Hash ribbons are a Bitcoin buy signal built from miner hashrate, and 18 of the 21 signals with a full 180-day record since May 2012 closed higher six months later. The signal fires when hashrate's 30-day average climbs back over its 60-day average and Bitcoin's 10-day price average holds above its 20-day.
We rebuilt the indicator from the Tuesday 15 September 2026 close back to 2012. That 85.7% hit rate compares with 64.1% for any day bought blind over the same years. Since 2024 the record reads 4 of 6, and the buys of 23 May 2025 and 23 July 2025 stood 14.7% and 22.0% lower after 180 days.
Hash Ribbons at a Glance
Metric | Details |
What it measures | 30-day vs 60-day simple average of Bitcoin hashrate |
The buy rule | 30-day back above 60-day, and price 10-day average above 20-day |
180-day record | 18 of 21 higher (85.7%), median +53.7% |
Blind control | 64.1% of days since 23 May 2012 higher 180 days on |
Since 2024 | 4 of 6 higher at 180 days |
15 September 2026 | Recovered 24 August, price leg below its trigger |
What Are Hash Ribbons?
Capriole published the indicator on 30 October 2019 as two simple moving averages of Bitcoin's hash rate, one covering a month and the other covering two. Hash rate measures the computing power pointed at the network, and Phemex Academy's guide to Bitcoin hash rate and network security explains where that number comes from. When the short average drops under the long one, miners are switching machines off faster than new ones arrive. Drawn on a chart the two lines twist over each other like a ribbon, which gave the hash ribbons indicator its name. The original post shows the same crossing on mining difficulty about two weeks later, so hashrate became the faster of the two inputs.
The trading logic comes from miner economics. A miner who can't cover the power bill sells coins and then shuts down, so miner capitulation puts forced supply on the market at weak prices.
Capriole's post puts the turn in plain words. When the one-month average crosses back over the two-month average, "the worst of the miner capitulation is typically over, and the recovery has begun."
How the Hash Ribbon Buy Signal Works
The hash ribbon buy signal needs two conditions to line up. Capitulation ends when the 30-day hash rate moving average crosses back above the 60-day, and a buy only prints once Bitcoin's 10-day price average sits above its 20-day. Capriole added that price leg in the 2019 post after one hash-only entry in January 2015 drew down 42%. The post credits the filter with removing most of those drawdowns without cutting returns.
Phemex Academy's walkthrough of trading crypto with moving averages covers how the price half of the rule gets built.
The two legs answer different questions. The hashrate leg tells you when miners stopped bleeding, and the price leg tells you when buyers came back to the chart. A trader who buys on the hashrate cross alone often gets in while the selling is still running, which is the January 2015 problem Capriole flagged. The word capitulation has a price-side meaning too, and the guide to spotting capitulation at a market bottom walks through that version.
How We Recomputed Bitcoin Hash Ribbons
Method: hashrate comes from the blockchain.com hash rate chart for its full history. Each average needs at least 75% of the calendar days in its window, because the feed skips 13 to 15 November 2025 and 17 to 23 August 2026. This page uses that one hashrate feed and no second source.
Prices are Phemex BTCUSDT perpetual daily closes from 21 December 2023. Before that date the study uses blockchain.com's daily market price shifted by one day, because its value for a given date matches the Phemex open of that date. Across the 1,000 days where both series exist, the shifted figure lands a median 0.054% from the Phemex close.
The raw ribbon whipsaws. Hashrate's 30-day average dropped under the 60-day 40 times since 2012, and several of those capitulations lasted one to three days in 2015 and 2025. We merged any recovery that lasted fewer than 10 days into the capitulation that followed it, which leaves 26 episodes. A buy is the first day on or after a recovery with the 10-day price average above the 20-day, taken only while the ribbon still reads recovered. Four of the 26 recoveries never produced one before hashrate rolled over again, and that leaves 22 buy dates.
Outcomes are the close 30, 90 and 180 days after each buy. The control buys every single day since 23 May 2012 and holds for the same span.
