Raney and Cordeiro met in 2017 through mutual friends in Seattle’s Bitcoin community and set up a software platform that became the genesis of Strix Leviathan’s strategy. “We started dabbling in crypto but were frustrated with the dearth of available tools that could allow us to trade it using hedge fund style strategies. So, we decided to build a more advanced and institutional-grade proprietary technology stack to trade our quantitative strategies and proprietary algorithms,” explains Raney.

The inspiration for the company name came from a children’s book, titled The Owl and the Octopus, and this concept remained albeit with different words: a Strix is a type of owl, and a Leviathan is an octopus. “These perceptive animals are representative of how intently we are watching and observing the market, and the many legs of an octopus symbolize how our system can execute across a variety of venues simultaneously,” says Raney.

The enduring power of commodities trend following

Strix Leviathan launched its NEST strategy in 2018, based on a trend following philosophy adopted by many CTAs and managed futures managers that have traded traditional markets since the 1970s, long before Bitcoin was conceived. “Philosophically, we believe human behaviour in aggregate does not really change – and we have not noticed it change during the institutionalization of the crypto market. We’ve seen the same behaviour over the past 10 years, even after an explosion of institutional interest and volumes since 2014,” explains Cordeiro.

The first CTAs traded only commodities because the earliest futures markets were commodity derivatives, and herein lies another parallel: “Cryptocurrency return distributions and autocorrelation patterns are like commodities. This is logical when cryptocurrencies are, in effect, digital commodities that give you block space on a Blockchain,” adds Cordeiro.

The strategy purely adopted a trend following model for the first few months in 2018 before adding a cross-sectional momentum model, both of which have been employed continuously since, with a typical split of 60% trend, 40% cross sectional.

In 2021, there was an opportunistic excursion into a supply/demand imbalance strategy that profited from mean reversion trades at times of high retail leverage and volumes. This model became less profitable as retail volumes declined and has been switched off for the time being, though Cordeiro mentioned it could be revisited in the right market conditions.

Proprietary trend models

Generalist trend following CTAs have also started trading crypto, typically just Bitcoin and Ethereum futures listed on CME, and often using the same models they apply to other markets. Strix Leviathan’s trend models are recognizably trend following, but they are also distinguished from generic models. “We sample price, volume and volatility data in different ways, partly to take account of sporadic liquidity,” says Cordeiro. 

Strix Leviathan has tailored its models to take into account the idiosyncrasies of the crypto market, including complex data processing and some apparent puzzles. “The top line shows billions of dollars of turnover per day, but markets can still be fairly illiquid. The size of price moves generally does not match volumes: one day Bitcoin can be up massively on billions traded, and another day it can be down a lot on a fraction of those volumes. 

We take a different approach to volumes to capture the disparities between liquidity and activity,” explains Cordeiro. For instance, Strix Leviathan’s models run in real time 24/7/365, including on US-based holidays when there are often big moves, even though the firm has not identified statistical differences in liquidity and behaviour on holidays. One popular theory is that people with day jobs are trading on holidays and weekends.

The application of machine learning is also somewhat nuanced. “A purely time-based model does not effectively incorporate machine learning, and we have developed our own ways of applying it,” says Cordeiro. Strix Leviathan’s machine learning algorithms learn and adjust by themselves, while the main models have historically been adjusted only infrequently.

Predominantly systematic

The quantitative approach is designed to take human emotion out of the investment process, but there is some leeway for discretion in carefully defined circumstances. NEST is 95% systematic. Discretion has been used about 5% of the time, most often to reduce risk and sometimes to increase it or enter new markets.

“We built NEST to capture alpha almost entirely systematically, though there are cases we would need to use discretion. For instance, in 2020, the market was dead and choppy with no directional bias and no volatility, which would have whipsawed the systems. In that case, we just reduced risk. In late 2021, we added a small discretionary trade in Solana, which was a newer coin at the time, and it made a big contribution to returns,” recalls Cordeiro.

Discretion would never lead to a trade in the opposite direction of the system, however. “That would make it hard to quantify behaviour,” Cordeiro points out.

Largely long, somewhat short

NEST is “long biased” over time in that its average exposure is net long, but it is not precluded from going net short within limits – and can sit in cash. In the last 6 years, NEST has quite often been long of all 20 coins it trades. It has occasionally only had short exposure but, as a US fund, it cannot generally short more than 5 or 10 coins within the top 20. This smaller roster of short-able coins means that the short book has rarely exceeded 20% of NAV, although it could theoretically reach 40% if more markets could be shorted. Net long is capped at 100% and NEST does not use leverage.

Top 20 universe

Long-term correlations between cryptocurrencies can reach 99% according to Cordeiro, but this only measures common directional movement; performance dispersion means that some coins can greatly outperform others. And in the short-to-medium term, the relationships are much less stable, generating a significant diversification benefit from trading multiple coins. 

