake | Rewards | Cumulative | Without restaking | Balance |
|---|---|---|---|---|---|---|
| 1 | 7 | 32 | 0.02320502 | 0.02320502 | 0.02319781 | 32.02320502 |
| 2 | 14 | 32.02320502 | 0.02322184 | 0.04642686 | 0.04639562 | 32.04642686 |
| 3 | 21 | 32.04642686 | 0.02323868 | 0.06966554 | 0.06959342 | 32.06966554 |
| 4 | 28 | 32.06966554 | 0.02325554 | 0.09292108 | 0.09279123 | 32.09292108 |
| 5 | 35 | 32.09292108 | 0.0232724 | 0.11619348 | 0.11598904 | 32.11619348 |
| 6 | 42 | 32.11619348 | 0.02328928 | 0.13948275 | 0.13918685 | 32.13948275 |
| 7 | 49 | 32.13948275 | 0.02330616 | 0.16278892 | 0.16238466 | 32.16278892 |
| 8 | 56 | 32.16278892 | 0.02332306 | 0.18611198 | 0.18558247 | 32.18611198 |
| 9 | 63 | 32.18611198 | 0.02333998 | 0.20945196 | 0.20878027 | 32.20945196 |
| 10 | 70 | 32.20945196 | 0.0233569 | 0.23280886 | 0.23197808 | 32.23280886 |
| 11 | 77 | 32.23280886 | 0.02337384 | 0.2561827 | 0.25517589 | 32.2561827 |
| 12 | 84 | 32.2561827 | 0.02339079 | 0.27957349 | 0.2783737 | 32.27957349 |
| 13 | 91 | 32.27957349 | 0.02340775 | 0.30298124 | 0.30157151 | 32.30298124 |
| 14 | 98 | 32.30298124 | 0.02342473 | 0.32640597 | 0.32476932 | 32.32640597 |
| 15 | 105 | 32.32640597 | 0.02344171 | 0.34984768 | 0.34796712 | 32.34984768 |
| 16 | 112 | 32.34984768 | 0.02345871 | 0.37330639 | 0.37116493 | 32.37330639 |
| 17 | 119 | 32.37330639 | 0.02347572 | 0.39678211 | 0.39436274 | 32.39678211 |
| 18 | 126 | 32.39678211 | 0.02349275 | 0.42027486 | 0.41756055 | 32.42027486 |
| 19 | 133 | 32.42027486 | 0.02350978 | 0.44378464 | 0.44075836 | 32.44378464 |
| 20 | 140 | 32.44378464 | 0.02352683 | 0.46731147 | 0.46395616 | 32.46731147 |
| 21 | 147 | 32.46731147 | 0.02354389 | 0.49085536 | 0.48715397 | 32.49085536 |
| 22 | 154 | 32.49085536 | 0.02356096 | 0.51441632 | 0.51035178 | 32.51441632 |
| 23 | 161 | 32.51441632 | 0.02357805 | 0.53799437 | 0.53354959 | 32.53799437 |
| 24 | 168 | 32.53799437 | 0.02359515 | 0.56158952 | 0.5567474 | 32.56158952 |
| 25 | 175 | 32.56158952 | 0.02361226 | 0.58520178 | 0.57994521 | 32.58520178 |
| 26 | 182 | 32.58520178 | 0.02362938 | 0.60883116 | 0.60314301 | 32.60883116 |
| 27 | 189 | 32.60883116 | 0.02364651 | 0.63247767 | 0.62634082 | 32.63247767 |
| 28 | 196 | 32.63247767 | 0.02366366 | 0.65614133 | 0.64953863 | 32.65614133 |
| 29 | 203 | 32.65614133 | 0.02368082 | 0.67982216 | 0.67273644 | 32.67982216 |
| 30 | 210 | 32.67982216 | 0.02369799 | 0.70352015 | 0.69593425 | 32.70352015 |
| 31 | 217 | 32.70352015 | 0.02371518 | 0.72723533 | 0.71913205 | 32.72723533 |
| 32 | 224 | 32.72723533 | 0.02373238 | 0.75096771 | 0.74232986 | 32.75096771 |
| 33 | 231 | 32.75096771 | 0.02374959 | 0.77471729 | 0.76552767 | 32.77471729 |
| 34 | 238 | 32.77471729 | 0.02376681 | 0.7984841 | 0.78872548 | 32.7984841 |
| 35 | 245 | 32.7984841 | 0.02378404 | 0.82226814 | 0.81192329 | 32.82226814 |
| 36 | 252 | 32.82226814 | 0.02380129 | 0.84606943 | 0.8351211 | 32.84606943 |
| 37 | 259 | 32.84606943 | 0.02381855 | 0.86988798 | 0.8583189 | 32.86988798 |
| 38 | 266 | 32.86988798 | 0.02383582 | 0.8937238 | 0.88151671 | 32.8937238 |
| 39 | 273 | 32.8937238 | 0.02385311 | 0.91757691 | 0.90471452 | 32.91757691 |
| 40 | 280 | 32.91757691 | 0.0238704 | 0.94144731 | 0.92791233 | 32.94144731 |
| 41 | 287 | 32.94144731 | 0.02388771 | 0.96533503 | 0.95111014 | 32.96533503 |
| 42 | 294 | 32.96533503 | 0.02390504 | 0.98924006 | 0.97430795 | 32.98924006 |
| 43 | 301 | 32.98924006 | 0.02392237 | 1.01316243 | 0.99750575 | 33.01316243 |
| 44 | 308 | 33.01316243 | 0.02393972 | 1.03710215 | 1.02070356 | 33.03710215 |
