Bitcoin Halving Cycles, Compared

The four-year cycle is the most durable story in bitcoin. Every halving cuts new supply in half, the story goes, and roughly eighteen months later the market peaks; the schedule is written in the protocol, so the rhythm repeats. It is a tidy narrative, it is repeated constantly, and it rests on a very small number of observations.

This article compares exactly two halvings: 11 May 2020 and 20 April 2024, both UTC, using nothing but our own daily closes, and is explicit about where the comparison runs out. Our price series now reaches back to 2011 and formally covers the 2012 and 2016 halvings as well, but those ran on a market a fraction of today’s size; including them would be a different comparison rather than more of this one. Two is enough to describe. It is not enough to generalise, and we will not pretend otherwise.

What the halving actually does

Every 210,000 blocks — roughly every four years at the protocol’s target block time — the reward paid to miners for a new block is cut in half. This is arithmetic in the software, not a policy decision, and the schedule is knowable in advance in blocks. It is only approximately knowable in dates, because blocks arrive faster or slower than the ten-minute target depending on how much hashing power is running.

What the halving does with certainty is reduce the rate at which new coins enter circulation. What it does not do is oblige anyone to buy. The supply reduction is public, dated and priced in by anyone who wants to price it in, which is the standard objection to the story and a fair one.

Two earlier halvings occurred, in November 2012 and July 2016. Our price series does now cover them, but on exchange volumes and liquidity so much thinner that the cycles are not comparable to the two below, and nothing in this article draws on them. The next halving falls at block 1,050,000, expected around April 2028 — an estimate that moves with block times rather than a fixed date.

The two cycles we can measure

The first thing our data says is that “four years” is already an approximation. The two halvings are 1,440 days apart, not 1,461. That is three weeks of slippage in a single interval, and it comes from block times running slightly ahead of target.

	2020 cycle	2024 cycle
Halving date (UTC)	11 May 2020	20 April 2024
BTC close on the day	$8,561.52	$64,940.59
Highest close after the halving	$67,525.83 on 8 Nov 2021	$124,773.51 on 7 Oct 2025
Days from halving to that high	+546	+535
Multiple of the halving-day price	7.9×	1.9×
First close at 2× the halving price	+190 days	never
Days closing below the halving price, +1 to +826	0 of 826	166 of 826
Deepest fall from the cycle high, within +826 days	−71.9% (18 Jun 2022)	−53.1% so far (1 Jul 2026)

Normalised to the halving date

Setting each halving-day close to 100 puts the two cycles on the same axis. The vertical scale below is logarithmic, because a linear one would flatten the 2024 cycle into a horizontal line next to the 2020 cycle.

Days from halving	2020 cycle	2024 cycle	Ratio
−365	81.4	42.0	—
−180	102.5	50.9	—
0 (halving)	100.0	100.0	1.00×
+90	136.4	102.6	1.33×
+180	173.1	103.8	1.67×
+365	661.9	131.2	5.05×
+547	782.0	165.0	4.74×
+730	339.9	113.7	2.99×
+826	281.4	98.7	2.85×

Where the two cycles agree

There is one genuinely striking similarity, and it is about timing. The 2020 cycle’s highest close came 546 days after its halving. The 2024 cycle’s highest close so far came 535 days after its halving. Eleven days apart, across intervals of roughly a year and a half.

Both cycles also rose in their first six months, both fell substantially in the months after their high, and the correlation of the two normalised paths over the first 820 days after each halving is 0.71 — the general shape rhymes.

Now the caveat, which is larger than the finding. Two matching dates out of two observations is one coincidence away from being nothing at all, and the 2024 peak may not be final — the cycle is still running. An eleven-day agreement between two events is the kind of pattern that feels like a law and carries no statistical weight whatsoever.

Where they do not agree at all

Everything about the size of the move is different, and the gap is not subtle.

