Takeaways

#
  • September was the only month with a negative average return in the authors’ Bitcoin sample. The low point on the average September path fell on the 26th.
  • Bitcoin trades every calendar day, so the researchers had to consider weekends, holidays and the time zone used to define a daily close.
  • The study separated 2011–2017 from 2018–2022 around the launch of CME futures to examine the influence of Bitcoin’s early years.
  • Hirsch was considering a small entry and a possible momentum signal. The late-September seasonal pattern gave him a window to watch.

Stock Trader’s Almanac editor Jeff Hirsch had a reason for publishing his first Bitcoin seasonality study on September 26, 2023. On the average path he had built with analyst Adrian Zduńczyk, that was the September low. When Spencer Israel asked him about it two days later on the Morning Show, Hirsch was still looking for an entry.

There was an apparent contradiction in calling the worst month a buying opportunity. September had produced an average loss of 6.45% in the authors’ sample, the only negative monthly average. But October and November looked much stronger. Hirsch’s interest was in what might happen coming out of September, rather than in owning Bitcoin through the decline that gave the month its poor record. The authors published the monthly and daily charts in their original report.

What counts as a September trading day?

Hirsch was accustomed to studying equity markets that close on weekends and holidays. For Bitcoin (BTC), September has 30 trading days every year. Put a Saturday at the beginning of the month and it still belongs in the chart. Even the boundary between two days requires a choice: midnight in New York and midnight in London split the same trading activity differently.

The intramonth chart compared three lines: the average path across all months, the path for September alone, and September 2023 through the latest observation available when they prepared it. Only the second line supplied the September 26 low.

Hirsch had approached the project skeptically. The earliest data they used began partway through 2010, and the available history amounted to about 13 years. He questioned how much the first few years, when speculative interest in a new asset could produce extraordinary percentage gains, resembled the market he was looking at in 2023.

A reason to split the sample

One chart compared 2011–2017 with 2018–2022 and plotted the current year separately. The dividing line marked a change in who could participate: CME launched its Bitcoin futures contract in December 2017. Looking at the two periods separately gave Hirsch and Zduńczyk a way to see whether the apparent strength late in the year came entirely from Bitcoin’s early run-up. The authors used different percentage scales so the large early moves didn’t swallow the later pattern.

Steve Strazza thought the shape resembled the seasonal charts for stocks. Hirsch saw the resemblance too, particularly to the Nasdaq, although Bitcoin’s swings were much larger.

When Strazza asked whether Hirsch was prepared to put money behind the research, Hirsch said he was interested in a small position and might use a moving average convergence divergence signal, or MACD, to time it. That borrowed from his equity process: the calendar specified when to start looking, but he still waited for momentum to turn. In stocks, his next seasonal window was opening in October and he wanted confirmation across three indexes. For Bitcoin, he was weighing how to enter a market he was only beginning to trade.

Questions and answers

#

Questions answered in the original 2023-09-28 episode.

Why would Hirsch look for a purchase during Bitcoin’s weakest month?

He was interested in a possible entry near the end of the historically weak period. In the sample, September’s average path reached its low on the 26th ahead of stronger average returns in October and November.

Link to this answer

Why didn’t Hirsch rely on the full-history average alone?

Bitcoin had only about 13 years of data. He questioned whether extraordinary gains in its earliest years represented repeatable behavior, and the study compared periods before and after CME futures became available.

Link to this answer

What did Hirsch want to see before entering?

He was looking for an opportunity to start small and said he might use MACD to time the entry. In his existing equity process he waited for momentum confirmation after a seasonal window opened.

Link to this answer

Adapted by Stock Market TV Editorial from the original video. Source date: . Article published: .