News
Why Halving Dates Drift: Difficulty Retargets and the Real Block Clock
A halving happens at a block height, not a date. Any calendar estimate assumes ten-minute blocks, and real blocks run faster or slower depending on hash rate between difficulty retargets.
Every halving countdown makes the same simplification: it multiplies the blocks remaining by ten minutes. That gives a date. But the protocol does not know about dates. The cut happens at a block height, and how fast the chain reaches that height depends on how much hash power is mining between difficulty adjustments.
The retarget rule
Bitcoin Core’s chain parameters set a target timespan of two weeks and a target block spacing of ten minutes. Dividing one by the other gives the retarget period: 2,016 blocks. At the end of each period, the node compares how long those blocks actually took with two weeks and scales difficulty to match. If blocks came too fast, difficulty rises; too slow, it falls.
There is a safety limit. In pow.cpp, the measured timespan is clamped to no less than a quarter and no more than four times the target before the new difficulty is computed. A single retarget can therefore move difficulty by at most a factor of four in either direction.
Why blocks still run fast or slow
Difficulty is always catching up. When new machines come online during a period, blocks arrive faster than ten minutes until the next retarget raises difficulty. When machines switch off, as after a halving, blocks arrive slower. Over long stretches of rising hash rate, Bitcoin has tended to produce blocks somewhat faster than the target, which pulls halving dates earlier than the simple countdown suggests.
That is why the estimated date on our supply events calendar moves. It is recalculated from the live tip height, which mempool.space reported at 976,417 when this note was written, against a next halving at block 1,050,000.
What the difficulty number does and does not mean
Difficulty is the cleanest public proxy for how much work stands behind the chain. It is often over-read. A large upward adjustment is described as miners “expressing confidence”; a downward one as “capitulation”. Both are mostly the lagged result of hardware deliveries, power contracts, seasonal conditions and price over the previous weeks. Difficulty measures the past fortnight, not intentions.
Hash rate, which difficulty tracks, measures the computing power securing the chain. Higher hash rate makes it more expensive for one party to rewrite recent blocks. It is one of the few Bitcoin metrics that describes security directly, which is why it is worth knowing, and also why it says little about price.
Why this matters for supply
- Supply per block is exact; supply per calendar year is approximate.
- A halving’s date estimate is a function of hash-rate trends, so it is a moving target until the final weeks.
- Annual issuance figures that assume exactly 52,560 blocks a year are approximations; our Supply Clock labels its rate as computed at the target block time.
The same logic applies to every subsidy chain we track, with different constants: see Litecoin and Bitcoin, Bitcoin Cash and our note on miner economics after a halving.
Sources checked for this page
- Bitcoin Core: pow.cpp (read 2026-11-18)
- Bitcoin Core: chainparams.cpp (read 2026-11-18)
- mempool.space API: tip height (read 2026-11-18)
Frequently asked questions
Why is a halving date only an estimate?
The protocol triggers a halving at a block height. Converting blocks remaining into a date assumes the target block time, while real block times vary.
How often does Bitcoin adjust difficulty?
Every 2,016 blocks. The target timespan is two weeks, which at ten-minute spacing equals 2,016 blocks.
Can difficulty change by any amount?
No. The measured timespan is clamped to between a quarter and four times the target before the new difficulty is computed, per pow.cpp.
Not financial advice. Ecoin Supply describes how crypto-asset supply works and what public sources report. Crypto-assets are volatile and you can lose all of the money you put in. Nothing on this site is a recommendation to buy, sell or hold any asset, and no figure here is a price prediction.