 “When we launched Strix Leviathan in 2018, a $500,000 trade in Bitcoin could move the market. Things have come a long way since then, but we still need to be careful about liquidity,” says Cordeiro, who would love to trade 200 coins if they were liquid enough. Strix Leviathan broadly trades the top 20 coins by market capitalization, which are mostly large-cap layer one coins. Most of the 10,000 plus other coins are not liquid enough for a dynamic trading strategy that might need to swiftly close or reverse positions.

Strix Leviathan trades with the typical major US-based counterparties and exchanges and can sometimes also trade with some UK-based exchanges. It only trades spot crypto for now, but would like to trade futures for capital efficiency, and did use perpetual futures at FTX.

Universe turnover 

There has been some turnover in the universe of coins traded, somewhat due to US regulatory reasons but more broadly thanks to new coin launches; and some coins have dropped out of the top 20. 

“A lot of coins did not exist in 2018 and the most actively traded ones may shift again in the next six years,” says Cordeiro. Back in 2018, Litecoin and Bitcoin Cash were in the top five, but now have a lower ranking. “We also stopped trading Bitcoin Cash for a period of time after we noticed some gaming of the difficulty algorithm following its halving. Miners were jumping back and forth between Bitcoin and Bitcoin Cash depending on the difficulty, which led to unpredictable market behaviour,” adds Cordeiro. 

Cordeiro prefers to trade a dynamic and fluid investment universe because he is not confident about predicting which currencies will emerge as the winners in this dynamic and innovative marketplace, or indeed how liquidity will evolve.

“Spot bitcoin ETFs will likely bring in more one-way flow, but ultimately global market makers are needed for two-way flow,” he says.

The market remains fragmented because regulatory constraints make it hard for any one entity to be active everywhere, leading to phenomena such as the Kimchi premium in South Korea. Longer term, Cordeiro entertains the possibility that “some harmonization of global regulations might change market dynamics”.

Alt coin alpha drivers  

NEST does not arbitrage pricing discrepancies between platforms, but does shop around for the best execution, and seeks to pick and upsize the best performing coins in its universe. Signals are equally weighted, but positions are not. “In the last few months, we have sometimes had as much as 70-80% in alternative coins, outside of Bitcoin and Ethereum,” explains Cordeiro.

The cross-sectional model overweights the coins with the strongest momentum. Wagering on outperforming coins, such as Solana and Avax, has been a key driver of performance, particularly during bull markets when the cross-sectional momentum model punches above its weight.

Playing defense during bear phases

NEST has often been in cash, or modestly net short, during bear markets or “crypto winters”, when its trend models have made disproportionate contributions to returns. 

“During the first crypto winter, in 2018, we sat out the entire months of November and December, when the market was down 5-10% nearly every day. And we were out of the market in March 2020 around Covid. We were also defensive during the Terra Luna collapse,” recalls Cordeiro.

Counterparty monitoring has also helped to avoid losses and preserve capital. “We had zero exposure to Three Arrows and Celsius, which did not pass our due diligence process,” says Raney. “In 2022 we had some exposure to FTX, but fortunately did not lose nearly as much as the broader market.”

Reasonably persistent trends are the best market backdrop for the strategy, which can handle substantial volatility. A random pattern of volatility is not helpful, however, and tends to result in noise-obscuring signals. 

Softening the blow of downdrafts is important due to the asymmetry of drawdown recovery. “A 33% loss can be recovered with a 50% gain, while a 90% loss requires a 1000% gain to break even again,” points out Cordeiro.

Perhaps the best benchmark for NEST is an equal-weighted index of the 20 coins traded, which Strix Leviathan does share with its investors. A more typical mainstream cryptocurrency index is less comparable because it will usually have a higher weighting – perhaps 70% – in Bitcoin and Ethereum, whereas NEST’s exposure is closer to 15-30% in these coins. 

Vision for financial inclusion and democratization

The Strix Leviathan team does not advocate a 100% buy and hold long-only approach, but nonetheless they do have strong faith in the crypto concept. Their vision is partly inspired by personal experiences and the aspiration to democratize finance and promote financial inclusion. 

While serving in the military in Iraq, Cordeiro read about Bitcoin in a news article and realized it could be very helpful for Iraqi individuals and business owners who saw their assets evaporate during the war. Later, when he discovered Ethereum, he was even more hooked on the concept of digital assets.

Raney, meanwhile, grew up in an area with a significant migrant worker population and, as she put it, was “saddened to see them having to first pay fees to cash cheques, and then pay another 20% or so to Western Union or MoneyGram to remit their money back to their families, who often lived in foreign countries”. 

Another formative experience came when working for a luxury yacht company in Singapore, where she perceived an injustice in banking regulations. “Cross border rules and restrictions made it very difficult for Malaysians to transfer wealth, purely due to their birthplace.” 

Raney says that by the time her children are grown up, she expects that digital asset technology will help power their every-day lives. 

“I expect that their finances will run entirely on Blockchain, though it will be in the background so they may not even be aware of the technology powering their banking.”