At the closed-door roundtable of the 2025 T-EDGE Global Dialogue, five frontline practitioners who manage huge sums of money established clear rules for digital asset management in the post-"1011 Black Swan" era.
Article by: Li Jingying
Article source: Barron's
On December 20th, at the 2025T-EDGE Global Dialogue, moderator Shigeru and five guests from asset management institutions, trading platforms, and frontline strategy practitioners engaged in in-depth discussions on the theme of "Digital Asset Management and Trading Strategies." The conference focused on three core topics: the evolution of on-chain and centralized trading strategies, the application of AI tools, risk control adjustments under liquidity volatility, and the selection criteria for quantitative teams. The guests shared highly valuable insights and experiences based on their extensive industry experience and market observations.
The five conference guests are:
Jason Huang: Head of Asset Management at JZL Capital. Founded in 2018, the company's core business is diversified asset allocation and quantitative arbitrage services. Currently, it is promoting a digital asset swap program and expanding its traditional compliant brokerage business to optimize its trading strategy.
Wang Jianbo: He entered the digital asset field in 2017, focusing on low-to-medium frequency quantitative strategies, integrating multi-dimensional and effective information, and achieving optimal decision-making through statistical models. His core direction is to create stable family fund products.
Charles: A representative from B7 Capital in Central, Hong Kong. The firm focuses on quantitative fund of funds (FOF) investments in digital assets, with an initial size of US$100 million, covering multiple strategies. It has conducted due diligence on hundreds of quantitative teams and has a deep understanding of market liquidity and how to handle extreme market conditions.
Wonder Xu: Has experience in institutional business at leading trading platforms and in primary and secondary quantitative investment. In 2024, he increased the coverage of CTA strategies and has extensive experience in institutional cooperation and market volatility management.
Spencer Fan: He entered the financial industry in 2008 with a traditional financial background. He founded a quantitative private equity firm in 2015 and entered the digital asset field in 2017. Later, he transformed into the research and development of mid-frequency quantitative strategies and launched a short-term CTA portfolio strategy tool. In the future, he plans to rely on blockchain technology to serve end users.
On-chain strategies are losing popularity due to security and efficiency concerns; centralized exchanges and brokerages are becoming the mainstream asset allocation method for institutional investors.
Host Shigeru: What new developments have occurred in on-chain strategies and centralized exchange strategies in the past year? In actual asset allocation or trading, how should we balance the combination of on-chain and centralized exchange strategies? Are large-scale model-driven AI trading tools effective in real trading? Are they just gimmicks or truly effective trading aids?
Jason Huang: He was involved in on-chain speculation and liquidity-related businesses in the early days, but due to issues such as the understanding of smart contract risks and the efficiency of decentralized wallet management, he has gradually withdrawn from on-chain fund allocation and shifted his focus to centralized trading platforms.
As a core trading scenario, it focuses on medium- and high-frequency strategies, with trading revenue accounting for 60%-70%. At the same time, in order to diversify risks, it is gradually transferring some funds to traditional compliant securities firms, forming a dual-scenario layout of "centralized trading platform + compliant securities firm", and expanding the allocation boundaries by leveraging the compliance qualifications and rich products of securities firms.
AI tools are used only as an auxiliary means to provide trading signal references and assist in code writing. All AI-generated code must be manually checked line by line before going online to ensure that the logic is rigorous. They do not participate in core trading decisions.
Wang Jianbo: On-chain and centralized trading belong to different ecosystems, with significant differences in their functional positioning and incentive orientation. On-chain trading focuses more on compliance requirements and ecosystem operation, with relatively limited trading value, and only a few opportunities may exist for high-frequency arbitrage; centralized trading platforms, with their mature mechanisms, rich products, and efficient fund management functions, remain the core choice for large-scale fund allocation.
For low- to medium-frequency strategies, traditional statistical models are sufficient, and complex technical tools are not necessary; over-reliance on them may increase model complexity and uncertainty. The value of AI tools is mainly reflected in scenarios such as assisting in code writing and text processing in fundamental analysis. Their core still relies on dimensionality selection and data structuring capabilities, and their application space is relatively larger in high-frequency strategies.
Charles: We maintain a cautious attitude towards participating in on-chain strategies. Our core concerns are focused on security risks and the difficulty of penetrating the underlying assets. Participation requires a deep understanding of private key management and smart contract logic. We only focus on compliant arbitrage opportunities on leading decentralized trading platforms and avoid getting involved in complex and nested related products.
