The short version: Bitcoin has about sixteen years of history, which means roughly sixteen observations of any given month. That is not enough to establish a seasonal pattern, and most of the ones you'll be shown this quarter are the shape of noise rather than the shape of a market.

Tuga appears at the back door at six o'clock. Every day, within a few minutes, with an air of having consulted something. For a long time I assumed she had an internal clock of unnerving precision.

She doesn't. She has cues: the light changing, a particular car arriving in the street, sounds from the kitchen, the fact that someone just sat down. Shift the routine — a day off, dinner late, a different car — and she's wrong, and entirely unbothered about being wrong, because she was never reading a clock. She was reading a mechanism that usually coincides with one.

It's November. For the next two months you will be shown a great many charts about which months are good for Bitcoin. This is the article about why most of them are Tuga at the back door.

1. The sample size problem, said plainly

Bitcoin has traded for roughly sixteen years, and the first few of those were a market so small and strange that including them tells you about a curiosity, not an asset.

So when someone says "October is historically strong" or "September is weak", the underlying evidence is about a dozen observations. Twelve numbers. For an asset whose monthly moves routinely range from minus thirty percent to plus forty.

With that much variance and that few data points, a couple of memorable years dominate the average completely. One extraordinary month — and we've had several, for reasons specific to those months — moves the "seasonal" figure for that month permanently.

This is not a subtle statistical objection. It's the whole thing. You cannot detect a modest recurring effect inside enormous noise using twelve samples. Nobody can. The honest confidence interval around any of these monthly averages is so wide it includes "this month is bad" and "this month is great" simultaneously.

That's why I'm not quoting you the monthly averages in this article. Printing them would lend them an authority the sample size doesn't support, and you'd remember the numbers and forget the caveat. Everybody does.

2. Why the patterns look so convincing anyway

Twelve months means twelve chances. Test twelve hypotheses against noisy data and some will look significant by luck alone. That's arithmetic, not cynicism. The month that happens to have won most often gets a nickname, and the nickname then does the persuading.

The nickname is the trick. "Uptober" is not a finding. It's a rhyme that makes a coincidence memorable and therefore repeatable. Once a pattern has a name it gets cited rather than re-tested.

Nobody publishes the failures. You'll see the chart in the years the pattern held. In the years it didn't, the chart quietly isn't posted, and the people who posted it last year are talking about something else.

And it's deeply convenient. A seasonal thesis gives you something to say, a reason to act, and a story that's exciting in the way "nobody knows" is not. Content needs patterns. The market does not supply enough of them, so the shortfall gets made up.

3. The honest version: mechanisms, not calendars

Here's where I'd push back on the pure sceptic position, because "all seasonality is nonsense" is its own overconfident claim.

Real calendar effects do exist in real markets, and the ones that survive scrutiny have a mechanism you can name:

• Tax-year selling. In jurisdictions with a fixed tax year, people realise losses before a specific date. That's a real flow with a real date.
• Fiscal and quarter ends. Institutions rebalance, close books, square positions. The dates are fixed and the behaviour is mandated.
• Reduced liquidity in holiday periods. Fewer participants means thinner books, which means moves travel further. A genuine, explicable effect on volatility — not on direction.

Notice what these have in common: you can explain why on a Tuesday, in one sentence, without referring to a chart. The mechanism came first and the calendar effect follows from it.

That's the test. Ask what the mechanism is. If the answer is "that's just what the data shows", you're being shown a coincidence with a nickname. If someone can point to an actual recurring flow with an actual date, that's worth a look — and worth checking whether the effect persists now that everyone knows about it, because publicised edges tend to get arbitraged into nothing.

4. What to do about it in practice

Nothing, mostly. If you're buying on a schedule, a seasonal argument is an invitation to break the schedule, and breaking the schedule is where the behaviour gap lives. The whole value of a mechanical plan is that it doesn't negotiate with narratives.

If a seasonal claim tempts you, price the mistake first. What does it cost if the pattern doesn't hold this year? If the answer is "I'd be buying later at a higher price" or "I'd be out during the move", that's the real risk of the trade, and it should be weighed against twelve data points.

Be especially careful in Q4. The end of the year generates more of this content than any other period, because it's when "what happens next" has the biggest audience and the least available evidence.

5. The uncomfortable half

I've argued from statistics, and statistics has limits too. "Sixteen observations is too few" is correct, but it means we cannot tell, not the effect is absent. A real seasonal effect could exist and be undetectable with this much data. My claim is only that you shouldn't act on something you can't distinguish from chance — not that I've disproved it. Those get conflated constantly, usually by people on my side of the argument.

Debunking is also a content format. Writing "that pattern is noise" is easy, feels rigorous, and costs nothing to be wrong about, because when the pattern happens to hold nobody comes back to check on the sceptic either. I benefit from the same asymmetry I've just criticised, and you should discount me for it accordingly.

And scepticism is not a strategy. Knowing that seasonality is unreliable doesn't tell you what to do; it only removes one bad reason for doing something. There's a type of person who becomes expert at dismantling everyone else's framework and never builds one, and they are not, in my observation, doing better than the person with a boring monthly buy and no opinions at all.