Bitcoin Broke $70,000 USD. Its Death Cross Record Is Weaker Than the Name Suggests
Bitcoin traded as low as $68,052.44 USD overnight into Friday, and the last daily bar that went lower was August 19, nearly three months ago, and that August bar reached $64,111.07 USD. Thursday’s completed session closed it at $71,342.97 USD ($101,763.11 CAD), down 3.5% from $73,964.47 USD on Wednesday. As of a live bar stamped 08:45 ET, BTC was at $69,750.67 USD, and $99,517.37 CAD on a separate CAD quote stamped 08:40 ET. The last daily close under $70,000 USD was August 19, at $69,266.19 USD.
That break puts Bitcoin into a technical state with a testable history, so we tested it, and the result is not the one the setup advertises. The state resolves into a death cross almost every time. Our roll-forward of the two averages puts this one in the middle of December if price holds anywhere near here. And the death cross itself, across the nine in our own events file, has a far weaker bearish record than its name implies.
The state Bitcoin just entered, defined precisely
Since September 8, on our own computation from daily closes, Bitcoin’s 50-day average has been above its 200-day average, the configuration usually called a golden cross. With Thursday’s close and Friday’s break, price is now below both lines. A golden cross regime with price underneath it is a cross that has failed.
The averages we calculate from daily closes sit at $80,566.29 USD for the 50-day and $73,043.15 USD for the 200-day, a gap of 10.3%. Those are our arithmetic on closing prices, not levels read off a chart. Anyone wanting to check the structure against the chart a trader actually watches can open the TradingView BTCUSD symbol page.
The state nearly always ends in a death cross, and that is mostly arithmetic
Our screen finds 12 episodes of this state in the daily series since September 2014, the twelfth being the one that began on Thursday. The crossover events file covers ten of the eleven prior episodes, the 2015 one predating the file, and nine of those ten ended with the 50-day dropping through the 200-day. In each of the nine, the gap from entry to the cross equals the episode length exactly, because the cross is what ended the episode. They ran 26 to 53 sessions. The exception is September 20, 2021: eleven sessions, price back above both averages, no death cross for 116 days.
Nine of ten sounds like a finding. It mostly is not. Once price sits below a 50-day average that is well above the 200-day, the 50-day falls on nearly every session price stays underneath it, while the 200-day moves far more slowly. The cross is close to built into the arithmetic of the two lines, and the resolution rate describes how moving averages behave more than it describes Bitcoin.
What it does buy is a timetable. Rolling both averages forward and assuming price simply holds flat, our arithmetic puts the cross 34 days out, around December 17, 2026, if Bitcoin holds $69,750.67 USD. Hold that $68,052.44 USD low and it is 30 days, around December 13. Recover to $72,000 USD and it is 42 days, around December 25. Anywhere in the $68,000 to $72,000 USD band lands the cross 30 to 42 days from here, the middle of December. It takes a rally to roughly $75,000 USD to push it into late January 2027.
Holding flat is a device for seeing where the two lines are heading, not a prediction that price will do it.
What the last nine death crosses actually delivered
Those nine crosses are the entire death cross record in our events file, so the follow-up measures cleanly: the return from the day each cross printed, forward.
| Horizon | Median | Mean | Positive | Best | Worst |
|---|---|---|---|---|---|
| 30 days | -2.1% | +2.7% | 44% | +36.7% | -22.7% |
| 60 days | +8.4% | +12.5% | 67% | +43.8% | -21.3% |
| 90 days | +25.6% | +16.0% | 56% | +59.7% | -25.9% |
| 180 days | +33.8% | +30.7% | 56% | +167.2% | -53.1% |
The first month is mildly negative: a median of -2.1%, with fewer than half the cases positive. No median past it is negative. The median is +8.4% at 60 days, +25.6% at 90 and +33.8% at six months. A signal named the death cross carries its conclusion in the name, and across our nine the name is not supported beyond the first month.
The same control test belongs here as anywhere else, and it sharpens the point rather than softening it. Against Bitcoin’s unconditional base rate the death cross is worse at 30 days, 44% positive against 58%, and after that it stops separating from the market at all: 67% against 59% at 60 days, then 56% against 61% at 90 and 56% against 67% at 180. The means tell the same story from the other side, running +2.7%, +12.5%, +16.0% and +30.7%, below the medians at the two longest horizons. So the honest reading is not that a death cross has been good for Bitcoin. It is that past the first month it has been close to indistinguishable from Bitcoin doing nothing in particular.