AI tools have been widely adopted in quantitative trading across various scenarios, including factor mining, signal generation, and text data processing. However, some models suffer from a "black box problem," making it difficult to accurately attribute the root cause of performance fluctuations, which poses challenges to risk control and strategy optimization.
Wonder Xu: The tiered nature of on-chain strategies means that high-frequency proprietary trading teams with smaller capital and more flexible strategies can explore some arbitrage opportunities in the on-chain market. However, for asset management institutions with larger capital, on-chain trading suffers from constraints such as insufficient security, lack of convenient tools, and limited capital capacity, and their returns are significantly affected by market liquidity.
Arbitrage strategies have a low correlation with AI tools, and traditional algorithms can already meet the needs. In high-frequency and CTA strategies, AI tools can improve efficiency in data processing and factor mining, but external tools have limited empowerment for mature teams and are more suitable for small and medium-sized teams with weaker development capabilities. At the current stage, they are unlikely to become a core competitive advantage.
Spencer Fan: On-chain spot trading will remain a niche market due to management pain points, but on-chain contract trading has certain development potential. Some emerging platforms have already approached centralized trading platforms in terms of user experience and have more relaxed compliance restrictions, making them suitable for global expansion needs. In addition, some prediction markets also present data-driven investment opportunities.
The platform's iterative improvements, including combined margin, multi-asset margin, and enhanced leverage limits, have improved capital utilization efficiency and transaction security, demonstrating the increasing maturity of its risk control system. Currently, its core value lies in improving development efficiency, rather than replacing human decision-making. The generated strategies or code require manual modification before use and cannot independently create comprehensive quantitative strategies. However, in the long term, it still has considerable room for development in niche areas of automated trading demand.
When liquidity disappears, "survival" becomes more important than "earning quickly".
Host Shigeru: In light of the current situation of insufficient market liquidity following the 1011 black swan event, what substantive adjustments have you made to the risk control mechanisms of your strategy products to address liquidity issues? Compared to before, have there been any changes in position sizing, drawdown tolerance, or stop-loss logic?
Jason Huang: The cryptocurrency rating system is based on core indicators such as market capitalization ranking, average daily trading volume, open interest, and listing duration to rate trading targets and set differentiated opening position limits to control exposure to high-risk targets from the source.
Hourly updates include the automatic reduction ranking of held stocks and changes in the balance of related support funds. For volatile periods such as the release of macroeconomic data, a 24-hour shift system involving domestic and international colleagues, combined with multiple monitoring methods, ensures immediate response during periods of significant market volatility. Addressing process vulnerabilities exposed during market fluctuations, the system improves the program's liquidation logic and exposure monitoring mechanism, clarifies position rebalancing requirements, and prevents excessive exposure to any single stock.
Wang Jianbo: The core risk control strategy is to focus on mainstream target transactions. These targets have better market depth and liquidity, which can effectively reduce transaction execution risks and proactively reduce exposure to niche targets.
When market trends are weak and liquidity is insufficient, raise the screening criteria for trading signals, selecting only trading opportunities with higher certainty, and significantly reduce trading frequency. This "conservative first" principle mitigates potential risks. Utilizing factor-driven strategy selection helps to offset the impact of insufficient liquidity to some extent, ensuring the strategy's relative stability under different market conditions.
Charles: Integrate risk control into the strategy and team selection process, comprehensively review the risk control details of the underlying strategy, require the team to provide clear risk response plans, and maintain a cautious attitude towards teams that lack experience in a complete market cycle.
Diversified allocation across multiple dimensions, including assets, strategies, and targets, avoids over-concentration. Low correlation between different assets and strategies helps hedge against risks arising from high market volatility and insufficient liquidity. The risk awareness and execution capabilities of strategy managers are core to the risk control system. The same strategy can exhibit significantly different risk performance under different managers; therefore, this is a crucial evaluation indicator for team selection.
Wonder Xu: Market volatility has made the team deeply aware of the liquidity risks of niche assets. We have now significantly reduced our exposure to niche assets and focused our trading on core assets with more secure liquidity, balancing risk and return.
Establish a dynamic adjustment mechanism to proactively reduce overall position size and leverage during periods of market euphoria and high speculation, avoiding passive risks caused by market corrections and sudden liquidity contractions. Closely monitor the team's risk control implementation throughout the cooperation process; for partners who do not comply with risk control requirements, even if they show excellent short-term returns, the cooperation will be decisively terminated to avoid long-term risks.
