Glossary

    Glossary

    Difficulty Adjustment

    The difficulty adjustment is Bitcoin's rule for retargeting its mining puzzle every 2,016 blocks - roughly a fortnight - so blocks keep arriving every 10 minutes whatever the hash rate does. It is the thermostat behind the block schedule.

    FoundationsLast reviewed 2026-10-01

    What it is

    Bitcoin aims to produce a block every 10 minutes, and it enforces the target with a self-adjusting puzzle. The puzzle is the proof-of-work difficulty: a threshold the winning miner's guess must beat. Every 2,016 blocks - about a fortnight at the target rate - the network compares how long the last period actually took with the two weeks it should have taken, and recalibrates. More hash power arriving pushes difficulty up; hash power leaving pushes it down.

    The adjustment is capped at a factor of four in either direction per retarget, which smooths the network through sudden swings in participation. The rule is why the block schedule holds steady, why block rewards halve on a roughly four year clock, and why a Bitcoin transfer averages 10 minutes whether the network has a thousand miners or a million. For Australian readers, it is also the co-author of mining revenue: the same hardware earns less bitcoin as difficulty climbs.

    Key takeaways

    • Every 2,016 blocks - roughly a fortnight - Bitcoin recalibrates its mining difficulty.
    • The target is a 10 minute average block time, whatever the hash rate does.
    • The adjustment is capped at a factor of four per retarget, in either direction.
    • Rising difficulty means rising competition for the same block rewards, not more blocks.
    • The retarget is why halvings land on a roughly four year clock and transfers average 10 minutes.

    How it works

    The retarget runs on a formula: new difficulty equals old difficulty multiplied by the expected time for 2,016 blocks divided by the actual time. The expected span is two weeks. If the last 2,016 blocks arrived every 9 minutes, the period finished in roughly 12.6 days, the ratio lands above one, and difficulty rises about 11 percent to push the average back toward 10 minutes. If miners left and blocks stretched to 12 minutes, the ratio falls below one and difficulty drops. The result is one number, set by the network itself, that no participant can vote on.

    The per-retarget cap of a factor of four exists for edge cases, and history has tested it. In mid-2021, China's mining ban removed roughly half the network's hash rate within weeks; block times stretched past 15 minutes, and the July 2021 retarget cut difficulty by roughly 28 percent - among the largest single adjustments on record. Timestamps are checked against the previous block to stop miners gaming the measurement, and the retarget lands at block heights divisible by 2,016. Difficulty is a setting, not a spend: the hash rate is what miners actually burn, and the adjustment chases it every fortnight.

    Downstream, the schedule drives everything an Australian reader touches: deposit confirmations arrive on the 10 minute rhythm, mempool congestion builds when blocks slow between retargets, and issuance stays predictable because the block count, not the calendar, pays the rewards. The hash-rate entry covers the security budget the adjustment is chasing.

    Why it matters

    For anyone mining in Australia, the difficulty chart is the other half of the revenue calculation. Australian electricity prices around 30 cents per kilowatt-hour mean a mid-range machine burning 3,250 watts costs about $23 a day to run. Whether it clears a profit depends on the bitcoin it wins per day - which the difficulty resets every fortnight. A colleague named Brent (yes, that Brent) bought a second-hand miner after watching a video. He calculated his revenue on last period's difficulty and met his first retarget before his first power bill.

    For everyone else, the adjustment is background plumbing that explains the two things they notice. Transfers sometimes slow for a fortnight, and mining is an industrial business rather than a hobby with a laptop. The ATO treats mining rewards as ordinary income at their AUD value when received, which makes the difficulty trend part of the paperwork too. The issuance side of the schedule - block rewards halving on a roughly four year clock - is the bitcoin-halving entry's subject, and the income treatment is the cryptocurrency-mining entry's.

    Real-world examples

    A retarget after miners arrive

    Suppose the last 2,016 blocks arrived every 9 minutes instead of 10, finishing the period in roughly 12.6 days against the expected 14. The ratio of expected to actual time is about 1.11, so the next period's difficulty rises roughly 11 percent. Miners earn the same block reward for harder work, and block times drift back toward 10 minutes within the fortnight. Nothing about Bitcoin's issuance changed - the same 3.125 bitcoin per block is simply contested by more hash power.