That is not licence to read it the other way. Nine observations is a very small sample, and the dispersion inside it is enormous: the best six-month outcome, +167.2%, came from the September 12, 2023 cross, and the worst, -53.1%, from January 14, 2022. Positive rates of 56% at 90 and 180 days are close to a coin flip. The defensible statement is that the cross has not been a reliable bearish marker in our data, not that it has been a bullish one.
The return record from the state itself, against the right control
A signal only means something next to a fair control. Comparing a drawdown state against every session in the history proves little, because the state is a drawdown by construction. The comparison that isolates something is the same golden cross regime with price above both averages: the identical trend configuration, the healthy version of it.
| Horizon | Failed-cross median | Failed-cross mean | Healthy median | Healthy mean |
|---|---|---|---|---|
| 30 days | -0.4% | +6.5% | +3.9% | +10.6% |
| 60 days | -4.6% | +3.1% | +11.7% | +22.2% |
| 90 days | +9.0% | +15.4% | +19.4% | +33.0% |
| 180 days | +29.4% | +26.9% | +38.9% | +75.6% |
Positive rates run 45%, 45%, 73% and 64% for the failed-cross episodes against 57%, 65%, 67% and 73% for the healthy regime. The healthy samples hold 1,700 sessions at 30 days and 1,663 at 180, but those windows overlap heavily, so they are far fewer independent observations than the counts suggest.
Two things in that table cut against a tidy conclusion, and they point opposite ways. On the mean, the failed arm sits below the healthy arm at all four horizons, including the two where the medians look like they converge, which undercuts the reading that the difference washes out by 90 days. Against that, the 90-day positive rate of 73% is above the healthy regime’s 67% and above the 61% unconditional rate, so median and hit rate disagree at the same horizon. That argues for leaning on the 90-day result less, not for picking whichever number suits an argument.
The 30 and 60-day result is fragile in a way worth putting a number on. Five of the 11 episodes were positive at 30 days and five at 60, against base rates of 58% and 59%. Getting five or fewer out of 11 at a 59% base rate happens about 27% of the time by chance alone, and only three episodes were positive at both horizons. Nothing there clears a reasonable bar for significance.
One objection worth testing is whether the result is an artifact of Bitcoin’s earliest years. Measured as the gap in percentage points between the signal median and the healthy-regime median, the full history runs -4.3, -16.3, -10.4 and -9.5 at 30, 60, 90 and 180 days. Restricting both arms to 2018 onward, which leaves ten episodes, it runs -1.8, -10.4, +3.1 and +6.0. The gap survives in reduced form at the two short horizons. At 90 and 180 days it does not shrink, it reverses, with these episodes beating the healthy regime.
Every episode, including the three that did not come back
Medians hide the episodes that matter most to someone holding the asset. The screen runs over 4,441 daily closes, the Yahoo Finance series from September 17, 2014, and returns 11 completed episodes once entries less than 30 sessions apart are merged, so one long drawdown is not counted as several signals. All 11 are below, with entry prices in USD.
| Entry date | Entry price | Sessions | 30d | 60d | 90d | 180d |
|---|---|---|---|---|---|---|
| 2015-08-18 | $211 | 27 | +8.9% | +28.2% | +56.7% | +92.9% |
| 2018-02-05 | $6,955 | 53 | +43.3% | -4.6% | +38.8% | +1.1% |
| 2019-09-26 | $8,119 | 30 | +13.9% | -12.0% | -10.4% | -17.0% |
| 2020-02-27 | $8,784 | 27 | -28.9% | -11.3% | +4.5% | +29.4% |
| 2021-05-19 | $37,002 | 31 | -3.3% | -14.1% | +20.8% | +71.8% |
| 2021-09-20 | $42,844 | 11 | +54.0% | +35.7% | +9.0% | -1.5% |
| 2021-12-13 | $46,737 | 32 | -6.0% | -9.3% | -19.0% | -39.3% |
| 2023-08-17 | $26,665 | 26 | -0.4% | +7.0% | +42.1% | +86.5% |
| 2024-07-04 | $56,978 | 37 | +6.5% | +3.7% | +6.4% | +64.0% |
| 2025-03-09 | $80,601 | 29 | -5.4% | +28.1% | +31.0% | +37.3% |
| 2025-10-17 | $106,468 | 30 | -11.5% | -17.5% | -10.3% | -29.7% |
Three were negative at both 90 and 180 days, the criterion we use for an episode that began a real bear phase rather than a dip inside an uptrend: September 26, 2019, December 13, 2021 and October 17, 2025. The last was negative at every horizon and 29.7% lower after six months. The worst single 30-day outcome, -28.9% from the February 27, 2020 entry, is the COVID crash. The best, +54.0% from September 20, 2021, is the autumn 2021 run.
