How do crypto quantitative fund managers obtain Alpha?

This AMA's Theme and Guest Introduction

Theme: How Do Crypto Quant Fund Managers Capture Alpha?

Host

Zheng Naiqian @ZnQ_626

• Founder of LUCIDA

• Champion of the mixed strategy group in the first season of the 2019 Bgain Digital Asset Trading League;

• April runner-up, May champion, and season third place in the 2020 TokenInsight Global Asset Quantitative Competition, Composite Strategy Group;

• Season third place in the 2021 TokenInsight x KuCoin Global Asset Quantitative Competition, Composite Strategy Group;

Guests

Ruiqi @ShadowLabsorg

• Founder of ShadowLabs & Investment Director at DC Capital

• Managing over $300 million in quantitative product management

• Market-making consultant for several exchanges and well-known projects

Wizwu @wuxiaodong10

• Multi-factor & subjective strategy fund manager at RIVENDELL CAPITAL

• Background in computer science and finance

• $20 million in non-traditional crypto strategies

• Focused on on-chain and off-chain data mining and neutral multi-factor strategies

What Does the Alpha Strategy Framework of Fund Managers Look Like?

Zheng Naiqian @LUCIDA:

LUCIDA is a multi-strategy hedge fund. We ensure our performance can traverse bull and bear markets by developing various low-correlation diversified strategies.

Taking our proprietary capital as an example, our return target is to outperform the spot price increase of Bitcoin during a bull market. Therefore, we first conduct a macro timing assessment of the market, determining whether the current market is at the bottom of a bear market or the top of a bull market. This assessment is very low frequency, roughly on an annual basis.

If we believe the current market is at the bottom of a bear market, we will convert all our capital into fully invested Bitcoin and hold it throughout the bull market. On this basis, we will enhance returns using quantitative strategies such as CTA, multi-factor strategies, and statistical arbitrage, which are also core sources of Alpha during a bull market. At the same time, we will dynamically adjust the capital allocation among these strategies based on the current market environment to ensure efficient use of funds.

If we believe the market has reached the top of a bull market, we will liquidate all Bitcoin and convert it into US dollars to weather the bear market. During the bear market, we will also use strategies like CTA and options volatility arbitrage to increase our dollar holdings until the next market cycle.

Thus, all contributions to Alpha fall into two main categories: 1. Macro timing judgments of bull and bear markets, which is one of our core competitive advantages. 2. Return enhancement through quantitative strategies. For example, if Bitcoin rises from $10,000 to $50,000, it is unrealistic to precisely buy at $10,000 and sell at $50,000. Therefore, we will use quantitative strategies to enhance returns, ensuring we can outperform Bitcoin's price increase.

Wizwu:

When it comes to Alpha strategies, it relates to the nature of our fund's capital. We have taken in a lot of native capital from the crypto space, all in crypto, so we have to passively earn Alpha. It is essentially a directional strategy, within which we have multi-factor strategies and some subjective strategies.

As an institution, we need to consider many factors when making subjective trades, including holding periods, liquidity of small coins, etc. These factors limit our selection of targets; having too many positions can lead to dilution, making it hard to outperform the market; having too few positions means competing with project investors for opportunities, so our framework is to do everything.

For example, if we discover a factor, different people may have different approaches to handling it—some neutral, some subjective, some quantitative—representing different trading ideas. Therefore, I combine subjective and multi-factor approaches. Since there are no precedents in the crypto market, we have both factor strategies from the stock market, which are data-driven, and value-based strategies, although we haven't found them; we also have some futures strategies, especially those analyzing inventory and supply-demand relationships. So, everything relies heavily on our understanding of data and trading cues.

However, we do not have a research department like the primary funds in the crypto space because we lack their resources and broad vision. We focus on being flexible and data-driven. Thus, different people in the market earn different types of money. This is somewhat similar to the futures market, where industries earn industry money, quant earns quant money, and subjective earns subjective money. The methodologies differ, and consequently, the profits differ.

Overall, we focus on crypto-denominated strategies, aiming for our strategies to achieve a Sharpe ratio of 3-4 and an annual return of over 10%, with macro timing being done less frequently or very low frequency. Based on this, we derive factors through market insights that can be applied to various strategies, including subjective and multi-factor ones.

