Last Updated on January 15, 2022 by Oddmund Groette
Yes, our backtests indicate that there is an end of month effect in Bitcoin and this can potentially be a viable turn of the month trading strategy.
The end of month effect, sometimes referred to as the turn of the month effect, is well known in US stocks and most other stock indices all over the world. We showed you how you can use this in a turn of the month trading strategy in a previous article.
In this article, we backtested to look for the same end of month effect in Bitcoin.
Let’s backtest the end of month effect in Bitcoin
We use data from Yahoo!finance. Unfortunately, the dataset starts as late as 2014, and thus our time frame is slightly limited.
To make a turn of the month trading strategy we start by doing optimization of both the entry and the exit: we enter at the close x days prior to the new month, and we exit n days after we enter a new month. We emphasize that we use trading days – not calendar days. They can be completely different.
This is the result of our optimization (100 different simulations):
The first column indicates at which last trading day of the month we enter, and the second row is the trading day of the new month we exit. For example, row 1 enters at the close on the tenth last trading day of the month and exits at the ninth trading day of the new month. The table is sorted by profit factor.
The fourth last column shows the max drawdown and that is pretty high for most simulations. The only way to reduce the drawdown is to reduce the holding period.
Let’s formulate an end of month trading strategy in Bitcoin:
- We go long at the close of the third last trading day of the month.
- We sell at the close of the third trading day of the new month.
Let’s see how this performs if we invest 100 000 at the start and compound this amount until today (logarithmic chart – read here for linear vs logarithmic charts):
The average gain per trade is 3.3%, the CAGR is 41.8% and time in the market is 16%. Max drawdown is 25%. Is this good? It depends. It’s much lower than HODL but at the same time, the holding period is reduced a lot.
Is this backtest curve fitted? Yes, to a certain degree it is. However, the average gain for any random period of the same length is 1.33% – much lower. The average gain for practically all entries and exits around the turn of the month is higher than a random period.
All in all, we would say that Bitcoin has a positive end of the month effect. This can hence be used as a turn of the month strategy.