| 45 | 315 | 33.03710215 | 0.02395708 | 1.06105923 | 1.04390137 | 33.06105923 |
| 46 | 322 | 33.06105923 | 0.02397445 | 1.08503368 | 1.06709918 | 33.08503368 |
| 47 | 329 | 33.08503368 | 0.02399184 | 1.10902552 | 1.09029699 | 33.10902552 |
| 48 | 336 | 33.10902552 | 0.02400923 | 1.13303475 | 1.11349479 | 33.13303475 |
| 49 | 343 | 33.13303475 | 0.02402664 | 1.1570614 | 1.1366926 | 33.1570614 |
| 50 | 350 | 33.1570614 | 0.02404407 | 1.18110546 | 1.15989041 | 33.18110546 |
| 51 | 357 | 33.18110546 | 0.0240615 | 1.20516697 | 1.18308822 | 33.20516697 |
| 52 | 364 | 33.20516697 | 0.02407895 | 1.22924592 | 1.20628603 | 33.22924592 |
| 53 | 365 | 33.22924592 | 0.00344128 | 1.23268719 | 1.2096 | 33.23268719 |
APR is what the protocol pays per year with nothing reinvested. APY is what you end up with once each reward starts earning too. Comparing one against the other is how two identical validators appear to pay different amounts.
If a dashboard already shows an APY, the restaking is in the number. Feeding it into a compounding projection counts the same effect twice and overstates a long lock badly. Switch the toggle to APY and the APR is backed out first.
A 12% yield paid in a token that falls 30% is a loss in every currency you spend. The reward column here is denominated in the asset on purpose, and any money figure is a price you typed and held flat.
This is arithmetic on the numbers you type, not investment, financial, or tax advice. It assumes the reward rate holds for the whole period, which no chain guarantees: issuance schedules, participation rates, and validator commissions all move. It does not model slashing, downtime penalties, unbonding periods, gas paid to claim or re-delegate, protocol changes, or tax on rewards, which many countries treat as income at the moment they are received. Staked assets can lose value faster than any yield replaces.
Once the stake is unbonded and sold, this prices the exit: fees on both sides, realized profit, and the break-even.
Work out the blended cost of the coins you are staking, so the yield is measured against what you actually paid.
The same compounding maths on money instead of coins, with recurring contributions, inflation, and a savings target.
Staking maths is two questions wearing one coat. The first is a conversion (what a rate becomes once its own rewards start earning), and it has an exact answer in both directions. The second is a projection, which is only ever as good as the assumption that the rate holds. This page keeps them apart: the conversion table is arithmetic you can rely on, the reward schedule is clearly a model, and the rate you type is read as an APR or an APY exactly as you label it.
Amount staked, then the rate, and say whether it is an APR or an APY. That toggle matters more than any other control on the page, because an APY already contains its own compounding.
Per block or epoch, daily, weekly, monthly, quarterly, annually, or the continuous limit. Take the validator commission off in the same panel and watch the gross rate become the net one.
Days, months, or years produces the rewards, the ending balance, the effective APY, and a row-by-row table with the running balance next to what the same stake would earn if nothing were ever restaked.
APY = (1 + APR/n)^n − 1 and its exact inverse, at every frequency from one restake a year to the continuous limit. The table converts your number in both directions at once, so a validator quoting APR and one quoting APY can finally be compared.