In the process of factor discovery, we like to transport some factors from the futures or stock markets for testing, and we also have our own trading experience.

Ruiqi:

We are a purely quantitative and fully automated team, so when we initially designed the structure of Alpha, we adhered to the principles of high engineering and high automation, relying heavily on data-driven execution. Internally, we divide our Alpha framework into execution Alpha and predictive Alpha.

Crypto exchanges are very fragmented, and there are many investment tools. For example, if I want to obtain a trading risk exposure, I can choose to trade futures or spot, or trade on different exchanges. Therefore, at the execution level, we will compare the capital costs of different markets, such as the prices of futures and spot, basis, transaction fees, slippage, borrowing costs, etc. After comprehensively comparing different costs, we will choose the tools with the lowest costs whenever possible. In this part, we can achieve an annualized return of about 5% to 20% through comparison and selection, which we categorize as execution Alpha.

The second part is predictive Alpha, which mainly refers to predictions at different levels, cycles, and targets, including time series and cross-sectional predictions. We will adjust our risk exposure based on these predictions across different targets.

However, there is a special case where predictive Alpha and execution Alpha have some overlap. For example, if I predict a direction, but this prediction may only solve 20% of the problem, the remaining 80% comes from whether I can execute it. This includes order placement techniques, execution probability analysis, conditional probabilities of capital costs, etc. These factors have both execution and predictive elements, and generally, we operate under this system to achieve breakthroughs in our Alpha.

When we conduct performance attribution, the contribution of these two types of Alpha varies. For example, for execution Alpha, our target is to outperform the benchmark by 5% to 20%, making this part relatively certain, but with limited upside potential. Predictive Alpha is different; for example, some of our high-frequency predictions have very thin profits per trade, which can mix with many execution Alpha strategies, but for some medium to low-frequency predictions, their contribution to predictive Alpha may be relatively high.

What Is Your Perspective on the Crypto Market? What Kind of Market Do You Think Crypto Is?

WizWu:

As mentioned earlier, we should earn different types of money in different markets. We earn money from logical analysis in futures, and similarly in Crypto. The characteristics of the crypto market itself are high volatility. For example, the yield on USDT can have a funding rate yield of at least 20% annually during a bull market. Therefore, if we want to make money, we need to think about how to earn around these characteristics. If we enter with USDT, we might first engage in arbitrage, which is a risk-free return.

Currently, in the crypto bull market, the risk-free yield on Pendle is around 30% to 40%. If we calculate the most precise Sortino ratio, the final deduction is the expected minimum return, and after that, the remaining return as a risk strategy is actually not much. This is also one of the reasons we pursue crypto-denominated Alpha.

My market view is that it's about hot money—wherever there is money to be made and clear logic, that's where we go to earn.

This year's market rotation rhythm in the crypto market is very similar to A-shares. In the past five to six years, A-shares have had a main line every year. For example, it started with carbon neutrality, and this year it's AI. However, historically, in my experience and review of the crypto bull and bear cycles, there has only been a main line this year. This year, we have AI and Meme. Before this, the crypto market had no main line; it was indeed a very dull market, which is a significant difference from this year. Therefore, if you can catch AI or Meme in the crypto space this year, you can make a lot of money.

When capturing the hotspots and sector rotation patterns in the crypto industry, momentum itself is a very important part. Besides data, we also pay attention to Twitter sentiment. However, if there are very few targets, the data we can focus on is still primarily value-related.

We have an internal tool similar to Wind. We have been working on factors for almost two years, storing market data and Twitter sentiment. However, we do not focus much on sectors because we do not capture sector rotations in that way; our factors will select coins with better elasticity within sectors to buy these assets.

Ruiqi:

We believe that the crypto market is a highly speculative market, primarily composed of a large number of continuous trades and occasional event-driven trades. This is also why we continue to participate in the market.

Compared to other financial assets or markets, its emotional trading and event-driven trading components are more pronounced. This makes it more suitable for capturing through quantitative methods, which aligns with our trading advantages.

As the market has developed, competition has intensified, whether in trade execution or prediction. Now, there is a flourishing of strategies, but there are still some structurally high opportunities. These structural trading opportunities are still filled with a lot of emotional and event-driven elements. The market is beginning to undergo structural differentiation.

