Halving

What a halving is

Every 210,000 blocks — roughly four years at ten minutes a block — the block subsidy paid to miners is cut in half. It started at 50 BTC per block in 2009 and is 1.5625 BTC after the next one. Nothing else changes: no vote, no announcement, no discretion. The rule is in the code every node runs, so new supply arrives at half the previous rate on schedule, and the issuance curve ends at 21 million.

The chart plots the price on a log scale with each halving marked, and shades the 500 days before and after it.

Halvings so far

Four halvings. Four data points. Enough to describe what happened, not enough to call it a pattern.

The next one

The date is an estimate, not a fixed point on the calendar. What the protocol fixes is a block height — 1,050,000 — and blocks do not arrive on a clock. Difficulty retargets every 2,016 blocks to pull the average back toward ten minutes, but hash rate keeps growing between retargets, so blocks tend to come in slightly fast and the height arrives slightly early. The figures above assume a flat four years from the last halving; the real date drifts by weeks.

Chart

The 500-day window

The shaded bands cover the 500 days leading up to each halving and the 500 days after it — a little under three years around each event. The claim people take from it is that buying into the run-up and holding through the window has historically worked out. That is a reading of four episodes, each with its own macro backdrop, market structure and set of leveraged blowups. The supply cut is real and mechanical; the price behaviour around it is four observations of a market that was a different size every time.

Taken as a bet rather than a rule, the edges are the interesting part: accumulating as the window opens and trimming as it closes is the trade the four past cycles would have rewarded. It is a bet on the shape repeating a fifth time, with a sample of four and no mechanism forcing it — the supply cut is scheduled, the price response is not. Size it like a bet you can be wrong about, not like a plan.

The Buy Thesis

The argument for accumulating in the 500 days before a halving is a supply argument. The cut is scheduled, mechanical and public: on a known block height the subsidy halves, and the number of new coins reaching the market each day halves with it. Nobody has to agree to it. The claim is that flow matters — that if demand does not fall at the same instant, the same buying meets half the new supply, and the price is what gives.

The second half of the claim is that the market does not price the cut in advance, even though the date is on everyone's calendar. What people point at is the run-up itself: measured from the day the window opened to the day of the halving, the price rose in 3 of the 4 cycles — 2012 -11.1%, 2016 +179%, 2020 +140%, 2024 +274%. That is the whole evidence base, and it is four numbers wide.

Daily closes from the same price dataset the chart uses, read at build time. Where the window opens or closes on a day the dataset does not cover, the nearest available close is used. A cycle whose window has not closed yet says so in the row instead of showing a number.

The Sell Thesis

The mirror argument is about getting out. Each cycle has put in a high and then given most of it back, so the thesis is to hold through the halving and realise somewhere in the 500 days after it, treating the close of the window as a deadline rather than a target.

The deadline is the part the data does not support. Taking the highest daily close in the two years after each halving, the high landed 2012 day 372, 2016 day 525, 2020 day 547, 2024 day 536 — inside the 500-day window in 1 of the 4 cycles. The other 3 topped after the window had already closed, by 25 to 47 days. What did hold is the second half: every one of those highs was followed by a drawdown of half or more within the same two years.

Two years after the halving is a fixed horizon, chosen so the run-up into the next halving cannot be counted as the previous cycle's high. Measured to the next halving instead, the 2020 cycle's highest close falls in March 2024, four years later.

What Argues Against Both

There are four halvings. Four events is not a sample; it is an anecdote with dates. Every one of them happened in a market of a different size, under different regulation, with different instruments available and a different set of leveraged failures on the way down. Both theses above are read off the same four lines.

The efficient-market objection is harder to wave off than the sample size. The halving is the most telegraphed supply event in any market: the block height is fixed in code, the date is estimated years out, and every participant can see it. A cut that everyone knows is coming should already be in the price. Nothing about the schedule stops that from being true.

The 2024 halving also landed in a different regime. Spot ETFs began trading in January of that year, three months before the halving, which changed who was buying and through what. Whatever the previous three cycles measured, they measured a market without that.

And the cut itself is getting smaller. What the halving removes is a share of new supply, and new supply keeps shrinking against the coins already in existence. In 2012 the subsidy fell from 50 to 25 BTC, which took annual issuance from 25.03% of the coins then in existence to 12.51% — a cut of 12.51% of circulating supply per year. In 2024 the same halving of the subsidy removed 0.83%, about 15 times less. The next one removes 0.40%.

Issuance is 52,560 blocks a year — 144 a day at ten minutes each — times the subsidy, over the coins already mined at that block height. Coins that are lost are still counted; there is no way to measure them.

None of this says what the window is worth. It says what it is: a scheduled, shrinking supply cut, with 4 price histories around it, 3 of which put their high after the window had closed.