Flip the toggle to APY and the rate is converted back to the APR that produced it before a single projection runs. Skipping that step is the most common error in yield maths, and it inflates a long lock by more than the commission ever will.
The validator cut comes off the reward rate, and the gross and net APR are shown side by side with the annual cost of the difference in the asset itself, the number that decides whether a cheaper validator is worth moving to.
The honest headline: the compounded run and the leave-it-alone run, same stake, same net rate, same length. Where the stake ends before a single restake fires, the page says the two are identical instead of inventing a gain.
Every row is derived from the closed form, so a yearly table and a per-restake table of the same stake agree to the last decimal and no rounding drifts down the column. Rewards inside an unfinished period accrue in a straight line, exactly as they do before a restake boundary.
Every number is one you typed, including the optional coin price used for the money column, which is labelled as your assumption throughout. Nothing is fetched, nothing is uploaded, and the page works for a rate that was current two years ago.
APR is the plain annual reward rate with nothing reinvested: stake 100 coins at 5% APR and a year later you have earned 5 coins, whether they were paid daily or all at once. APY is what you end up with when each reward is added back to the stake and starts earning too. The formula is APY = (1 + APR/n)^n − 1, where n is how many times a year rewards are restaked. At 5% APR restaked daily the APY is 5.127%; restaked annually the two are the same number, because there is nothing for the reward to compound into.
Because the compounding is already inside it. If a dashboard advertises 8% APY and you run that through a compounding projection, you have applied the effect twice, and over a five-year lock the error is larger than most commissions. This page handles it with a toggle: mark the rate as an APY and it is converted back to the APR that produces it at your chosen frequency before anything else happens, then reported both ways so you can see the pair.
It is a cut of the rewards, not of the stake. A 10% commission on a 4% gross APR leaves you a 3.6% net APR, and the panel shows both figures plus the cost of the gap in coins per year, which is the form that makes it comparable to the transaction fees of moving to a different validator. On some chains the operator also charges a fixed amount per epoch before the percentage, so treat a very small stake with more suspicion than the percentage alone suggests.
Less than yield marketing implies. The distance from annual to daily compounding is real but modest, and the distance from daily to continuous (the mathematical ceiling) is tiny: at 5% it is under a hundredth of a percentage point. The frequency table on the page shows the whole curve flattening. Where frequency does matter is when restaking costs a transaction fee: past a certain point you pay more in gas than the extra compounding returns, which is the argument for auto-compounding validators and liquid-staking tokens.
Only as the length you enter, and that is a genuine limitation worth stating plainly. Most networks make you wait to get your coins back (commonly somewhere between a couple of days and a month), during which the stake usually stops earning, cannot be sold, and remains fully exposed to the price. The projection assumes you stay staked for the period you typed and can leave at the end of it. If your chain has an unbonding queue, add it to the length you are committing and treat the final stretch as earning nothing.
Easily, and it is the most important thing on this page. Rewards are paid in the asset you staked, so the yield is a coin count, not a currency return. A 12% yield on a token that falls 30% is a loss in every currency you actually spend. That is why the reward figures here are denominated in the asset and any money column is labelled as a price you assumed and held flat. Against that, the risk is not only price: slashing can burn part of the stake outright for validator downtime or double-signing, and the funds are illiquid while you would most want to sell them.
Partly from transaction fees and partly from new issuance. The issuance part is a transfer, not a gain. When a chain mints new coins to pay stakers, every holder who is not staking is diluted, and a staker earning the same rate as the inflation rate is holding their share of the supply steady rather than growing it. This is why a real-yield figure subtracts the inflation rate from the nominal one. This calculator projects the nominal reward you receive; if you want the dilution-adjusted view, enter the reward rate minus the network inflation rate as your APR.
Up to 100 years of stake length and a reward rate up to 10,000%, both clamped with a visible notice rather than silently accepted, and a schedule capped at 240 rows. The row length picks itself from the length of the stake (restake, day, week, month, quarter, or year), and you can override it. Hitting the cap trims only the table: the rewards, balances, and APY above it always cover the entire period.
No, and it never will. There is no price feed and no chain connection, so both the reward rate and the optional coin price are yours to type. The chain presets in the panel are illustrative typed defaults from mid-2026 to save you filling four boxes; they are not live network rates, and every one of them moves with participation and validator policy. Check your validator dashboard and overwrite them.
Explore the growing collection for calculations, documents, writing, and everyday work.