First, in terms of market predictability, the effectiveness of pricing on old assets has further improved. Specifically, we can see that previously, a trend might take several hours or even a day to ferment, but now a trend might end in just 10 minutes, with significant errors caused by different factors being quickly corrected. However, we find that there are still good Alpha opportunities in new assets.

If we also participate in some Altcoins, we will find that everyone's narratives will include some new assets, whether in competitions, startups, or new trends. You will find that the factors used previously are still effective on these assets. However, new assets present challenges in acquisition. For example, your technical implementation, data access, and the stability of trading models may be lacking.

How Do Different Factors Contribute in the Crypto Market? What Are the Underlying Sources of Returns for These Factors?

Wizwu:

The characteristic of the crypto market is that funding rates are high, meaning the basis is large. For futures, the basis can be understood as the monthly difference. If we consider them as the same thing, the fluctuations in the monthly difference in the crypto market are significant. Arbitrage characteristics are built around this, and alternative factors may also be based on this logic.

Additionally, due to the high volatility of the market, some coins in differentiation have high elasticity, so the real profits depend on timing. We tried this momentum and found that neutral momentum only earns about the same as Bitcoin during a bull market. If you do not time the market, it is difficult to see good excess returns. This is also closely related to the trading mechanisms in crypto.

Moreover, the data provided by exchanges and some off-market data differ from traditional markets. Therefore, many of our excess returns come from these unique aspects and strategies that have been overused in traditional markets.

Ruiqi:

One representative of emotional factors is the momentum factor, which essentially involves chasing trends and selling on dips. The profits from this factor primarily come from the market's overreaction.

For example, when retail investors see a certain coin rising, they usually believe that this upward trend will continue, and they rush to buy. At this point, we can ride the wave and profit. Additionally, we can engage in momentum reversal trading, based on judgments of market overreactions, to preemptively position for reversal operations. The core of these trades lies in exploiting the market's overreactions to gain profits.

The profits from event-driven factors mainly come from the repricing of assets, which requires a certain reaction time. For instance, by monitoring data on Twitter or potential data from major events, we can react quickly after an event occurs. For example, when CPI data is released, Bitcoin prices may experience significant fluctuations. In such cases, reacting quickly and trading can yield profits.

From a high-frequency trading perspective, many traders are insensitive to trading costs, leading them to conduct large trades entirely in a single market. This behavior can significantly impact the market, creating arbitrage opportunities. The liquidity factor is long-term effective in high-frequency markets and is one of the important tools for fund managers to capture Alpha.

What Do You Think Are the Methodological Differences in Gaining Alpha in the Crypto Market Compared to Traditional Financial Markets? How Can We Capture More Alpha in Crypto?

Zheng Naiqian @LUCIDA:

In recent years, I have clearly felt that people are perhaps the most critical element of Alpha. Despite the development of the crypto industry, the average level of practitioners in the crypto industry, especially secondary market participants, shows a significant gap compared to the A-share market.

The second point is data; the infrastructure of this market is simply too poor. There is almost no comprehensive data provider, like Wind or Bloomberg in the A-share market. The data quality is poor and highly fragmented. Obtaining data is a headache for many teams, but without data, how can you build models?

I believe that if institutions have a significant advantage in talent and data compared to their peers, it will be a stable source of excess returns.

Wizwu:

The crypto market has several notable characteristics compared to traditional financial markets: high volatility, high elasticity of small coins, and strong speculative sentiment. To gain Alpha in the crypto market, we must explore strategies around these characteristics.

A core issue is that the risk-free arbitrage returns in the crypto market are too high. This is devastating for value factors in the crypto market because there are very few projects that can provide stable USDT dividends. Therefore, when we try to calculate value, PE, or price-to-earnings ratios, we find that no matter how we calculate, the returns in US-denominated terms are far inferior to arbitrage returns. Thus, it is not feasible to use traditional financial market value factors to measure Alpha in the crypto market.

In the crypto market, the core values we need to focus on differ from those in traditional markets. In traditional stock markets, factors like value and price-to-earnings ratios are core, whereas in the crypto market, we may pay more attention to the price-to-dream ratio, which reflects optimistic estimates of future expectations and everything derived from achieving those expectations.

