All Eyes on BTC’s 50-Week Moving Average

This alert was originally sent directly to clients of Galaxy Trading and Galaxy Asset Management on Aug. 27, 2026. Trade or invest with Galaxy to receive the most timely research directly in your inbox.

Introduction

Historically, two moving averages have framed every bitcoin bear market – often, the 200-week moving average (MA) has set a floor, and the 50-week moving average has formed the ceiling. BTCUSD has only printed below the 200-week moving average on 56 of the 642 weekly closes since the MA came into existence. The metric has been a key feature of each bear market low: in 2015 and 2018, the bear market bottom was within 9% of the 200w MA. Same for the local bottom (and perhaps the cycle’s) in June 2026. It’s not a perfect bottom indicator, or perfect long-term support, but it’s pretty close.

The 50-week moving average has similarly served as ceiling resistance in bear markets. In four of the five completed bear markets, once the 50-week moving average was first broken to the upside, the bear market bottom was definitively “in.” (The only exception is during the mini-bear market in between April and November 2021 all-time highs, a period that most Bitcoin analysts don’t even consider to be a “full bear market.”) Essentially, retaking the 50w MA has previously confirmed the end of a bear market.

This paper examines the history of the 200w bear floor and 50w bear ceiling.

Today, the 50-week moving average sits at $81,796. Last week, BTC rose 23.5% on the week and closed at $77,593, but this week, BTCUSD traded as high as $81,265, mere spitting distance from the 50-week. On the Aug. 30 weekly close, the 50w MA will likely be $81,110, and a weekly close above that level would be a strong indicator that the bear market bottom is “in” and the uptrend can resume.

Exhibit 1. Bitcoin's daily close against its 50-week and 200-week moving averages, on a logarithmic scale, with the seven qualifying bear episodes shaded and the corridor between the two averages filled. Since 2014, the floor has met three bear-market lows within 9% (2015, 2018, 2026), sat far beneath one (2021) and been broken by one (2022); every completed bear that lost the ceiling ended with a weekly close back through it.

What Counts as a Bear Market?

Every stretch between an all-time high and the next is a drawdown. If we define bear markets as drawdowns of 50% or more that lasted more than 90 days, seven periods qualify: 2011, 2013, 2013-2015, 2017-2018, the 2021 mid-cycle drawdown, 2021-2022, and the current drawdown that began after the Oct. 6, 2025, all-time high. These periods are identified in the shaded areas of the chart above and detailed in the table below.

Table 1: The seven bear market episodes. *Current episode is provisional. Its low and duration are measured as of Aug. 25, 2026. Dates and prices are UTC closes.

Exhibit 1.1. The same series as Exhibit 1 on a linear price axis. The log scale of Exhibit 1 makes every cycle comparable; the linear scale shows the present corridor at true proportions – a 50-week ceiling falling toward a 200-week floor that has never turned downward.

The 200-Week Moving Average Sets the Floor

The 200w MA is a bit unusual because it has only ever risen. Across all 642 weeks since it came into existence, the 200w MA has never printed a weekly decline. Bear markets have typically rested on the 200w MA as support, with one major failure in 2021-2022.

• 2013–15. The Jan. 14, 2015, low at $176 came in 8.6% below the average as it stood that day ($192), but bitcoin closed the week 9.8% above it. The average was not lost on a weekly basis until the August–September 2015 retest, when five weekly closes printed beneath it by at most 1.4%, in runs of no more than three weeks. From there, the recovery began.

• 2017–18. The Dec. 15, 2018, low closed at $3,185.07. The 200-week average stood at $3,185.93 that day. Across the entire period, including the COVID crash, the weekly close fell below the average exactly once (March 15, 2020, 2.8% below); on a daily basis, the March 2020 panic reached 10% below it, and the daily close spent seven days beneath it in the whole episode, the low included.

• 2021. The mid-cycle bear bottomed +113% above the average. The floor was never in play.

• 2021–22. The weekly close first broke the average on June 19, 2022, recovered it after six weeks, then lost it for 30 consecutive weeks from August 2022 to March 2023. The Nov. 9, 2022, low sat 34.2% below the average; 47 weekly closes and 322 daily closes in the episode printed beneath it. Even here the average was a meaningful level: the June 2023 and August–October 2023 dips beneath it, 6.7% deep at worst, were both recovered before the run to new highs.

In summary, 56 of 642 weekly closes have been below the 200-week moving average since 2014, and 36 of those weekly closes happened from August 2022 to March 2023. Excluding that period, the 200w MA has been undercut on only 20 weekly closes in 12 years, and never by more than 6.7%. This track record of the 200w MA acting as long-term support is what justifies our use of the term “floor” in this report: not that the level is never pierced, but one that has been pierced briefly and shallowly except in one case.

Table 2: The 200-week floor across all seven bear markets. * Current episode is provisional. Daily closes (the low, the deepest daily close) are compared with the 200-week average as a weekly chart would have showed it that day: the previous 199 completed weekly closes plus that day's close. Weekly closes are compared with the completed average of their own week. Counts cover the whole episode, including the recovery after the low.