• Amplitude differs roughly fourfold. The 2020 cycle reached 7.9× its halving-day price. The 2024 cycle has reached 1.9×. At the one-year mark the 2020 path stood at five times the 2024 path.
• One cycle never looked back; the other has, repeatedly. In the 826 days after the 2020 halving, bitcoin did not close below its halving-day price on a single day. In the same window after 2024 it has done so on 166 days — including 25 July 2026, which leaves the 2024 index at 98.7.
• They entered from opposite conditions. The 2020 halving arrived 60 days after a crash that took bitcoin to $4,800 on 12 March 2020 — 44% below the halving-day close it would recover to within two months. The 2024 halving arrived after a long recovery, with price already 2.6× its June 2023 low.
• The drawdowns differ. Measured over the same 826-day window, the 2020 cycle gave back 71.9% from its high; the 2024 cycle is 53.1% below its high at the deepest point so far, and its window is not finished.

Put plainly: if you had taken the 2020 cycle as a template in April 2024 and sized a position on it, the shape would have been roughly right and the magnitude wrong by a factor of four. In practice, magnitude is the part that determines outcomes.

The asset that does not halve

Ether has no halving. Its issuance schedule has nothing to do with bitcoin’s block subsidy. So it makes a useful control: if the halving were the dominant driver of the move, the asset without one should not have kept pace.

Indexed to the same two dates, ether stood at 2,548 at +547 days in the 2020 cycle — a 25× move against bitcoin’s 7.8× at the same point — and at 123 at +547 days in the 2024 cycle. At +826 days the two cycles read 1,022 and 59.

That does not disprove a halving effect; ether may simply ride bitcoin’s liquidity. But an asset with no supply event outrunning bitcoin threefold inside bitcoin’s own post-halving window is hard to reconcile with a story in which the supply cut is doing the work. Whatever moved both markets in 2020–21 was substantially larger than the difference between 12.5 and 6.25 new coins per block.

Two observations are not a cycle

The honest summary of the table above is short. We hold two post-halving paths. They agree closely on when the high arrived and disagree by a factor of four on how high it went. The interval between the halvings was 1,440 days rather than the four years everyone quotes. And the second cycle is still running, so half of what is written here may read differently in a year.

Where the “four-year cycle” claim goes wrong is not in noticing a rhythm — the rhythm is visible. It goes wrong in the step after, where an observed rhythm becomes a schedule you can plan around. With two measured cycles, any statement of the form “the market always does X by month Y” is describing two events and calling it a law.

Why our index still uses a 1,460-day window

The Hodlometer Index ranks every component against its own trailing 1,460 days — four years. Given everything above, that deserves an explanation, because it looks like the cycle belief we just spent an article declining to hold.

It is a choice about memory length, not about cycles. A percentile window has to be long enough to contain both a hot and a cold regime, or “unusual” stops meaning anything; and short enough that a market structure from a decade ago is not still setting the scale. Four years satisfies both, and the fact that it lands near the halving interval is a convenience rather than a justification. A 1,200-day or 1,700-day window would produce a very similar index.

The distinction matters. The window says: compare today with the recent past. It does not say: the recent past will repeat on schedule. If our index depended on the second claim, the two cycles compared here would already have broken it.

Where the current cycle stands

On 25 July 2026, 826 days past the 2024 halving, bitcoin closed at $64,099 — 48.6% below the all-time high of $124,773.51 set on 7 October 2025, and fractionally below the halving-day close of $64,940.59. The 2020 cycle at the same point stood at 281 on the same index.

Two readings of that are available. One says the current cycle is late and following the same downward leg the last one did, just with less altitude to give back. The other says the two paths were never close enough for the comparison to carry weight. Our data supports the second more comfortably than the first, and neither is a forecast.

For the metrics that describe where price sits now rather than where the calendar says it should be, see the pieces on the Mayer Multiple and the 200-week moving average, or the Hodlometer Index that combines them.

Frequently asked questions