An introduction to Capriole’s autonomous, fundamentals only, Bitcoin trading strategy which analyzes over 35 on-chain and macro market data points
- Authored by: Mick Herfkens & Charles Edwards
Capriole Investments is a quantitative asset manager. While we gain a lot of insight and learnings from visual charts and individual metrics, at the end of the day our assets are managed by automated algorithms which take a “weight of the evidence” approach. No single metric drives our investment thesis.
In this short piece we would like to introduce our Macro Index strategy and see what it is saying about the fundamentals of the Bitcoin market today.
The Capriole Macro Index
The Capriole Macro Index is an algorithm which uses machine learning to weight and assess an array of fundamental data metrics which we have found give a strong indication of Bitcoin’s relative value throughout historic cycles. The model only looks at on-chain and macro-market data.
Uniquely, price data and technical analysis is not considered as an input in this model.
Tons of Data
We hand selected over 35 of the most powerful Bitcoin on-chain and macroeconomic data metrics. 10 years of this data is fed into the Capriole Macro Index machine learning model. The inputs include a lot of the familiar names such as:
- Hash Ribbons
- Supply Delta
- Puell Multiple
- Realized Profit and Loss
- Dormancy Flow
- Whale and Fish growth
…as well as other traditional finance macro indicators, like AAII sentiment, and much more.
The Machine Learning Model
Following processing of the metrics, the machine learning model assesses each metric’s predictive power and outputs importance weights for each metric. These weights are systematically re-calibrated on a daily basis to give more weight to metrics that have provided more explanatory power of Bitcoin’s history and less weight to those that have provided less insight. The Capriole Macro Index score is then simply equal to the sum product of the weighted-metric scores. Finally, the output is run through a Hodrick-Prescott (HP) filter to remove noise.
This results in our final “Macro Index” score.
Defining Bitcoin Regimes
From the Capriole Macro Index score, we define four Bitcoin regimes based on the index level and slope, namely “Recovery”, “Expansion”, “Slowdown” and “Contraction”. The regime names are reasonably self explanatory. In Slowdowns and Contractions the Capriole Macro Index (and therefore fundamental values of Bitcoin metrics) are in decline. In Recovery and Expansion, they are growing.
When plotting the Capriole Macro Index score against its regimes, a “clock-like” structure can be seen which highlights the cyclical nature of Bitcoin through time over the last decade.
The Macro Index Clock shows the cyclical nature of Bitcoin’s expansion and contraction cycles. Source: Capriole Investments
The Result
We run a long-term autonomous trading strategy on the Capriole Macro Index. This strategy takes long-only positions in Bitcoin. In Slowdowns and Contractions, cash is held. Backtesting suggests the Capriole Macro Index strategy performs well against Bitcoin buy-and-hold with superior long-term returns to Bitcoin and less downdraw over the last 8 years.
Yes. On-chain analysis can work in crypto investing.
Backtest results of the Capriole Macro Index show significant outperformance against Bitcoin buy-and-hold over time. Source: Capriole Investments
So how does the Capriole Macro Index look? The below chart shows the Index plot for the last 10 years. As you can see the current score is in Contraction (red) and quite low (-1.62). This low score suggests we are in a long-term deep value region relative to Bitcoin’s history. However, all prior cycles did go lower before bottoming. That’s not to say that the Capriole Macro Index must go lower this cycle, but it’s worth considering that we historically have.
The Capriole Macro Index for Bitcoin is still in “Contraction” today, but the rate of decline has dropped. Historically we are in a value region.Source: Capriole Investments
The Capriole Macro Index overlaid on Bitcoin’s Price history. Source: Capriole Investments
While we are still in Contraction today, the rate of fundamental decline has been slowing recently. We may well be very close to entering Recovery. When that happens, the Capriole Macro Index strategy will take a long-position in Bitcoin.
The Capriole Macro Index is just one of our in-house autonomous investment strategies. Hopefully the above provides some insight into how we consider fundamental data and the “weight of the evidence” in taking positions. Regardless of how you manage your investments, it’s not a bad idea to look at a spectrum of data and achieve “confluence” before making a decision.
Disclaimer on Backtests
Any Backtest performance returns presented represent hypothetical returns and are meant to simulate how a strategy would have performed during the period shown had the strategy been implemented during that time. Backtested/simulated performance returns are hypothetical and do not reflect trading in actual accounts. Backtest returns are provided for informational purposes only to indicate historical performance had the strategy been implemented over the relevant time period. Backtested performance results have inherent limitations as to their relevance and use. One of the limitations of hypothetical performance results is that they are generally prepared with the benefit of hindsight. In addition, hypothetical trading does not involve financial risk, and no hypothetical trading record can completely account for the impact of financial risk in actual trading, such as the ability to withstand losses or to adhere to a particular trading program in spite of trading losses, all of which can also adversely affect actual trading results. There are numerous other factors related to the markets in general and to the implementation of any specific trading program which cannot be fully accounted for in the preparation of hypothetical performance results, all of which can adversely affect actual trading results. Any and all of these factors mean that no representation is being made that strategies presented here will achieve performance similar to that shown, and in any case, past performance is no guarantee of future performance.
