Uptober and its seasonal probability

Let’s cut through the Uptober hype with some cold, hard data. You’ve got a chart of Bitcoin monthly returns and you’re eyeing October. From 2013 to 2024, October closed green 10/12 times (only 2014 and 2018 were red). If this month Oct-2025 closes red, that’s 3/13 about 23% odds of a down month. Nice edge, sure, but not destiny.

The seasonal probability 70/30:

In traditional markets, many calendar effects win around 60-70% of the time over long samples, good, but far from guaranteed. Two well studied examples: the Halloween indicator (“Sell in May”) shows higher Nov-Apr returns across many markets, but not every year, updated research still finds persistence, not certainty. Likewise, the turn of the month effect concentrates a big chunk of average monthly return into a few days, yet it also fails regularly. In short, seasonality is probabilistic, think 70/30, not 100/0.

Reality check:

markets don’t care about calendars, they care about cycles, liquidity, and macro. If we’re in a prolonged consolidation (post-halving churn, shifting policy), that ~23% downside month isn’t some anomaly, it’s the cost of playing a probabilistic edge. The log trend can stay intact while October still prints red.

Bottom line:

Use seasonality as context, not a trigger. Based on my risk metric, BTC $109K = risk 48. It's still under 50 and today is Sunday. I'm DCAing in, sticking to my plan.

Source:

• Halloween indicator (“Sell in May”) https://www.aeaweb.org/articles?id=10.1257%2F000282802762024683
• turn of the month effect: https://www.sciencedirect.com/science/article/abs/pii/S1057521903000073