A specific example of a value factor is in Layer 2 (L2) solutions, such as MATIC, where changes in the number of native token addresses holding between 10 to 100 USDT often indicate market trends. When a public chain is about to welcome a blockbuster application or mass adoption, an increase in these small holders is usually a positive signal, resonating well with market sentiment and price, and often occurring early. Addresses like these essentially represent individuals, reflecting the issue of user quantity. From the perspective of this factor, you can think of addresses holding balances between $10 and $100 as more likely to represent real users.

Ruiqi:

I have summarized a few differences:

1. Information asymmetry caused by market fragmentation: The fragmentation of the crypto market leads to information asymmetry. Non-professional investors find it challenging to understand market conditions, making arbitrage opportunities particularly evident.

2. Chasing trends and market volatility: Unlike traditional financial markets, crypto assets are often traded across multiple regional markets. This fragmentation makes chasing trends and being reactive more common, with investors frequently switching their focus and engaging in irrational trading.

3. Market manipulation: Market manipulation is more prevalent in the crypto market than in traditional markets. For most ordinary investors, it is challenging to exploit this phenomenon for trading or to design trading strategies. However, some high-frequency trading firms can manipulate the market on a larger scale than in traditional markets to gain Alpha. Such behavior is illegal in traditional markets and can lead to imprisonment.

Differences in the Asset Management Product Landscape of the Crypto Market Compared to Traditional Financial Markets

Zheng Naiqian @LUCIDA:

I have noticed that over 80% of secondary teams are engaged in very neutral arbitrage strategies, leading to significant homogeneity between strategies.

From an investment perspective, the principles of the strategies themselves are not complex, and if you are doing relatively low-frequency trading, you do not need to invest much effort in trade execution. This results in over 80% of products competing in the arbitrage space, making it particularly unappealing to engage in CTA or options/multi-factor strategies compared to statistical arbitrage. The same applies to high-frequency trading; you can optimize all your trading details, but in terms of managed scale, there is still a significant deviation compared to arbitrage. So, do you think arbitrage products will become the mainstream of the market in the future?

Wizwu:

Not just in the crypto market, but in traditional financial markets, bond trading is also a significant portion. The trading volume of different levels of bonds is also considerable, so arbitrage trading will always exist. As long as it can be operated under a semi-compliant premise, the arbitrage returns in the crypto market can reach at least two to six times those in traditional markets, providing a very high capacity and return space for arbitrage trading, so this situation will continue.

As for other strategies, such as CTA strategies, they are also a large-capacity option. These strategies may only be genuinely recognized by the market after arbitrage returns decrease. At that time, when we look at the Sharpe ratio of such strategies, it will look very appealing. Currently, arbitrage returns are calculated in US-denominated terms, thanks to the unified accounts of exchanges, we can also run similar strategies using crypto-denominated terms. Therefore, our current direction is to use USDT for arbitrage and crypto for risk, which is the best allocation method.

Ruiqi:

I generally agree with Wiz's viewpoint.

First, the market is highly fragmented, and there are so-called barriers to capital entry, which may be difficult to resolve in the next two to three years. Therefore, in the foreseeable future, arbitrage opportunities will continue to exist. Even if the arbitrage space decreases, arbitrage trading volume and capital capacity will still be a major part of the market.

However, by then, arbitrage may not exist in the form of asset management products. More likely, it will be self-operated by high-frequency quantitative teams, with high-frequency teams directly taking the profits without additional profit distribution to the market. This is likely to be the case. For some asset management projects, they may settle for providing adjusted risk-return ratios that are reasonably good, such as statistical arbitrage and CTA strategies. In the next two to three years, such a soil may begin to emerge.

Zheng Naiqian @LUCIDA:

The structure of crypto asset management products is significantly different from that of A-shares. I have observed that the most mainstream products in A-shares are index-enhancing products, whether they benchmark the 300, 500, or 1000 broad-based indices. Products based on index enhancement should be the best sellers. Most of these index-enhancing products rely on multi-factor models.

However, I find that such products are almost non-existent in the crypto market. I know that there are fewer than 10% of teams developing multi-factor strategies. Why is the proportion of teams developing multi-factor strategies so low?

Wizwu:

The reason is that the yield on USDT in the market is simply too high. For example, on PENDLE, I almost buy all USDT there. In this case, I wouldn't choose my own strategy. Because when my strategy subtracts 30% risk and divides by volatility, its performance is even worse than the Sharpe ratio and other indicators in the traditional futures market.

