# Why the reward halves every two years, and what that means if you are deciding today > Canonical: https://blazephoenix.xyz/learn/the-halving-schedule > License: CC BY 4.0 (attribution + link) · © 2026 BlazePhoenix > Updated: 2026-08-13 The staking budget is 180,000,000 tokens — 18% of supply — spread over eight two-year periods, each paying half its predecessor: 90,000,000 in period one down to 703,125 in period eight, 179,296,875 paid in total, with a 703,125 remainder the curve never reaches (exactly one more period's worth). Half the whole budget is paid inside the first two years — the dilution arithmetic that matters more than the halving itself. Staking rewards here are not a rate. They are a budget: a fixed number of tokens, divided into periods, with each period paying exactly half of the one before it. Nothing about that can be changed by a vote, a market condition, or a decision made later. This article prints the whole schedule, checks that it adds up, and then does the arithmetic that actually decides what you receive — which is not the halving. ## The budget and the curve The total supply is 1,000,000,000 tokens, of which 180,000,000 — 18% — is set aside for staking rewards. That budget is spread across eight periods of two years each, which is sixteen years, with each period paying half its predecessor: ``` period years emission cumulative share of budget 1 0 - 2 90,000,000 90,000,000 50.0000% 2 2 - 4 45,000,000 135,000,000 75.0000% 3 4 - 6 22,500,000 157,500,000 87.5000% 4 6 - 8 11,250,000 168,750,000 93.7500% 5 8 - 10 5,625,000 174,375,000 96.8750% 6 10 - 12 2,812,500 177,187,500 98.4375% 7 12 - 14 1,406,250 178,593,750 99.2188% 8 14 - 16 703,125 179,296,875 99.6094% ``` Check the halvings by hand down the column: 90,000,000 / 2 = 45,000,000, then 22,500,000, then 11,250,000, then 5,625,000, then 2,812,500, then 1,406,250, then 703,125. Eight entries, each exactly half the last. Then check the total: adding all eight gives 179,296,875, not 180,000,000. The shortfall is 180,000,000 - 179,296,875 = 703,125 tokens, which the halving curve never pays out. ## Why the leftover equals the last period exactly That the remainder is 703,125 — precisely the amount of the eighth period — is not a coincidence, and seeing why is worth a minute because it tells you the shape of the whole curve. After the first period, 90,000,000 has been paid and 90,000,000 remains, which is what period two would pay. After period two, 45,000,000 remains, which is what period three pays. At every step, what is left is exactly the size of the next payment. A halving schedule always has this property: the remaining tail equals the next instalment. So after eight periods, what remains is what a ninth period would have paid: 703,125. The schedule stops there, and that tail is recoverable rather than emitted. It is worth stating plainly rather than rounding away, because the difference between "180,000,000 will be distributed" and "179,296,875 will be distributed" is 703,125 tokens, and if a document tells you the first, it has not checked. ## How steep this actually is Compare the ends of the curve. The first period pays 90,000,000 and the eighth pays 703,125, and 90,000,000 / 703,125 = 128, which is 2 raised to the power of 7 — seven halvings between the first period and the eighth. Convert to a daily figure so the periods are comparable. Taking a two-year period as 730 days and ignoring leap days: ``` period 1: 90,000,000 / 730 = 123,287.67 tokens per day period 2: 45,000,000 / 730 = 61,643.84 tokens per day period 3: 22,500,000 / 730 = 30,821.92 tokens per day period 8: 703,125 / 730 = 963.18 tokens per day ``` And look at the cumulative column again: half the entire budget is paid inside the first two years, three quarters inside the first four, and 93.75% inside the first eight. Whatever this programme does, it does most of it early. ## The arithmetic that decides what you receive Here is where most explanations of halvings stop, and where the useful part begins. The schedule sets the numerator. Your reward is a share of it, and the share has a denominator that the schedule says nothing about. Suppose you stake 100,000 tokens of weight while the total staked weight is 10,000,000. Your share is 1%, so across period one you receive 90,000,000 x 0.01 = 900,000 tokens. Now period two arrives and the emission halves to 45,000,000. Suppose that in the meantime the total staked weight has doubled to 20,000,000, while your own row is unchanged. Your share is now 0.5%: ``` period 1: 90,000,000 x 1.0% = 900,000 period 2: 45,000,000 x 0.5% = 225,000 900,000 / 225,000 = 4x less ``` Your income fell to a quarter, and only half of that came from the halving. The other half came from other people staking. Two multiplications, one of which is published sixteen years in advance and one of which nobody can tell you. This is why any figure presented as an annual rate on a staking page is a reading, not a promise. It is computed from the emission for the current period divided by the amount currently staked, and the second number changes every time anybody acts. The emission is a fact; the rate is a snapshot. ## What it means for someone deciding today Three honest statements follow, and no recommendation follows from them. The schedule rewards being early, and it does so by an amount you can compute exactly: a token of weight in period one earns 128 times what the same token of weight earns in period eight, if the total staked weight is the same in both. That last clause is doing a great deal of work and is almost certainly false. The schedule is a ceiling on dilution, not on price. It caps how many new tokens can ever enter circulation from staking — 179,296,875, and not one more — but it says nothing whatever about what a token will be worth. A reward paid in the same token you staked is a bigger slice of an amount this article cannot predict. And the lock is the part you cannot undo. Choosing a long lock buys a larger multiplier, as the previous article showed, and it does so by removing your ability to react to anything, including a fall in the rate you were expecting. The multiplier is the compensation for that; whether it is adequate compensation is a judgement, and it is yours. **Verify it yourself:** Add the eight period emissions above with a calculator. You must get 179,296,875, and 180,000,000 minus that must be 703,125. Then read the token contract's totalSupply() on a block explorer: divided by 10^18 it must be 1,000,000,000, of which the staking budget is 18%. Related: https://blazephoenix.xyz/learn/what-a-stake-records · https://blazephoenix.xyz/learn/bzpx-token · https://blazephoenix.xyz/learn/what-is-staking · https://blazephoenix.xyz/learn/the-dual-accumulator · https://blazephoenix.xyz/learn/check-it-yourself