Purpose of the strategy and our motivation
When it comes to Bitcoin, it seems like we only hear about two polar opposite ways to play the market : Holding forever and day trading. Both have their merits. The meteoritic ascent of Bitcoin in the last decade was such that it created new wealthy people out of nothing. Indeed, some early adopters had their few thousand bucks investment grow into a multi-million hoard inside of a few years. This wealth creation story started a strong narrative around the reward of holding this asset no matter what. But, the extreme volatility of Bitcoin makes it actually a very hard investment to hold on to throughout all the ups and downs of its cycle. This harsh reality has reinforced the glory around never selling your bitcoin by adding an extra layer of pride in having the fortitude to live through the pain of a massive move down. The crypto community has even created a typo driven name for this type of courageous person: the Hodler.
On the other hand, Bitcoin’s high volatility is very tempting to the day trader. If played right, the roller coaster that is Bitcoin’s price can be an incredible opportunity for an active trader to quickly build wealth.
One of our main motivations for creating our investing strategy originated in this binary narrative of holding vs day trading, as we actually think that the most efficient way to hold bitcoin is probably somewhere in between these two extremes. The Bitcoin cycles have been so defined and predictable (on the long term) in the past that we see no reason to sit through the pain of the long downtrends. But we also think that the high price volatility is extremely difficult to successfully ride on a daily or weekly basis. For example, during an uptrend phase, sometimes you can see a -20% drawdown that recovers within 2 daily candles. This behaviour in the price action is pretty hard to play without grinding capital.
Our second motivation was more scientifically driven. If you have read about our backgrounds you already know that the majority of our team are academic researchers (and a university professor) in the field of robotics, who have many peer reviewed contributions to machine learning, model based predictive control, and advanced sensor processing. As researchers we like numbers, stats, and facts.
And while I personally like the whole Bitcoin story a lot, I’ve always been uncomfortable with the tools currently in the hands of retail investors. No other asset has more self-proclaimed gurus that give daily analyses that actually amount to almost nothing and that are backed by almost nothing. I have heard/read stuff like “we know that this cycle is not over because it’s not how we feel at the top”. We have seen a guru saying in 2020 that they know the exact date that this cycle would peak based on the length of the 2 previous cycles (if you know your stats, you know that you cannot conclude much with two data points). We have also seen people with millions of followers trying to pump Bitcoin using simple technical analysis wrong and ignoring data and the surrounding news and economics. With the FTX bust it seems that this phenomenon of very misguided investors is not exclusive to retail investing. Indeed, Caroline Allison CEO of Alameda Research (the sister company of FTX) said in an interview when asked about how she was using her Stanford background in Mathematics said that she actually only used elementary school math for her trading and avoided stop loss for risk management. See where this got them.
We find this situation of very poorly guided investing very ironic since the decentralized and public nature of Bitcoin offers a gold mine of data that exceeds everything we have ever seen in any other investing asset that has ever existed. To fully understand our excitement here you need to know that although people often think that the main limitation to AI in most fields is the current hardware or algorithms, it’s the lack of good data that is more often responsible for poor AI performance. In many fields, it is pretty hard to get a good amount of quality data. Over the past few years, I had several graduate students that spent their whole summer collecting and labeling robot data. Even Google had to build a whole room of robots and have them run 24/7 for months to acquire 800K of data points. It is also what they did for pretty much a decade with their automated car division, Waymo. But unlike traditional stock market assets that offer little visibility outside of the price action in itself, Bitcoin offers a wealth of information through the public blockchain that gives an accessible view of what is currently happening under the hood. This data, called on-chain data, allows us to track several things like where and when each part of each coin is exchanged and at what price, what type of investors are currently buying or selling, how many new network participants have been added, what is the current computational power of the network, what are the current economic of mining, etc…
Through the work we did on building a hedging strategy for the IOFund (an excellent technology investment fund that we recommend), we came to fully realize the power of using alternative data for assessing the current situation of an asset compared to only looking at its current price action. We really believe that on-chain metrics are the future of crypto investing and think that they are still massively underused. Indeed even famous Bitcoin supporters like the great Cathie Woods and her company Ark Invest only started to promote on-chain metrics as a tool in 2021.
In this context, we decided to use our engineering background (and our non-elementary school mathematical knowledge!) to start looking more deeply into these on-chain metrics in order to build ourselves a good Bitcoin investing strategy. Our optimal results came not from day trading, but from something still more active than simply Hodling. Indeed, with an average of 4 trades per year over the last 10 years, this strategy beats the buy and hold over the same period by 954x or (59749x if you short Bitcoin instead of exiting the market during the downturns). These numbers are crazy high, but to be honest Bitcoin’s returns in its infancy were also crazy high, and our results are over a 10 year period. As Bitcoin matures and loses its explosive nature, we expect to see this number go down, but we are extremely confident that we will continue to significantly outperform Hodling while avoiding the emotional abyss of living through the cyclical downtrends.
(P7.1_2022)
Comments