Therefore, I believe that when the risk-free returns in the market are so high, people will naturally choose risk-free returns. When calculating in this way, the ratio for measuring strategy standards must subtract a risk-free return. When we use the true risk-free return of this market (annualized 30%) for calculations, everything becomes futile; no matter how we calculate, it is meaningless.

Our multi-factor strategies have become more diversified. Initially, we did design according to neutral multi-factor strategies in A-shares or traditional futures. However, over time, it has gradually diversified and become very varied, incorporating more subjective factors. I think the core reason is that the market's drawdown cycles are very short, and changes happen very quickly. In this case, implementing multi-factor strategies presents some framework issues. We cannot rely solely on the recent two years of market conditions to prove that a factor is long-term effective.

In traditional markets, we may explore a factor, and it must be tested not only in A-shares but also in US stocks. If it has been effective in US stocks for 20 years and in A-shares for 5 years, we can say it is an effective factor that can be used for large capital operations. However, in the crypto market, using such a factor to create a neutral strategy makes it difficult to have such verification opportunities. We may only be able to look at one or two years of backtesting, which is not very reasonable in terms of framework.

Ruiqi:

My feelings may differ, which also depends on our understanding of this framework.

I have observed that more people are engaged in time series trading on mainstream coins, such as trend trading on Bitcoin and Ethereum. However, if we consider trend trading on 100 targets, there are very few teams doing that. There are many doing time series trading, but few doing cross-sectional trading; this is what I have observed.

If I were to attribute this, I believe there are several main reasons:

First, the data length issue. Most assets may have only gone through one cycle, with no longer data available for validation and backtesting.

Second, even for assets that have gone through multiple cycles, such as EOS, it became inactive after 2017-2018, making it difficult to include it in the target pool. There are many similar assets in the crypto market, and very few assets can complete several cycles while maintaining activity and liquidity; basically, only Bitcoin and Ethereum can do that. Others, like Solana, have also been dormant for a long time and have only recently become active.

Third, relatively speaking, the effectiveness of time series factors may be more pronounced in practice than that of cross-sectional factors. The underlying logic is that the response to emotional momentum has long been present, and we can effectively plan using traditional trend trading frameworks. In contrast, the relative strength of cross-sectional factors is unstable because many targets themselves are unstable. They do not experience multiple bull and bear cycles like traditional commodities or stocks, making the relative strength comparisons more stable. In the crypto market, a target that is strong in one wave may disappear in the next wave, making it impossible to verify whether its relative strength comparison exists.

What Do You Think Is the Measure of Value for Crypto Assets? Where Does the Value of Crypto Assets Lie?

Ruiqi:

From the current situation, the value of the crypto market is equivalent to attention. In other words, it is now an attention-driven market. Regardless of the underlying logic of a project, as long as it can attract attention, it can gain value. This may have some similarities with the market momentum mentioned by Wiz, but I think it is not entirely the same. Simply put, it resembles a product of an attention economy.

In the long run, we hope, and many practitioners and VCs are working towards a direction where future value reflects actual applications and the competitiveness of ecosystems. But at least in the current state, the market does not entirely reflect this.

Bonus: What Do You Think of the Market Now? How Do You Think Bitcoin Will Perform in the Near Future? (This is just a subjective guess, not responsible for accuracy).

Wiz:

If I were to guess, it will oscillate at this position, with not much upward space. Even if it breaks new highs, the increase might only be around 30%, and then it may have to pull back. From the current level, I believe that major risk assets globally may not have much room for growth. This is just a guess, and it is quite revealing.

Ruiqi:

I would be more optimistic because I think we haven't started lowering interest rates yet. Although I didn't have faith in Bitcoin before, I now consider myself a half-believer. Therefore, I think it is possible to reach $150,000 within this bull market cycle in the next two years.

About LUCIDA & FALCON

Lucida (https://www.lucida.fund/) is an industry-leading quantitative hedge fund that entered the crypto market in April 2018, primarily trading strategies such as CTA/statistical arbitrage/options volatility arbitrage, currently managing $30 million.

Falcon (https://falcon.lucida.fund) is a next-generation Web3 investment infrastructure based on a multi-factor model, helping users "select," "buy," "manage," and "sell" crypto assets. Falcon was incubated by Lucida in June 2022.

For more content, visit https://linktr.ee/lucida_and_falcon