Bias strategies for Bitcoin trading: price analysis and investment outlook

In this new article we are going to analyze the average price trend of the most important and largest cryptocurrency (by capitalization) in the crypto landscape, Bitcoin.

We will then go on to construct a strategy based on the indications of a proprietary software, the Bias Finder, which is capable of quickly identifying on any market the existence of biases, i.e., recurring behaviors within the historical period considered. 

What are bias strategies on Bitcoin (BTC)

Seasonal strategies fall within the bias category; they are strategies that take their cues from the historical average trend of a market in order to determine in which months (days or weeks, as the case may be) it is appropriate to buy or sell that particular asset.

In the case of Bitcoin and thanks to the Bias Finder, the calculation is soon done. Unfortunately, one does not have many years of history, as Bitcoin‘s first transaction was in 2009, right after the subprime mortgage crisis. At that time, there was no cryptocurrency exchange (Binance and Coinbase for example came many years later). 

That is why the data from the early years of Bitcoin’s life are not very clean (as well as being very different from what they are today). and those that can be considered reliable for use in the various studies are not very deep.

In any case, data from 1 January 2018 through 31 July 2023 will be used.

Annual and Monthly Biases on Bitcoin (BTC)

The images below show different biases, the first is the “annual bias,” from which the market trend is shown for all months of the year (Figure 1). In contrast, the next figure shows the “monthly bias,” which identifies the best days within the month to buy or sell Bitcoin.

imm 1 20
Figure 1. Annual Bias of Bitcoin (BTC)
imm 2 22
Figure 2. Monthly Bias of Bitcoin (BTC)

In Figures 1 and 2 as mentioned, it is possible to appreciate what were the average movements of Bitcoin during the year and during the months, respectively. In the first graph we can see that the absolute worst month to buy Bitcoin is May (famous is the saying “Sell in May” related to stock markets), with declines starting as early as the second half of April and persisting until mid-June. The best months to buy Bitcoin, on the other hand, seem to be October, February, and July (also good in 2023). 

The most positive days of the month, visible in Figure 2, are those that straddle the end of the month and the beginning of the following month. Particularly between the 22nd day of the month and the third day of the following month, there is a rise in prices (again on average), considering the 5 years between 2018 and 2023. 

Moreover, it seems reasonably certain that the first part, between the first and 15th day of each month, is the least volatile and frivolous, while in the second part, between the 20th day and the end of the month, volatility increases. What will help will be the following figures, in which the year-by-year performance will be analyzed in order to assess the stability of the average data obtained from the tests done earlier. 

imm 3 15
Figure 3. Year-to-year Bias of Bitcoin (BTC) analyzed year by year
imm 4 8
Figure 4. Monthly Bias of Bitcoin (BTC) analyzed year by year

In Figures 3 and 4 we see, in different colors, the monthly and annual bias broken down year by year: in green is 2023 (which in fact stops at the end of July), in purple is 2022, in yellow is 2021, and finally in orange is the average of the years between 2018 and 2020.

From these results we are able to confirm how the month of October seems great for buying Bitcoin. Even during 2022, a very difficult year that saw major declines, if taking into account the month of October alone, Bitcoin still managed to limit the damage and stop the descent. Good results for October also in 2021 and during the three-year period 2018-2020.

February is also confirmed as a positive month, which saw rises in all years except 2022. 


The final days of the month remain constant during all years of the backtest. By contrast, the early ones denote a lack of resilience in the last 2 years, 2022 and 2023, where we see more “sluggish” profit curves (Figure 4) between the second and third day of the month.  

Backtest of a trading system bias on Bitcoin (BTC)

At this point, it is necessary to get our hands dirty and develop a strategy with the information obtained. The system will buy $100,000 worth of Bitcoin in the months of February and October. Clearly, in 4 1/2 years of history there will be few occurrences where the strategy can make trades, but for academic purposes we are seeking confirmation through this backtest with a somewhat more practical slant. The trades will be closed on the first day of the following month, 1 March and 1 November, respectively. 

imm 5 6
Figure 5. Equity line strategy bias on Bitcoin (BTC)
imm 6 4
Figure 6. Average trade bias strategy on Bitcoin (BTC)
imm 7
Figure 7. Performance Summary bias strategy on Bitcoin (BTC)

As was to be expected, the results are good (Figures 5-6-7) although not very consistent. In fact, 12 trades are few to represent a reliable statistical sample. However, it is pleasing to see how without additional filters, nor the insertion of stop losses, the strategy manages to obtain a good equity line, with a profitability percentage at 75% and an overall profit of about $160,000, against a drawdown of -$38,000.

The average trade is plentiful to say the least and exceeds $13,000. 

As a further test to try to add substance to these tests, in the following lines we are going to set up a second strategy, different from the previous one, but based on the same results uncovered with the Bias Finder.

Backtest of a trading system bias of the highs on Bitcoin (BTC)

Specifically, the second strategy will make entries on the highest highs of the last 100 hours, in February and October alone. The trades will be closed at the end of the day, thus an intraday strategy. 

In Figures 8, 9 and 10 it is possible to note how, with this configuration, the system makes many more trades than the previous strategy, with an average trade exceeding $600 in 110 total trades.

The total profit reaches $68,000 with a maximum drawdown of -$20,000. 

Again the equity line is increasing and is obviously very similar to that of the previous system. 

imm 8
Figure 8. Equity line strategy bias of the highs on Bitcoin (BTC)
imm 9
Figure 9. Average trade strategy bias of the highs on Bitcoin (BTC)
imm 10
Figure 10. Performance Summary bias strategy of the highs on Bitcoin (BTC)

Conclusion: bias trading system on Bitcoin (BTC)

In conclusion, even this second strategy, which is more consistent than the first, maintains good results and confirms the possibility that there may be recurring and potentially exploitable behavior with trading systems within these specific months, namely February and October. Despite several tests, the certainty that this type of bias will continue into the future is unknown, but with a little patience one can witness the life of these strategies in out-of-sample, and only then evaluate what to do. 

Until next time and happy trading!

Andrea Unger