The Truth Behind Bitcoin’s “10 AM Sell-Off”: A Deep Dive Into Market Data

Understanding the Long-Standing Bitcoin Market Theory

For years, cryptocurrency traders and enthusiasts have debated a peculiar phenomenon in Bitcoin markets—the supposed “systematic selling at 10:00 AM” Eastern Time. This theory has circulated widely across trading forums, social media platforms, and within the crypto community, suggesting that Bitcoin regularly experiences price drops around this specific time each day. Many market participants have attributed these alleged declines to various actors, including institutional market makers, exchange-traded fund (ETF) flows, or even coordinated manipulation efforts. However, Vetle Lunde, the Head of Research at K33 Research, a respected cryptocurrency research firm, recently decided to put this popular theory to the test with hard data and rigorous analysis. His comprehensive study examined an enormous dataset spanning over 606,000 minutes of Bitcoin trading activity between January 1, 2025, and February 26, 2026, aiming to determine whether there’s any statistical validity to these long-standing claims. The findings challenge many of the assumptions that traders have held for years and offer a more nuanced understanding of Bitcoin’s intraday price movements.

What the Data Actually Reveals About 10 AM Trading

The results of Lunde’s analysis paint a picture that directly contradicts the popular narrative. When examining the full dataset covering the entire period, Bitcoin’s performance at 10:00 AM Eastern Time was actually quite strong compared to other times of day. Specifically, the average return for Bitcoin during any given minute throughout this period was approximately -0.003 basis points—essentially flat with a slight negative bias. In stark contrast, the average return at the 10:00 AM minute was calculated at 0.207 basis points, a decidedly positive figure. To put this in perspective, when ranking all 1,440 minutes in a trading day by performance, the 10:00 AM minute ranked as the 359th strongest. This means that roughly 75% of all other minutes during the day performed weaker than 10:00 AM. This finding fundamentally challenges the notion that Bitcoin systematically drops at this time. If anything, the broader dataset suggests that 10:00 AM has historically been a slightly better-than-average time for Bitcoin performance, placing it in the upper quartile of intraday timeframes. Lunde emphasized that this extensive dataset “quickly refutes” the widely held belief that there’s a systematic price drop occurring at this specific hour. The sheer volume of data points analyzed—over half a million minutes—provides statistical confidence that the findings aren’t merely the result of random chance or limited sampling.

The Recent Exception That Fueled the Theory

While the comprehensive long-term data doesn’t support the 10 AM sell-off theory, Lunde’s analysis did uncover something interesting when examining a more recent and narrower timeframe. Between November 1, 2025, and February 26, 2026—a period of approximately four months—the pattern looked quite different. During this shorter window, the average return at 10:00 AM minutes dropped significantly to -1.41 basis points, a notable decline. When ranked against other minutes during this period, 10:00 AM performed as the 35th worst minute of the day, placing it within the weakest 2.36% range of all intraday minutes. Additionally, negative returns were observed at the 10:00 AM minute on 53.85% of the days examined during this period, meaning that more than half the time, Bitcoin was indeed declining at this specific moment. However, Lunde pointed out an important caveat: 12% of all minutes throughout the day had an even higher frequency of negative returns than 10:00 AM. This suggests that while 10:00 AM might appear as a “negative outlier” when looking at this specific short-term sample, it’s not sufficiently exceptional to directly indicate market manipulation or systematic selling. This recent pattern likely explains why the theory gained such traction among traders—they were experiencing a genuine short-term phenomenon but extrapolating it as a permanent market feature without examining the longer-term historical context.

The Bigger Picture: Context Matters

One of the most important insights from Lunde’s research is that focusing on a single minute in isolation can be misleading. When he expanded the analysis to examine the time windows surrounding 10:00 AM, the picture became much more balanced and less alarming. Looking at the ten-minute window from 09:55 to 10:05 AM, the average return was -0.0033%—barely negative and essentially flat. Expanding further to the twenty-minute window from 09:50 to 10:10 AM, the average return actually turned positive at +0.0017%. Even more tellingly, the ten-minute period from 10:00 to 10:10 AM showed an average return of +0.0038%, a clearly positive figure. When examining even wider intervals, such as the full hour from 09:30 to 10:30 AM or from 10:00 to 11:00 AM, any losses remained quite limited and not statistically significant. This contextual analysis reveals an important truth about market microstructure: individual minutes can show volatility and negative returns without indicating any systematic pattern of manipulation. The fact that the periods immediately surrounding 10:00 AM show neutral to positive returns suggests that any weakness at that exact minute is likely absorbed and reversed within minutes, indicating normal market volatility rather than coordinated selling pressure.

The Randomness of Weak Performance Times

Another crucial finding that undermines the manipulation theory is the apparent randomness of poorly performing minutes throughout the trading day. Lunde specifically examined which minutes performed worst during the November 2025 to February 2026 period and discovered that they didn’t cluster around “round” times like 10:00 AM, which would be expected if manipulation were occurring at psychologically significant moments. Instead, the worst-performing minutes during this period included seemingly random times such as 10:12 AM, 09:41 AM, 17:13 PM (5:13 PM), 10:46 AM, and 14:02 PM (2:02 PM). These irregular times lack the psychological significance that round numbers possess, suggesting that their poor performance results from general market volatility rather than coordinated actions by specific market participants. If market makers or large institutional players were systematically manipulating Bitcoin prices, they would likely choose more recognizable times—round hours or half-hours that are easier to coordinate and program into trading algorithms. The scattered nature of weak performance across various random minutes throughout the day points instead to the natural ebb and flow of trading activity, news events, and the chaotic dynamics of a 24/7 global market responding to information from multiple time zones.

The Real Driver: US Market Hours and Volatility

Perhaps the most significant finding from Lunde’s research is the clear relationship between Bitcoin volatility and US trading hours, particularly the opening of traditional stock markets. The analysis revealed that Bitcoin volatility increases dramatically when US markets are open, with particular intensity in the minutes surrounding major events like macroeconomic data releases and the stock market opening bell. The period from 09:31 to 09:37 AM Eastern Time—the first seven minutes immediately following the opening of the US stock market—represents the absolute peak of Bitcoin volatility during the day. This finding has important implications for understanding Bitcoin’s market microstructure. Despite being a decentralized, global cryptocurrency that trades 24/7 across markets worldwide, Bitcoin’s price action remains closely tied to US financial markets and the traditional trading day. This connection likely stems from several factors: the concentration of institutional Bitcoin trading and ETF activity during US market hours, the release of US economic data that affects risk sentiment globally, and the simple fact that the largest pools of capital and liquidity are often active during the US trading session. The increased volatility during these hours doesn’t indicate manipulation but rather reflects the natural concentration of trading activity, information flow, and decision-making by major market participants. In his conclusion, Lunde was clear and definitive: the comprehensive data analysis did not provide clear evidence of intentional and systematic selling at 10:00 AM or any other specific time. While short-term patterns can emerge and create the appearance of systematic behavior, these are better explained by the natural volatility associated with US trading hours and the randomness inherent in financial markets rather than coordinated manipulation efforts. For traders and investors, this research offers an important lesson about the dangers of pattern recognition bias—our human tendency to see intentional patterns in what may actually be random or cyclical market behavior.