Spencer Fan: Based on changes in the market landscape, we adjust the weight allocation of different factors through objective data models to reduce interference from subjective judgments and ensure that the strategy maintains relatively stable performance in different market environments.
The focus is on increasing allocations to market-neutral market-making strategies and event-driven strategies. These strategies have a low correlation with overall market liquidity, effectively hedging return volatility and enhancing the portfolio's counter-cyclical capabilities. A balance must be struck between handling extreme market conditions and ensuring consistent returns. Reasonable volatility thresholds should be set based on risk tolerance, and the ability to cope with extreme market conditions should be improved through factor optimization and strategy diversification.
Teams that haven't experienced a complete bull and bear market are no longer on the list of top investors.
Host Shigeru: How do you, our guests, view the current state of development of quantitative and trading strategy teams (given the increase in the number of teams but insufficient institutionalization)? As asset management institutions or trading platforms, what are the core standards or indicators you value when selecting and evaluating quantitative and trading strategy teams?
Jason Huang: Currently, the situation in the industry is that competition among some traditional arbitrage strategies is becoming increasingly fierce, and investors are less interested in these strategies and prefer to choose cooperation models that can reduce management costs. Quantitative teams are facing pressure on their profit margins and profit-sharing ratios, investors are increasingly demanding stability in their cooperation, and the trend of funds concentrating on top-tier teams is becoming more and more obvious.
The selection criteria for the team are as follows: First, the ability to control long-term volatility, with a focus on the team's performance and recovery ability in extreme market conditions; second, the ability to iterate and optimize strategies, assessing the frequency of strategy updates and the changing trends of core indicators; third, the efficiency of emergency handling and communication, including the response speed under extreme market conditions and the quality of daily communication; and fourth, differentiated competitive advantages, focusing on the team's unique strategic logic and technical expertise.
Wang Jianbo: The widespread use of technological tools has lowered the entry barrier to the industry, but it has also exacerbated the differentiation. Some teams lack core capabilities, their strategies are highly homogenized, and it is difficult to achieve sustained profitability. The life cycle of high-frequency strategies is gradually shortening, while medium- and low-frequency strategies, due to their reliance on more stable factors and cyclical logic, show stronger sustainability. The importance of multi-asset allocation is becoming increasingly prominent.
The selection criteria for teams are as follows: First, the ability to mine and integrate factors, especially the ability to understand and apply key information such as macro data and industry fundamentals; second, the adaptability of strategies to assets, assessing whether the team can design suitable strategies based on the characteristics of different assets; third, a long-term orientation, prioritizing teams that focus on low-to-medium frequency, multi-dimensional, and multi-asset strategies; and fourth, professional ethics and integrity, which are the foundation for long-term cooperation.
Charles: The number of teams is growing rapidly, but their quality varies. Many new teams lack experience in the complete market cycle and only perform well in specific market environments. Their performance is prone to fluctuations after market style changes. Top teams have achieved continuous market share growth thanks to their comprehensive strategy system, technical support, and risk management experience. The "Matthew effect" in the industry is becoming increasingly apparent. The market adaptability of strategies has become a core competitive advantage.
The selection criteria for the team are as follows: First, the background and capabilities of the core team, including academic foundation, industry experience, learning ability and character; second, the team's risk response and debriefing capabilities, with teams that have experienced market risk events and can summarize and optimize them being more trustworthy; third, a combination of quantitative and qualitative evaluation, focusing on both core performance data and soft factors such as risk control awareness and decision-making mechanisms; and fourth, the transparency and explainability of the strategy, avoiding the selection of strategies that cannot be attributed at all.
Wonder Xu: There are still some teams in the industry with inaccurate data, using simulated data to replace actual transaction data and misleading partners; there are significant differences in the institutional operation capabilities of teams, and many teams lack sound risk control and compliance processes, making it difficult to meet the large-scale cooperation needs of institutional investors; strategy diversification is becoming a trend, and teams with the ability to deploy multiple strategies have a greater competitive advantage.
The selection criteria for the team are as follows: First, data authenticity and background investigation, verifying the team's past performance and work experience through multiple channels; second, small-scale testing and gradual cooperation, testing the team's performance with a small amount of capital before deciding whether to expand the scale of cooperation; third, risk control system and execution capabilities, focusing on measures such as volatility control, stop-loss logic, and position management; and fourth, long-term tracking and dynamic evaluation, continuously monitoring the team's performance during the cooperation process and taking timely measures to address potential risks.