What the Hash Ribbons Record Shows
Holding period | Signals higher | Median move | Any day bought blind |
30 days | 15 of 21 (71.4%) | +7.1% | 57.4% |
90 days | 15 of 21 (71.4%) | +26.6% | 59.2% |
180 days | 18 of 21 (85.7%) | +53.7% | 64.1% |
The edge grows with the holding period. A buy on the signal beat a blind entry by 21.6 percentage points at 180 days and by 12.2 points at 90, which fits an indicator built to find cycle lows and not the next few weeks. The 19 August 2022 buy at 20,843 shows the cost of that design, down 20.0% after 90 days before it finished 16.8% higher at 180.
The early signals carry the reputation. The 10 January 2019 buy at 3,646 closed 245.2% higher 180 days later, and the 12 July 2020 buy at 9,296 gained 337.5% over the same span. Since the start of 2019 the count is 12 of 15, with the misses on 7 August 2021 at 44,634 and on the two 2025 dates.
The recent record is thinner. Six signals since 2024 have a full 180 days behind them and four closed higher, a 66.7% rate against the 64.1% a blind buyer collected. At 30 days only two of those six were up. The Bitcoin hash ribbons buys at 107,229.1 on 23 May 2025 and 118,697.0 on 23 July 2025 were the two highest entry prices in the study, and both were underwater six months on.
Every figure before 21 December 2023 rests on blockchain.com prices alone, and 21 scored trades across 14 years is a small sample for any win rate.
What Did Hash Ribbons Show on 15 September 2026?
The latest buy printed on 24 August 2026 at a close of 78,950.4. Hashrate's 30-day average had spent most of the 77 days from 8 June below its 60-day, and it crossed back inside the seven-day gap in the feed. So the recovery date could move by a day or two with a cleaner series.
On the 15 September bar the 30-day average read 935.1 EH/s against 922.2 EH/s for the 60-day, a gap of 1.41%. The daily estimate for that bar came in at 1,032.7 EH/s. Phemex's read on what a record difficulty jump signals for BTCcovers the mining side of that climb.
Price has gone the other way. Bitcoin closed the Tuesday 15 September session at 75,601.9 on Phemex, 4.24% under the signal price, and its 10-day price average of 77,760 slipped below the 20-day at 78,269. The ribbon still reads recovered while the momentum leg has switched off. The longer trend points the other way, because the 50-day average crossed above the 200-day on 8 September, the setup Phemex Academy explains in its guide to the Bitcoin golden cross and death cross.
The previous recovery never reached a buy at all. Hashrate crossed back on 23 May 2026, price momentum stayed negative, and the ribbon rolled into capitulation again on 8 June before the price leg turned.
How to Read Hash Ribbons Without Overtrusting Them
Treat the hashrate line as an estimate. blockchain.com infers it from how fast blocks arrive, so a lucky or slow run of blocks moves the number without a single machine switching on. Daily readings between 6 and 12 September 2026 ranged from 843 to 1,045 EH/s, which is why the study only trusts the averages. Hashrate also falls for reasons that have nothing to do with price, and Phemex covered one of them in its report on tariffs on mining equipment and hash rate.
The signal is slow by design. A 60-day average needs weeks of lower readings before it flags anything, and 22 buys in 14 years means you wait a long time between entries.
Read the result window as part of the signal, because a trade you judge after one month judges the wrong thing. A buy that looks wrong after 30 days has been wrong six times in 21 at that horizon and still delivered at 180 days in most of them.
Bottom Line
Hash ribbons earned their name on buys near the 2019 and 2020 cycle lows, and the 180-day record since 2024 sits within three points of buying any day at random. The 24 August 2026 signal is the next test of which record is real. Its 30-day result lands on 23 September, and the price leg was already pointing down on the 15 September close. A trader who takes the signal is making a six-month call on miners, and the first month has said very little about how those calls end.
Disclaimer: This article is for informational purposes only and is not financial advice. Cryptocurrency trading involves substantial risk. Always do your own research before making investment decisions.
