From Fading Edges to Fixed Gains
Mar 3, 2025
The Kickoff
February may be short, but the markets didn’t hold back—signals aged, fixed-income strategies gained momentum, and shifts in data trends kept quants on their toes. As trends settle and cycles play out, the search for alpha continues. In this edition, we explore what’s holding firm, what’s fading, and what’s next as we step into March.
The Compass
Here's a rundown of what you can find in this edition:
Newest partner additions to the Quanted data lake.
From conferences to a crisp new website—what we’ve been up to
Why fixed-income strategies are gaining traction among hedge funds
Insights from our chat with Conrad Poulson, former CEO of Huq Industries
Round-up of the latest news impacting the markets
How to quantify the life-cycle of a quant signal before it becomes a liability
Why we love Mark Fleming-Williams' "The Alternative Data Podcast"
A resonating quote from Ray Dalio to take into March
Insider Trading
We’ve been making moves this month, both online and offline. Our new website is live—built to make it clearer and easier for both quants and data vendors to see exactly how Quanted fits into their workflows.
Meanwhile, we’ve been on the ground at BattleFin in Miami and Eagle Alpha in New York, catching up with partners and meeting new faces. One theme kept coming up in conversations: having data isn’t enough anymore. The real challenge (and opportunity) is how fast funds can find and test new data. The usual approach was agreed to be slow, expensive, and often a shot in the dark when it comes to whether the data will actually improve a model. Good thing there are alternatives out there ;)
Overall this month has been about building momentum—expanding our reach, refining our website, and continuing to deepen our data coverage. More updates are in the works, and we’re looking forward to keeping up the pace and delivering even more value to quants and data providers.
On the Radar
On the partnership front, we’ve added four new data partnerships to the lake, signing agreements with EDI, ETF Global, Intrinio, and SIX. This expands our range of data sources across market, fundamental, alternative, and economic data—bringing more breadth and depth to the features available for quants to test and identify which have the most impact on their models.
Provides corporate actions and pricing data with global coverage, offering accurate insights into dividends, mergers, and stock splits—critical for quants who require clean, timely data for model development.
Provides ETF analytics with proprietary risk and sentiment indicators, offering deeper insights into fund flows, factor exposures, and market trends beyond standard ETF datasets.
Offers a comprehensive suite of financial data APIs, including real-time and historical market data, fundamentals, and options pricing, facilitating seamless integration into financial models and applications.
Sources and consolidates market data from all over the world. From wealth management to media organisations, customers have been using SIX to gain a comprehensive view of global markets for over 90 years.
The Tradewinds
Expert Exchange
We recently sat down with Conrad Poulson, former CEO of Huq Industries, where he led the company in transforming mobility data into actionable intelligence for government, finance, retail, and real estate. Over nearly a decade, Conrad helped scale Huq into a leading provider of mobility insights, serving 300+ local councils, investors, and major corporations.
With a career spanning leadership roles at Orange, startup mentorship at Techstars, and pioneering mobile innovations, Conrad brings deep expertise in data-driven decision-making and high-growth leadership. As he moves on from Huq, we discussed his reflections on the industry, the evolution of mobility data, and what’s next in his journey.
Over your nearly 10 years leading Huq, how have you seen the alternative data industry evolve? What’s been the most defining or memorable moment for you?
Over the last 10 years we’ve seen the development of more consistent engagement models and pricing levels however, I still think, we are in the foothills when it comes to exploring and extracting value from alternative data. The defining moment for me was our first 6 figure sale as it was an objective validation of the value and potential of our data.
Throughout your career, what were the 3 key skills you honed that elevated your performance most?
Keeping focused on the big issues — It's easy to get caught up in the day-to-day, but the ability to filter noise and stay laser-focused on what truly moves the needle is essential.
Identifying and attracting great people — The best ideas and strategies mean little without the right people to execute them. Identifying the right people to work with is, I think, a deeply human skill and one that I’m always trying to refine.
Building great culture requires constant effort — I want to work in environments where people feel motivated and empowered. Getting the balance between leading and empowering your team takes practice
Could you provide a couple of technical insights or innovations in recent mobility data that you believe would be particularly interesting to investment professionals?
One of the biggest historical challenges in using mobility data for investment decisions has been entity resolution—accurately mapping raw data points to real-world locations and, in turn, linking those locations to relevant investment entities. Recently, I’ve seen promising advances in AI-driven tagging and classification techniques. Refining these tools is key to improving the accuracy of location attribution, reducing false signals, and allowing investors to gain more reliable insights into foot traffic patterns.