Exhibit 2. The daily close as a multiple of the 200-week moving average since the average came into existence in May 2014, taken as a weekly chart would show it each day. The floor is the 1.0x line; red fill marks the time spent beneath it; the deepest dips are labelled. The 2022-23 break is the only sustained one.

Exhibit 3. How far below the floor each bear low printed (left bars), beside the deepest weekly close of the whole episode (right bars) and the number of weekly closes beneath the average. The 2018 low missed the average by 0.03%; 2022 is the outlier at 34.2%.

The 50-Week Moving Average Is the Ceiling

Historically, the 50-week MA has played the opposite role. It is lost early in bear market drawdowns: in the five completed bear markets that lost it, the first weekly close below the 50w MA came five to 36 weeks after the all-time high. From that week until the week before the low, it functioned as a ceiling. Across the five bears (106 weeks in total) the weekly close finished above the 50w MA only three times, and all three of those were in the mini-bear between the 2021 “Twin Peaks.” Rallies that did not clear the 50w MA got close: in 2018, the first bounce after the loss stopped 0.6% below the 50w MA; the July 2018 rally reached 1.9% below the 50w MA; and in 2014, the best attempt came in 6.6% below the 50w MA. The 2013 bear was an exception: the crash from $231 to $66.34 ran its course with the low still 45% above the 50w MA.

But the reclaiming of the 50w MA has been instructive for when bear markets end. During the six completed bear markets, the weekly close crossed back above the 50w MA 13 times. These crossings can be categorized into three groups:

1. Five were the first crossing after a bear-market low (Jan. 8, 2012; Oct. 25, 2015; May 5, 2019; July 25, 2021; and March 19, 2023) and none of them was followed by a lower low.

2. Six were later re-crossings that occurred inside recoveries whose low an earlier 50w MA reclaim had already confirmed (133% to 204% above bear market lows that were already in).

3. And two upward crossings occurred before the low was in and ultimately failed to hold (Dec. 26, 2021 and March 27, 2022, 48 and 139 days, respectively, after the November 2021 all-time high, with the price only 25% and 31% below the ATH). Each of these held for a week or two.

Ultimately, in four of the five bears that lost the 50w MA, the first weekly close back above it was the real one. The final reclaiming of each episode held for 21 to 139 weeks and the prior all-time high was regained 86 to 575 days after the first post-low 50w MA retaking.

Table 3: The 50-week ceiling across seven bear market episodes. * Current episode: the 50-week average was lost in the week ended Nov. 16, 2025, and has not been reclaimed. "Weeks loss to low" counts weekly closes from the week the average was lost through the last week that ended before the low. "Closest rally" is the highest weekly close relative to the average among the weeks in that span that finished below it, excluding the week the average was first lost (2021's span is that week alone). "Days to prior high" counts from the first post-low reclaim to the day the previous all-time-high close was regained. In 2011 the average was only months old and sat far beneath the crash.

Exhibit 4. The daily close as a multiple of the 50-week moving average since June 2011. Shaded spans are the bear episodes; purple fill marks time beneath the average; diamonds mark the weekly closes that crossed back above it - green when the low was already in, hollow red when a lower low followed.

Exhibit 5. Weeks spent under the ceiling between losing it and the low, with the number of weekly closes that broke above it in that span. Three of 106 weeks in the completed bears; 0 of 33 so far in the current one.

Exhibit 6. The price of confirmation: every weekly close back above the 50-week average inside a completed bear, placed by days from the low and by the gain from the low to that close. The first reclaim after the low came 5-284 days after it; the only two failures happened before the low, in the 2021-22 bear.

The Three Major Bear Markets

Charts of BTCUSD during the 2013-2015, 2017-2018, and 2021-2022 bear markets show both averages doing their jobs. In 2014–15 (Exhibit 7 below), the 50-week average was lost slowly, 36 weeks after the peak, then held every rally for 22 weeks while the 200-week average rose to meet the January 2015 low; the reclaim in October 2015 came 284 days later and held for 135 weeks. In 2018 (Exhibit 8) the ceiling was lost in May and rejected two rallies within 1.9% of it before the December low landed on the floor; the May 2019 reclaim then confirmed the low, and the one weekly close beneath the floor, in March 2020, was recovered the following week. In 2021–22 (Exhibit 9), the ceiling was lost five weeks after the peak, punctured twice (three weekly closes above it) in the months that followed, and then held for the rest of the decline (while the floor gave way in June 2022 for six weeks and, after a 3-week recovery, stayed broken for 30 weeks). The March 2023 reclaim, 130 days after the low, held for 139 weeks.

Exhibit 7. The 2013-15 bear. The 50-week average was lost in August 2014 and capped every rally for 22 weeks; the January 2015 low printed 8.6% below the 200-week average on the day but the week closed above it; the August-September 2015 retest dipped five weekly closes beneath the floor by at most 1.4%. The October 2015 reclaim held 135 weeks.