One limit belongs right beside those columns. The entry price in this screen is depressed by construction, because the state is defined by price sitting below both averages. Some part of the 90 and 180-day recovery is a low starting point rather than anything the screen knows.
Bitcoin’s all-time closing high was $124,752.53 USD on October 6, 2025. Friday morning’s live bar is 44.1% below that, and the $68,052.44 low is 45.5% below. The 2026 low to date was $57,747.77 USD on July 1, so a three-month low here is not a new bottom for the year, and from the $75,747.58 USD close on Saturday, November 7, the six-session decline is 7.9%. Where the line between a correction and a bear market sits matters for position sizing rather than labelling, which is what Correction vs Bear Market: How Markets Behave works through.
The Canadian crypto names did not move together
Thursday’s TSX session is worth separating out, because three of the Canadian ways to hold this exposure went in different directions. Purpose Bitcoin ETF (BTCC-B.TO) closed at $14.06 CAD, down 4.42%. HIVE Digital (HIVE.TO) closed at $2.35 CAD, down 6.37%. Hut 8 (HUT.TO) closed at $88.40 CAD, up 1.63%, and is up 9.7% across November 10 to 12, from $80.55 CAD.
The fund that holds the coin moved with the coin, and close to it: Bitcoin in Canadian dollars fell 3.4% on Thursday, from $105,391.52 CAD to $101,763.11 CAD, against the fund’s 4.42%. The two miners moved opposite ways on the same day. What the spread establishes is that “Canadian crypto stock” is not one trade, which is the premise behind ranking the names on what they actually hold in Canadian Crypto Stocks: Ranked on Bitcoin Per Share. It also decides what a December cross would mean for a Canadian holder: the fund carries the coin’s record and nothing else, while the miners carry whatever their other businesses are doing at the time. For what else sits inside one of those businesses, our November 10 report on HIVE Digital’s fiscal Q2 2027 results covered its AI revenue.
The backdrop a Canadian holder is sitting in
Bitcoin’s 3.5% Thursday decline came alongside gold and both equity indices. The S&P/TSX Composite closed at 36,137.20, down 0.75% from 36,411.70. The S&P 500 closed at 7,592.40, down 0.35% from 7,619.38. Gold’s continuous front-month series closed at $3,885.30 USD, down 2.88% from $4,000.50 USD. Ethereum closed Thursday at $2,016.83 USD ($2,876.79 CAD) and was under $2,000 USD on the 08:45 ET live bar at $1,975.56 USD, or $2,818.49 CAD on the 08:40 ET CAD quote.
Two sessions into this episode, the resolution is the part that is close to mechanical. Unless price climbs back above both averages, the 50-day keeps falling and the cross arrives, on our roll-forward around the middle of December if Bitcoin stays in the high $60,000s USD. What the cross has been worth is the part that rests on nine observations: a mildly negative first month, then a record that stops separating from Bitcoin’s ordinary behaviour, sitting alongside eleven observations of the state itself that run below a healthy golden cross regime on the mean at every horizon. The date is the thing to hold us to. If price stays near here, the cross prints in the middle of December, and the reader who remembers this piece will already know what the last nine delivered.
Disclaimer: The content on bestcanadianstocks.ca is for informational and entertainment purposes only and does not constitute financial advice. Past performance is not indicative of future results. Always consult a qualified financial advisor before making investment decisions. Bitcoin and Ethereum prices marked Friday are live bars from Yahoo Finance, stamped 8:45 a.m. ET for the USD quotes and 8:40 a.m. ET for the CAD quotes on November 13 2026, not closes, and the USD and CAD quotes are separate pairs taken five minutes apart. Crypto trades continuously. The November 12 2026 figures for Bitcoin, Ethereum, the S&P/TSX Composite, the S&P 500, gold’s continuous front-month series (GC=F) and the Toronto listings of Hut 8, HIVE Digital and the Purpose Bitcoin ETF are completed daily closes. Every percent change is our own arithmetic on those figures. The forward-return study is our own screen over Yahoo Finance daily closes from September 17 2014, defining the state as the 50-day average above the 200-day with price below both, merging episodes closer together than 30 sessions, and comparing it against the same golden-cross regime with price above both averages. Those control windows overlap one another and are not independent observations, and the screen’s entry price is depressed by construction. The same comparison was recomputed over sessions from January 1 2018 onward as a check that the result is not an artifact of Bitcoin’s earliest years. The death-cross record covers the nine crosses in the same computed series and is compared against the same unconditional base rate. The projected crossing date is our arithmetic rolling both averages forward on the assumption that price holds flat at the stated level, and is not a forecast. Moving averages and crossover dates are computed by us from daily closes and are not levels read from a charting platform.