    The July 2021 collapse in difficulty

    When China banned mining in mid-2021, roughly half the network's hash rate went offline within weeks and block times stretched past 15 minutes. The July 2021 retarget measured the slow period and cut difficulty by roughly 28 percent - among the largest single adjustments on record. Blocks returned to a 10 minute average within days, and the hash rate rebuilt over the following months. The factor-of-four cap was nowhere near tested; the fall never needed it.

    The electricity arithmetic at 30 cents per kilowatt-hour

    A mid-range ASIC drawing 3,250 watts consumes about 78 kilowatt-hours a day, which at 30 cents per kilowatt-hour costs roughly $23.40 AUD. Whether that machine earns more than $23 of bitcoin depends on the reward it wins per day at the current difficulty - and the difficulty is recalibrated every 2,016 blocks. A retarget that lifts difficulty 11 percent lifts the electricity cost per bitcoin won by the same proportion. The schedule, not the price alone, decides which side of zero the machine sits on.

    Common mistakes

    Mistake: Calculating mining revenue on last period's difficulty.

    Why it bites you: The next retarget resets the arithmetic every 2,016 blocks. Brent's revenue projection met its first retarget before his first power bill arrived.

    Mistake: Reading fast blocks as a healthy network.

    Why it bites you: Blocks arriving under 10 minutes mean too much hash power for the current difficulty, and the retarget will correct them. On schedule is better; fast is just loud.

    Mistake: Expecting the adjustment to respond instantly.

    Why it bites you: The retarget waits for the full 2,016 block period. A sudden hash rate change means stretched or compressed blocks for up to a fortnight first.

    Mistake: Confusing difficulty with hash rate.

    Why it bites you: Difficulty is the setting the network adjusts; hash rate is the spend miners choose. One chases the other every fortnight, and the lag between them is the schedule noise.

    Ivy's take

    I ran the retarget arithmetic through a decade of adjustment history so you don't have to. The pattern is beautifully boring: the difficulty chases the hash rate, overshoots slightly, and corrects. The largest single adjustment on record - the roughly 28 percent cut of July 2021 - happened because a nation-state's mining fleet switched off. Even that resolved inside a fortnight of blocks.

    The recurring Brent conversation is the YouTube miner who calculates revenue on a good week and is surprised by a fortnight that resets the rules. My position: if the business case only survives one difficulty setting, it is not a business case. The schedule is the most honest part of the network - it adjusts to reality on a clock, which is more than most portfolios manage.

    Related jargon, defined

    Hash rate
    The total computing power a proof-of-work network spends securing its ledger.
    Block
    A batch of confirmed transactions sealed with a hash and chained to its parent.
    Block reward
    The newly issued bitcoin a winning miner claims, currently 3.125 per block.
    Retarget period
    The 2,016 block span between difficulty adjustments, roughly a fortnight.
    Mining
    The proof-of-work race to add blocks, paid in electricity and rewarded in bitcoin.
    Mempool
    The waiting queue of broadcast transactions not yet in a block.

    Frequently asked questions

    How often does Bitcoin adjust its difficulty?+

    Every 2,016 blocks, which is roughly a fortnight at the 10 minute target. The adjustment lands automatically at the first block height divisible by 2,016.

    What happens when blocks arrive too fast?+

    The next retarget raises difficulty to push the average back to 10 minutes. Until then, blocks keep arriving fast and new bitcoin is issued slightly ahead of schedule.

    Can difficulty fall?+

    Yes - when hash power leaves the network, as it did after China's 2021 mining ban, when difficulty fell roughly 28 percent. The fall is capped at a factor of four per retarget.

    Does the difficulty adjustment affect my transaction fees?+

    Indirectly. When blocks slow between retargets, the mempool backs up and fees rise; the retarget restores the schedule and fees settle.

    Is the difficulty adjustment written into the Bitcoin whitepaper?+

    The whitepaper describes proof-of-work and a difficulty targeting mechanism; the specific 2,016 block retarget lives in the software. Bitcoin.org's developer documentation documents the rule as implemented.

    Disclaimer: This information is general in nature and does not constitute financial or legal advice. Always consult a qualified professional for your specific situation.

    Keep learning

    Ivy Sinclair-Wren

    Ivy Sinclair-Wren

    Financial Chaos Analyst

    Connect on LinkedIn

    Ivy Sinclair-Wren is a Financial Chaos Analyst covering crypto markets, investing behaviour, and the Australian habit of checking prices during family dinners. Based in Melbourne, she reviews AUD Crypto Tracker's coin guides for accuracy, ATO alignment, and ruthless utility — no hype, no predictions, just what the data says.