Spencer Fan: Global digital asset allocation funds continue to grow, creating broad market demand for quantitative teams; the industry landscape is showing regional division of labor, with teams in different regions focusing on local market resources and clients; cutting-edge trading technologies and strategy iteration capabilities have become core competitiveness, and public events such as quantitative competitions have become important platforms for teams to showcase their strengths.
The selection criteria for teams are as follows: First, core quantitative indicators, including basic assessment dimensions such as profitability and volatility control; second, practical performance and market adaptability, prioritizing teams that have experienced a complete market cycle and can maintain stable performance in different market environments; third, technical strength and strategic barriers, evaluating the team's R&D capabilities, coding skills and unique strategic advantages; and fourth, cooperation compatibility, including communication efficiency, compliance awareness and cooperative attitude.
Overall, the roundtable reached the following consensus: In terms of trading scenarios and tool applications, centralized exchanges remain the core scenario for large-scale capital allocation. On-chain trading has growth potential in the contract field and niche demand scenarios, but security and liquidity issues remain the main constraints. AI tools are currently mainly for auxiliary functions, playing a role in code writing, data processing, and signal reference. They are not yet mature enough to replace human core decision-making. The future development focus is on improving code accuracy and strategy generation capabilities. Quantitative teams need to view their value rationally and avoid over-reliance.
In terms of risk control and liquidity management, insufficient liquidity and black swan events will become the norm in the market, making dynamic adjustments to risk control mechanisms crucial for survival. Core strategies include: focusing on highly liquid assets and reducing exposure to smaller cryptocurrencies; establishing a multi-dimensional risk monitoring and early warning system to improve response speed to extreme market conditions; hedging risks through strategy diversification and asset allocation; and balancing responses to extreme market conditions with normal return performance within risk tolerance limits, avoiding any extreme.
In terms of quantitative team development and selection, intensified industry competition is driving quantitative teams to transform towards institutionalization, specialization, and differentiation. Teams that rely solely on a single strategy or follow trends will find it difficult to achieve sustainable development. Team selection should adopt a comprehensive evaluation system of "quantitative + qualitative". Quantitative indicators focus on long-term returns and risk control data, while qualitative indicators focus on soft factors such as the team's core capabilities, risk control awareness, and compatibility with partners. At the same time, small-scale testing and long-term tracking are necessary to ensure the safety and sustainability of the partnership.
About Crypto Quant 2026: A new brand jointly launched by Barron's China and DeAI Expo, and co-hosted by MetaEra, CGV, and Crypto Alpha, Crypto Quant brings together the strengths of the finance, media, AI, and digital asset ecosystem. Its core objective is to build a long-term evolving digital asset management system using institutional-grade standards. Its initiation stems from three current market trends: AI changing decision-making methods, quantitative trading becoming the mainstream execution tool, and compliance determining the institutional path. It focuses on the synergy of AI, quantitative trading, and compliance systems to build sustainable asset management capabilities.
The brand comprises two core components: First, the Digital Asset Management Forum to be held in Hong Kong in late April 2026. Prior to the forum, closed-door exchanges will be held in the Middle East, the United States, and Asia, targeting institutional investors. The forum will review market changes and explore future trading paradigms, focusing on topics such as trading structures, the evolution of AI strategies, and cross-jurisdictional compliance. The core focus will be on algorithmic liquidity, institutional rights confirmation, and long-term capital collaboration. Second, the Global Digital Asset Quantitative Trading Competition, which will involve more than 60 days of live trading from January to April 2026. Registration has been open since the end of 2025. Participating teams will compete with real accounts and funds on mainstream trading platforms. The competition is divided into comprehensive and arbitrage groups. Teams with stable performance and strong risk control capabilities will be selected through multi-dimensional evaluation. The competition provides a unified tracking system, allowing institutions to observe and select partners over the long term. Rewards include long-term value resources such as prize money, institutional recommendations, and media exposure.
CQ2026 is not just a competition, but a platform for screening capabilities in the real market. It aims to connect global strategy teams and capital, and invites teams and investors who are interested in AI quantitative finance and hope to promote the institutionalization of strategies to participate and jointly explore the long-term path of digital asset management.
Registration link: https://forms.gle/AKxLs1zoowFFuVh57
