The Math That Explains How Multi-Strategy Hedge Funds Make Money

Episode Summary

Executive Summary: Bloomberg’s Odd Lots interviews Dan Murillo of Freestone Grove Partners about how multi-strategy hedge funds create returns, manage crowding, and structure incentives. Murillo argues that success comes from optimizing many small business decisions—team size, correlation, comp, and capital allocation—not from a single magic edge, and that quantitative discipline is essential for understanding both markets and human behavior.

Main Topics: What quant means in investing: Murillo defines quant as disciplined, repeatable decision-making supported by modeling, forecasting, attribution, and risk tools—applied not just to markets but to human behavior inside firms. Why multi-strat funds can outperform: The hosts question how many pods can produce more than average returns. Murillo explains diversification math: a firm’s aggregate Sharpe improves with more risk takers, but only up to a limit set by correlation. The 'Dan’s math' framework: Murillo’s core thesis is that firm design choices—number of PMs, incentives, netting structure, and technology investment—can be analyzed systematically to maximize performance and reduce unnecessary complexity. Hiring, skill detection, and internal development: He says the best hiring process focuses on understanding the mechanism of skill rather than chasing past returns, and that firms should prefer developing talent internally when possible. Crowding and market impact: Murillo argues crowding is not inherently bad for an individual manager because being early and then crowded is how one gets paid; he distinguishes firm organization from actual investment strategy. Compensation and capital allocation: The discussion explores how team-based netting, direct incentives, and portfolio size affect both firm economics and allocator outcomes, with smaller netting groups preserving incentives at lower cost. AI and the future of research: Murillo sees AI as an extension of the ongoing arms race in data and analytics, useful for summarization and consensus detection, but not a substitute for true differentiated investment insight.

Key Arguments: Multi-strategy funds do not need to scale endlessly for performance; beyond a point, more people add little unless correlation is reduced. Correlation among PMs is the key limiter of firm-wide returns, and even modest correlation sharply caps the achievable aggregate Sharpe ratio. Investors should ask how skill is generated, not just whether past returns were strong, because realized returns are noisy and often misleading. Management access still matters in short-horizon trading because deep firm understanding helps identify subtle earnings surprises and catalysts. Crowding is a feature of being early in an idea: managers profit as others adopt the same view, then move on before the unwind. Compensation design has real economic consequences; netting at the team level can preserve incentives while reducing total comp cost versus paying each PM separately. AI can help organize and summarize information, but because it is trained on consensus text, it is not inherently a source of differentiated alpha.

Data Points: Episode length: 5 minutes or less - Bloomberg’s Stock Movers promo described the format of short audio reports Years in buy-side hedge fund work: About 25 years - Murillo described his career background Example mean Sharpe ratio per risk taker: 0.75 - Used in the diversification math example Example correlation between risk takers: 10% - Murillo’s first diversification scenario Maximum aggregate Sharpe at 10% correlation: About 2.3 to 2.4 - Upper bound even with an infinite number of risk takers in the example Maximum aggregate Sharpe at 20% correlation: 1.6 - Shows how a small increase in correlation materially lowers firm-wide performance potential Example number of risk takers: 45 - Illustrative platform size in the math example Example team size: 3 - Used to show how team structuring gets to about 95% of maximum Sharpe Efficiency reached with 45 risk takers in teams of 3: About 95% of ultimate maximum - Shows diminishing returns to adding more pods Comp cost increase in fully individualized netting: About 25% more - Comparing paying each of 36 risk takers separately versus netting as one group Comp cost increase with team netting: About 5% to 6% more - Middle-ground structure that preserves incentives with lower cost Illustrative predictive correlation between views and returns: 1% - Used to explain how weak predictability can still generate strong outcomes over many names Illustrative predictive correlation between views and earnings: 10% - Used to show why analyzing underlying drivers is easier than analyzing returns directly Human scale example: $100 million vs 10x larger - Murillo explained how psychological and implementation constraints change with asset size Firm access example: 2 slots vs 15 slots - Illustrated that large firms do not automatically get proportionally more corporate access

Pivotal Quotes: "“There comes a point where adding more people actually doesn’t make any difference.”" — Dan Murillo: Explaining the limits of scale and the role of correlation in multi-strategy platforms "“Crowding in is how you get paid.”" — Dan Murillo: Describing why being early in a crowded trade can be advantageous for individual managers "“The right way of doing quant is where you also mix these two together.”" — Dan Murillo: Defining quant as a blend of mathematical discipline and human judgment

Implications: The episode argues that elite multi-strats are built through process discipline, not magic scale. For investors, it suggests focusing on correlation, incentive design, and true skill provenance. For job seekers, it highlights data fluency, rigor, and grind tolerance.