Bitcoin closed at $75,747 on Friday, 3.5% above its 200-day simple moving average of $73,191. It is already below its 50-day ($81,975) and 39.3% under its $124,753 record close from October 2025. If the slide continues, the next headline writes itself: "Bitcoin loses the 200-day, bear market confirmed." Every crypto market wrap treats that line as the border between bull and bear, and the advice that rides with it is always the same: get out below it, get back in above it.
We tested the line against every daily close since the 200-day average first existed for BTC in April 2015. The short version: losing the 200-day is a weak sell signal and a decent risk-control tool, and those are not the same thing. After an independent break below the line, BTC's median 30-day return was -0.25%, against +3.13% on a random day. But 90 days later it was still up 72.7% of the time, and a strategy that only held BTC above its 200-day finished with a lower annual return than buy-and-hold (59.6% vs 63.5%). What it bought you was a smaller worst-case loss: -69.7% instead of -83.4%.
What we measured
Data: yfinance BTC-USD daily OHLC, from April 2015 (first valid 200-day average) through this week, 4,235 days. Bitcoin trades every day, so 30, 90 and 180 days are calendar days.
- Break below: BTC closes under its 200-day SMA after closing at or above it the day before.
- Entry: the next day's open. The signal day's close is not tradable once you know it.
- Baseline: every day in the same window, also entered at the next open.
- Overlap: breaks cluster, because price chops around the line. We count an independent episode only when there has been no other break below in the prior 90 days. That turns 37 raw breaks into 11 episodes. Eleven is a small sample, so treat every forward-return number below as suggestive, not proven.
- Trend filter: hold BTC while the prior close is at or above the 200-day, sit in cash otherwise, switch at the next open, 0.1% cost per side.
What the data says
Forward BTC return after a break below the 200-day (11 independent episodes) vs every day:
| Horizon | Break: median | Baseline: median | Break: mean | Baseline: mean | Break: up | Baseline: up |
|---|---|---|---|---|---|---|
| 30 days | -0.25% | +3.13% | +2.81% | +6.46% | 45.5% | 57.9% |
| 90 days | +6.33% | +10.41% | +14.17% | +22.5% | 72.7% | 60.8% |
| 180 days | +29.33% | +30.9% | +25.26% | +53.12% | 54.5% | 66.6% |
Three things stand out.
The first month is the weak spot. The 30-day window is the only one where a break below the line shows a clear tilt: median negative, hit rate under half, versus a positive median and 57.9% up on a normal day. That is what the "get out" crowd is pointing at, and it is real, just small.
After 90 days the break mostly stops mattering. At 90 days the median after a break (+6.33%) is lower than baseline (+10.41%), but BTC was higher 72.7% of the time, more often than on a random day. At 180 days the medians are nearly identical: +29.33% vs +30.9%. The means are lower after a break mostly because the baseline includes BTC's parabolic years, which lift its average far above its median.
Most breaks are fake. Of the 37 raw closes below the line, 22 (59.5%) closed back above it within 10 days. If you sell every first close below the 200-day, you will mostly be selling into chop and buying back higher.
The reclaim is the stronger signal. Breaks above the line (17 independent episodes) had a 90-day median of +18.39% and a 180-day median of +41.02%, up 68.8% of the time at 180 days. Above the line is where BTC has done its compounding.
The trend filter: less return, smaller hole
| April 2015 to now | Hold above 200-day | Buy and hold |
|---|---|---|
| Annual return (CAGR) | 59.6% | 63.5% |
| Max drawdown | -69.7% | -83.4% |
| Annual volatility | 51.8% | 64.9% |
| Sharpe (no cash yield) | 1.16 | 1.08 |
| $1,000 became | $225,544 | $298,643 |
The filter was in the market 61.2% of the time and made 37 round trips. It gave up a few points of annual return and roughly a quarter of the ending balance. In exchange, volatility fell by about a fifth and the worst peak-to-trough loss went from -83.4% to -69.7%. Risk-adjusted, it is a slight win (Sharpe 1.16 vs 1.08). In dollars, it is a loss.
Note that -69.7% is still a catastrophic drawdown. The 200-day is slow: in a parabolic run like late 2017, price sits so far above the average that by the time it crosses, most of the fall has already happened.
It depends heavily on the era:
| Period | Filter CAGR | B&H CAGR | Filter max DD | B&H max DD |
|---|---|---|---|---|
| 2015-2017 | 311.2% | 332.3% | -35.4% | -35.4% |
| 2018-2021 | 9.9% | 34.6% | -66.3% | -81.5% |
| 2022-2026 | 27.1% | 10.8% | -32% | -66.9% |
From 2018 to 2021 the filter was a disaster: whipsawed through the 2018-2019 chop and the March 2020 crash, it returned 9.9% a year against 34.6% for holding. From 2022 to 2026 it was the opposite: it sidestepped most of the 2022 bear market and the 2025-2026 slide, more than doubling buy-and-hold's annual return with half the drawdown. Anyone quoting the 200-day rule from only one of those windows will give you the wrong answer with total confidence.
And 200 is not special. We ran the same filter on 50, 100 and 150-day averages:
| SMA length | Filter CAGR | B&H CAGR (same start) | Filter max DD | Round trips |
|---|---|---|---|---|
| 50 | 70.4% | 57.3% | -58.7% | 112 |
| 100 | 61.6% | 58.5% | -62.1% | 74 |
| 150 | 74.9% | 64.5% | -61.5% | 38 |
| 200 | 59.6% | 63.5% | -69.7% | 37 |
The 200-day was the worst of the four on return and on drawdown. The faster averages did better, but the result is not smooth (150 beats 100, which loses to 50), which is what you see when a parameter is fitting noise. The defensible conclusion is that some trend filter has cut BTC's drawdowns. It is not that 200 is the magic number.
The actionable rule
A first close below the 200-day is not a sell signal on its own: 59.5% of them reversed within 10 days. Treat it as a warning for the next month, when returns after a break have run below baseline, not as the start of a bear market. If you use a trend filter on BTC, use it to cap the drawdown, not to add return, and expect to give up some upside in the years BTC rips. Watch the reclaim more closely than the break. Closes back above the line have been followed by stronger 90 and 180-day returns than breaks below.
Where it fails
Five of the 11 independent breaks were followed by a 90-day return above the baseline median, and all five of those rallied more than 20%. Selling them was the wrong call:
- August 2015 at $211: +46.6% over 90 days, +80.5% over 180.
- August 2023 at $26,665: +42.2% over 90 days, +86.7% over 180. This was the start of the run to the 2024-2025 highs.
- March 2025 at $80,601: +31% over 90 days.
Even when the break was right, it could be late or painful. The March 2020 break (late February, $8,784) was followed by a -53.3% drawdown inside 180 days, the worst in the sample, and then a +29.3% gain by day 180. The break that mattered most, mid-December 2021 at $46,737, led to -19% over 90 days and -39.3% over 180. The most recent one, mid-October 2025 at $106,468, led to -10.3% over 90 days and -29.7% over 180. BTC got back above the line in August, and is now sitting just above it again.
With only 11 independent episodes, one or two new breaks can move the medians a lot. Bitcoin's history is also short compared with stock indexes, and the early years were dominated by a tiny, illiquid market. The 2022-2026 window, when the filter did best, is also the period when BTC traded most like a large risk asset.
How to watch it
The 200-day tells you about BTC's own trend. It does not tell you what the stock market is doing to risk appetite, and since 2020 that has mattered for crypto. Market Health tracks the equity regime (breadth, VIX and trend) that has been moving BTC more and more.