Bitcoin difficulty adjustment: live tracker and how it works
Tracker · Bitcoin network · Published 14 October 2026 · By TradeSignalsHQBitcoin has no clock and no central operator, yet blocks keep arriving roughly every ten minutes whether a few machines or millions are mining. The difficulty adjustment is the reason. Every 2,016 blocks, each node looks at how long that period took and makes the puzzle harder or easier so the next period lands closer to two weeks. This page tracks the current period live and explains the rule, including the limits and quirks that the headline number leaves out. It is educational only.
The estimate assumes the rest of the period keeps the pace seen so far, so it swings most in the first few days. "T" means trillion. Source: mempool.space.
What difficulty measures
To add a block, a miner has to find a block header whose hash is at or below a number called the target. Hashes are effectively random, so the only way to find one is to try again and again with small changes. Bitcoin's developer documentation puts it simply: the lower the target, the more hash attempts will be needed on average (Bitcoin developer guide: block chain).
"Difficulty" is the same information turned into a friendlier number. It is the easiest-ever target divided by the current target, so a difficulty of 2 means the target is half the original and each block needs twice the work on average (Bitcoin Wiki: difficulty). When people say "difficulty rose 5%", they mean the target got about 5% smaller and the expected number of hashes per block grew by 5%.
The retarget rule, step by step
- Wait for the boundary. Difficulty only changes on blocks whose height is a multiple of 2,016. In Bitcoin Core the check is literally "Only change once per difficulty adjustment interval" (pow.cpp).
- Measure the period. Take the timestamp of the last block of the period and subtract the timestamp of the first. The ideal is 1,209,600 seconds — two weeks — which is 2,016 blocks at 600 seconds each (chainparams.cpp).
- Clamp extreme results. If the period took less than a quarter of two weeks, it is treated as a quarter; if it took more than four times two weeks, it is treated as four times. One adjustment can therefore move difficulty by at most a factor of four either way.
- Scale the target. Multiply the old target by actual time ÷ ideal time. Slow period, bigger target, easier blocks; fast period, smaller target, harder blocks.
- Every node checks it. No one announces the new difficulty. Each node computes it from the block headers it already has, and rejects blocks that do not meet it.
The white paper describes the intent in one line: proof-of-work difficulty is set by a moving average targeting an average number of blocks per hour, so that if blocks are generated too fast, the difficulty increases (Nakamoto, 2008, section 4).
Three worked examples
In terms of difficulty, the rule is new difficulty = old difficulty × 14 days ÷ actual days (before the clamp).
| The last 2,016 blocks took | Average block time | Multiply difficulty by | Change |
|---|---|---|---|
| 13 days | 9 min 17 s | 14 ÷ 13 | +7.69% |
| 18 days | 12 min 51 s | 14 ÷ 18 | −22.22% |
| 2 days | 1 min 26 s | 14 ÷ 2 = 7, clamped to 4 | +300% (the cap) |
The live estimate works the same way. If blocks in the current period have averaged 570 seconds rather than 600, the estimate is 600 ÷ 570 − 1 = +5.26%. The 2-day row is not hypothetical for young networks: when a coin shares Bitcoin's design but has little hashrate, a sudden arrival or departure of miners can hit the cap.
The limits and the off-by-one
The developer guide notes that the timestamps compared are those of the first and last block of the period, so the measured span covers 2,015 block intervals rather than 2,016 — a long-known quirk that slightly skews the result (developer guide). The effect is small, about 0.05%, and it stays because changing it would need every node to upgrade together.
Two other limits matter more in practice. First, difficulty looks backwards. It reacts to the last two weeks, so a sharp drop in hashrate — a heat wave in a mining region, a ban, a halving that makes old machines unprofitable — is felt as slower blocks for days until the next boundary. Second, block timestamps are set by miners within rules, not by a shared clock, so a single block's time can be off by minutes; only the average over many blocks is meaningful.
Difficulty and hashrate
Nobody can count the machines mining Bitcoin. What the network does reveal is how much work went into each block, and from that you can estimate the total. A common approximation is hashrate ≈ difficulty × 2³² ÷ 600, in hashes per second (Bitcoin Wiki). With difficulty at 132.7 trillion, its level in mid-October 2026, that gives about 950 EH/s (exahashes per second, or 10¹⁸ hashes per second).
That figure assumes blocks come at exactly ten minutes. When they come faster, the real hashrate is higher. That is why data sites, including the 3-day average we show from mempool.space, can read noticeably above the difficulty-implied figure late in a fast period, and why the next adjustment then tends to be upward. More on what hashrate means for miners' revenue is on our hashrate and hashprice page.
How to read the estimate
- Early in the period it is noise. After 100 blocks, a few lucky or unlucky hours move the average a lot. After 1,500 blocks, the estimate is usually close.
- A rising difficulty means more competition. For a miner with a fixed fleet, each upward step is a cut in expected bitcoin earned, unless the price or fees rise to compensate.
- It does not predict the price. Difficulty follows miners' investment decisions, which follow past prices and power costs. Treat it as a description of the network's security budget, not a trade trigger.
- It sets up the halving. The halving does not change difficulty directly; it changes miners' revenue, and the next adjustment records how many of them kept their machines on. See the halving countdown.
Questions people ask
- How often does Bitcoin difficulty change?
- Every 2,016 blocks, which is about every two weeks when blocks arrive at the target pace of ten minutes. The exact date depends on how fast the blocks come.
- How much can difficulty change at once?
- At most a factor of four in either direction in a single adjustment, because the measured time is clamped to between a quarter of and four times the two-week target.
- Why is the estimated change so different from day to day?
- It extrapolates from the blocks seen so far in the period. Early on, a handful of fast or slow blocks moves the average a lot; the estimate settles as the period fills up.
- Does higher difficulty make transactions slower?
- Not on average. The adjustment exists so that average block times return to about ten minutes. Transactions wait when blocks are full and their fee rate is too low, which is a separate issue.
Sources
- Bitcoin developer documentation, Block chain: proof of work and the difficulty adjustment.
- Bitcoin Core source code, pow.cpp (GetNextWorkRequired, CalculateNextWorkRequired) and kernel/chainparams.cpp.
- Bitcoin Wiki, Difficulty.
- S. Nakamoto, Bitcoin: A Peer-to-Peer Electronic Cash System, 2008, section 4 "Proof-of-Work".
- Live data: mempool.space API (difficulty adjustment, hashrate).
Risk warning. Educational content only, not financial or investment advice. Crypto assets are highly volatile and you can lose all the money you put in.