The current Bitcoin 200-week moving average is $66,841 (October 18, 2026), and the $82,116 close sits at 1.23× the line.
Nearly four years of weekly closes, averaged. It moves so slowly that, in our series from 2014 to 18 Oct 2026, it hasn’t fallen in a single week. This page plots price against it, the ratio between the two, how much the average rose over the last four weeks, and every week that closed below it.
Today’s count: up to the weekly close of 18/10/2026, 55 of 644 weekly closes since 2014 were below the average.
“Bitcoin 200-week moving average: $66,841, with price at 1.23× the line, 18 Oct 2026 close.” NodeWitness, https://nodewitness.com/en/charts/bitcoin-200-week-moving-average-chart/
Think of mean sea level in a harbour. Waves, the price, rise and fall every day, and storms push them far up or down; the mean level, measured over nearly four years, barely moves and, up to the 11 Oct 2026 weekly close, had only risen. What matters is how far the waves stray from it. At each cycle’s highest close, price stood at 15.4 times the average in 2017, 3.9 in 2021 and 2.3 in 2025, and weekly closes below it have been rare.
Top panel: price and the average on a log scale, with weeks that closed below shaded Monday to Sunday. Middle: price divided by the average, where 1 means price is sitting on the line. Bottom: the average’s 4-week change in percent, the figure the well-known 200-week heatmap uses for its colours.
Because the average doesn’t fall, the cycle shows up in how fast it rises. It speeds up in strong rallies, reaching 27.1% in the four weeks to 31 Dec 2017, and slows in long bear markets: up to the weekly close of 18 Oct 2026, its slowest pace was 0.97%, on 1 Mar 2026.
The coloured version, the 200 Week Moving Average Heatmap, was created by PlanB in January 2019, according to Look Into Bitcoin, which publishes a slightly modified version [3]: each price point is coloured by how much the average rose over the past month. On the plain average, Galaxy Research calls it the floor of Bitcoin’s bear markets, while stressing it isn’t a perfect one [1].
In our series, the 14 Jan 2015 low closed 10.8% below the average, yet the week closed above it; the 15 Dec 2018 low stayed 1.6% above it. 2022 was the exception: on 21 Nov 2022 price closed at 0.65 times the average, the lowest reading in the series, and between 21 Aug 2022 and 12 Mar 2023 it strung together 30 weekly closes below. Up to 17 Oct 2026, the year’s lowest close was $58,534 on 30 Jun, 6.5% under the average, and up to the weekly close of 18 Oct 2026 only three weeks of 2026 had closed below it.
Runs of consecutive weeks whose Sunday close was below the average, newest first. Dates are Sundays (week close).
We start from our USD daily close series [2]. Weeks run Monday to Sunday (UTC) and the weekly close is Sunday’s close. Each week’s official average is the simple mean of the last 200 Sunday closes; the first complete one is from 22 Jun 2014.
Between Sundays we show the live average for the current week: the previous 199 Sundays plus that day’s close, divided by 200. On Sundays it equals the weekly value. It’s what a weekly chart would have shown on that day, and it’s the convention that comes closest to the reference we checked against [1]. The 4-week change compares each day’s average with its value 28 days earlier.
We compared our series with a public Galaxy Research report [1], published on 28 Aug 2026 with data up to the 23 Aug 2026 weekly close: the average comes out at $23,944 on 9 Nov 2022 (report: $23,951) and $13,974 on 20 Jul 2021 ($13,981). To compare counts, we cut our series at that same close: up to the 23 Aug 2026 weekly close we count 55 weekly closes below the average out of 636 weeks, against the report’s 56 of 642 (its average starts in May 2014, ours on 22 Jun 2014). Over that period, neither our average nor the report’s fell from one week to the next. That count stays frozen at 23 Aug 2026; today’s, covering every week up to the latest close, is the one shown above the chart and in the FAQ.
The average needs four years of data, so it only exists from 2014 and misses Bitcoin’s first two cycles. The result shifts slightly depending on which day is taken as the week’s close. And an average that hadn’t fallen in a single week up to 11 Oct 2026, or a price that had rarely closed below it, says nothing about what comes next: in 2022–23 it spent 30 straight weeks underneath. This page describes the series; it is not investment advice.
How much supply would sit at a loss at each price below this line, from $60,000 down to $10,000, is measured from our node in our McGlone vs Fidelity piece, on the clash between Bloomberg Intelligence’s strategist and Fidelity’s, who reads price through a power law. The short definition is in the glossary, and where the ratio sits today next to ten other indicators is in Bitcoin cycle top and bottom indicators today.
Not in our series. From its first value on 22 Jun 2014 to the weekly close of 18 Oct 2026, the weekly average never fell from one week to the next, matching Galaxy Research’s count up to 23 Aug 2026. What changes is how fast it rises.
Up to the weekly close of 18 Oct 2026, 55 out of 644: five in 2015, one in March 2020, 46 between June 2022 and October 2023, and three in 2026. The up-to-date count is shown above the chart.
A way of drawing the same average where each price point is coloured by how much the average rose over the past month. PlanB created it in 2019. We show that figure as it is, as a 4-week change, in the third panel.
Because they depend on which day counts as the weekly close and on the price series. We use Sunday’s close in UTC and, between Sundays, a “live” average that includes the day’s close. With that convention our figures land within $10 of Galaxy Research’s on the dates we compared.
The rest of the cycle metrics are on Cycle & Score.