Exhibit 8. The 2017-18 bear and its recovery. The ceiling was lost in May 2018 and rejected rallies within 0.6% and 1.9% of it; the December 2018 low closed $3,185.07 against a 200-week average of $3,185.93 that day. The March 2020 crash produced the episode's only weekly close beneath the floor; the May 2019 and May 2020 reclaims bracket the whipsaws around it.

Exhibit 9. The 2021 mid-cycle bear and the 2021-22 bear. The only two failed reclaims on record (hollow red) came in the first months after the November 2021 top; the floor was broken for 30 straight weeks from August 2022; the March 2023 reclaim, 130 days after the low, held for 139 weeks. The 2023 dips beneath the floor were retests from above.

The Corridor Between the 200w and 50w Moving Averages

Because the ceiling falls through a bear market (as the lower prices pull the average down) while the floor keeps rising (because 200w MA has never turned down), the corridor between the two averages narrows as the bear market drawdown ages (Exhibit 10 below). In the weeks of the 2015, 2018, and 2021 lows, the 50w MA stood 2.48, 2.4, and 2.31 times the 200-week average. In the week of the November 2022 low, it stood 1.31 times and the ceiling went on to fall below the floor for 47 weeks (from the week ended Feb. 19, 2023, to Jan. 7, 2024, bottoming at 0.84). That is the only inversion of the two averages on record. In the week of this year’s June 30 low, the price 50w/200w was 1.41x, narrower than at any bear low for which both averages exist (except November 2022’s inversion). The ratio has since narrowed further to 1.27.

The corridor can be narrow at the bear market low for two primary reasons: 1) a bull market that never ran far ahead of its long-term trend, or 2) a long bear market that has dragged the 50w MA down substantially. In 2022 the second scenario produced the narrowest corridor at any bear low on record following a 77% drawdown.

The current bear market sets itself apart because of the starting point. At the Oct. 6, 2025, all-time high, the 50w MA stood 1.88x the 200w MA and the price 2.3x the 200w MA, compared to 2.9x and 15.6x at the December 2017 peak, 2.1 and 5.5x in April 2021, and 2.6 and 3.9x in November 2021. This bull market ran far less ahead of the trend than any prior bull market, so its bear had less distance to fall. The 53.1% drawdown is just barely deeper than the 2021 mid-cycle bear’s drawdown (53.11% vs. 53.08% in 2021). These two drawdowns are the shallowest of the seven we track in this report (of course, the current drawdown % is provisional).

Exhibit 10. The 50-week average as a multiple of the 200-week, by week, with each bear-market low marked. The corridor was 2.31-2.48x wide at the 2015, 2018 and 2021 lows, 1.31x in 2022 and 1.41x at this June's low; it inverted once, from February 2023 to January 2024.

The Current Market

The current bear market drawdown has so far behaved similarly to the 2015 and 2018 bear markets at the floor, and like every bear market but one at the ceiling (Exhibit 11 below). The 50w MA was lost in the week of Nov. 16, 2025, five weeks after the October all-time high, when BTCUSD was trading 25% below the ATH. In the 33 weeks from that November 2025 cross to the June bear market low, no weekly close finished above the 50w MA, and the five rallies that came within 14% all stalled. The Jan. 18 rally made it within 7% of the 50w MA. At the time of writing on Aug. 27, BTCUSD is trading around $80k, just a mere 2.2% below the 50w MA of $81.8k).

Last week (which ended Aug. 23, 2026) closed at $77,593, up 23.5% on the week and just 5% below the 50w MA of $81,796. This was the closest that any weekly close has come to the ceiling since it was lost in November 2025. And the average is also falling toward the price at roughly $987 a week on average over the last two months. (The weeks now leaving the moving average’s window are the $108,707-$123,524 closes of September to November). For this week, which will end on Sunday, Aug. 30, 2026, a close above $81.1k would put the week above its own 50w MA for the first time since the week ended Nov. 9, 2025.

Exhibit 11. The current episode. The floor held at the June 30 low; the ceiling has capped every weekly close for 41 weeks and stood at $81,796 on last week's close, with the weekly closing price 5.1% beneath it. The dashed level, $81,110, is the Sunday close that would put the week ending Aug. 30 above its own 50-week average.

Crossing above the 50w MA has been a very strong indicator historically that a bear market is over and a bull market is beginning, though that claim is not without conditions. Of the 13 weekly crossings back above the 50w MA in completed bear markets, only two were followed by a lower low. And in four of the five bear markets that lost the average, the first 50w MA upwards crossing was the real one (the one that correctly confirmed the bottom was “in”). The five first reclaimings that held occurred 103 to 690 days after the prior all-time high, at levels 45% to 76% below the prior all-time high. A reclaiming at $81,110 on the Aug. 30 weekly close would come 328 days after the all-time high.

In this bear market, the floor has held, and a weekly close above the 50w MA ceiling could be a very strong indicator that the bottom is in.