- Yeetum Weekly Quant Report
The latest news from a quantitative finance perspective on cryptocurrencies
The analysis has been performed using python language, with an extended use of matplotlib, seaborn, plotly, pyportfolioopt, and sklearn.
We are publishing a weekly report on our market analysis. For additional information about us and our team of analysts visit Yeetum.
The Bitcoin hype seems to have reached its peak, at least for this cycle. After the cycle has corrected, investors will average down on Bitcoin's future prosperity.
What we are witnessing right now is just another step in the Bitcoin life cycle. The market volume for Bitcoin has increased substantially in the past few years, allowing investors to perform a perpetual price discovery. Bitcoin will increasingly look attractive as its own asset class, and its price will keep correcting until a level of stability in its volatility is reached.
Cryptocurrencies are volatile and highly correlated assets, which makes them very challenging to properly manage from a portfolio optimization perspective. Conventional finance techniques, such as mean-variance allocation methods are ineffective when applied to cryptocurrencies: there is a need for more complex mathematical tools that deal with this level of complexity.
A simple buy and hold position is not an advisable strategy in the world of cryptocurrencies. Given the Bitcoin fall, and the evident bearish effect that will be propagated onto the other cryptocurrencies, if investors do not wish to terminate the current position, hedging against this inevitable risk is advisable.
Because the next step in the Bitcoin phase seems to be foreseeable, which is the end of its current cycle, with proper caution, shorting can be advisable as a short-term strategy for Bitcoin and correlated cryptocurrencies.
Our data has been downloaded using python code from Coinbase. It has been further analyzed to provide a comprehensible report.
The data comprehends a total of 12 different cryptos available in our current portfolio: BTC, ETH, OMG, DAHS, BAT, LINK, EOS, ETC, BCH, LTC, ZEC.
Screenshot of the original cryptocurrency dataset
BTC has been omitted, being too high would have made the other lines invisible
As we can see from the graph above, all cryptos are likely to move together because of their high correlation coefficient. Their returns follow the same logic:
Returns for every cryptocurrency, BTC is included as the other lines will remain visible
As mentioned, all cryptocurrencies are very volatile, which makes them a very risky asset, potentially dangerous when overlooked and not managed properly: however, very profitable when managed correctly.
The heatmap representation of the Pearson Correlations between cryptos shows that all assets are highly correlated with each other (>.5 correlation). This means that conventional diversification methods such as the efficient frontier are not effective, as they are unable to operate correctly given this kind of data. Hence the use of more advanced and less orthodox diversification methods is required to provide safety nets to investors.
Heatmap of the Correlation Matrix
After compressing the 12 dimensions (the number of cryptos we are analyzing) we can visualize the correlation matrix in a cartesian plane: each point represents a different asset, while the distance between the points the intensity of their relationships. As we can see, BTC and ETH are very highly correlated, while cryptos such as LINK and OMG are the least correlated, and may provide a diversification opportunity.
3D correlation Matrix
This week we have been selecting two main methods of allocation: HRP, a Machine Learning portfolio optimization method called Hierarchical Risk Parity, and the common efficient frontier.
The efficient frontier is a conventional way to minimize risks, but, at the same time, maximizing profits.
ZEC, 0.08Expected annual return: 157.5%
Annual volatility: 74.8%
Sharpe Ratio: 2.08
This algorithm takes into consideration the clusters within the correlation matrix of the assets, and then group them accordingly.
Clustering Machine Learning allocation method
ZEC, 0.06Expected annual return: 116.1%
Annual volatility: 73.3%
Sharpe Ratio: 1.56
As mentioned in the data, with the fall of Bitcoin, all other cryptocurrencies have been following the trend. Ethereum, the closest crypto to Bitcoin, is going down on the same path.
It is unlikely that, at least for the close future, we are going to see a turnaround in the price of Bitcoin, given the similarity with the previous phase of the BTC life cycle. However, given that the market cap for Bitcoin has kept increasing and there is no hint that it will cease to do so, there is still potential, in the long term, for even higher performance of Bitcoin.
- Date of publication:
- Wed, 01/13/2021 - 13:06
Click on the link - it will be copied to clipboard