How do you see the growing adoption of alternative data changing the relationship between financial data vendors and investment teams?
AI-driven tools are lowering the barriers to extracting insights from complex datasets, making alternative data more accessible to a broader range of investment professionals. I think this shift will likely lead to greater adoption of alternative data across all tiers of the investment landscape, not just hedge funds with dedicated data science teams. At the same time, as AI simplifies data extraction, there’s a risk of over-reliance on automated insights without proper due diligence. Investment teams will need to be increasingly discerning about how they validate and contextualise the data they use.
Aside from AI, what major trends do you see having the most profound impact on mobility data over the next 5, 10, and 25 years?
In the short to medium term, I expect the most profound developments in mobility data to come from advancements in predictive analytics. There’s growing research exploring the ability to forecast human movement at scale, which could unlock tremendous value across industries, enabling investors to anticipate patterns before they materialise and governments & urban planners better able to predict and manage infrastructure demands.
Now that you’re transitioning from Huq, what is the next major project or initiative you’re working on?
My focus is now on helping businesses identify and protect the unique data pools they are creating—many companies don’t even realise the proprietary value they hold. Beyond that, I work with data owners to maximise the commercial potential of their assets, whether through strategic partnerships, monetisation models, or operational optimisations.
Before we wrap up, is there anything else you’d like to share with our readers?
Feel free to connect with me on LinkedIn if you'd like to continue the conversation.
Numbers & Narratives
Fixed Income Hedge Funds: Why the Surge Makes Sense
Institutional investors are ramping up fixed income hedge fund allocations—62% plan to increase exposure in the next 12 months, per RBC Global Asset Management. The report highlights interest rate policies (58%) and geopolitical risks (60%) as the biggest drivers.
This aligns with what we’re seeing in the data. Central banks are navigating inflation without triggering liquidity crises, leading to sharp repricings across sovereign and corporate bond markets. The volatility in rate expectations creates persistent mispricings in swaps, repo markets, and structured credit—prime ground for hedge funds running systematic relative value strategies.
On the credit side, dispersion is increasing. 45% of investors cite improved market liquidity, but it’s uneven. Higher dispersion in CDS indices vs. cash bonds, along with structural changes in market-making, means there’s more room for arbitrage in credit derivatives and distressed debt. Selectivity is also rising—47% of institutional investors won’t consider funds under $100M AUM. That’s partly a function of execution constraints. Many of these strategies depend on tight spreads and high turnover, meaning scale matters.
The opportunity is clear: Investors are betting on funds that can navigate structural dislocations with precision. Quants with models tuned to evolving rate regimes and liquidity cycles will find this an optimal environment.
Market Pulse
Tariffs & Tumult
The latest escalation in trade tensions is forcing quants to rethink market structure, volatility regimes, and risk modeling. Trump's tariffs—10% on China, 25% on Canada, Mexico, and a potential 25% on EU imports—have triggered a broad market sell-off, with the Nasdaq dropping 2.8% and the Hang Seng plunging 3.3%, reflecting investor anxiety over global growth. The auto sector is under particular strain, as BMW (-2.6%) and Porsche (-2.4%) slid on fears of retaliatory EU tariffs, which could further disrupt supply chains already weakened by post-pandemic constraints. Meanwhile, Nvidia’s 8.4% decline, despite strong earnings, underscores investors shifting away from high-multiple tech stocks amid tightening financial conditions and rising import costs. The dollar’s 0.8% appreciation signals capital rotation into safe-haven assets, while Brent crude’s 1% decline reflects weakening demand forecasts. This combination of equity stress, currency divergence, and commodity softness creates an environment of volatility clustering, where correlation breakdowns and factor rotations become more unpredictable. For quants, this requires rethinking hedging strategies, especially as trade-driven inflation risks challenge traditional macroeconomic assumptions.
Confidence in Crisis
The resurgence of stagflation risks is forcing quants to rethink macro regime classifications and stress-test portfolios against a structurally different inflation-growth dynamic. U.S. consumer confidence plunged 7 points to 98.3 in February, its lowest in eight months, reflecting rising anxiety over economic stagnation and price pressures. Inflation remains anchored at 3%, but 12-month expectations have surged to 6%, the highest since May 2023, eroding real yields and pressuring rate-sensitive assets. Across the pond, UK consumer sentiment collapsed to a ten-month low as private-sector employment contracted for the third consecutive month, with job losses at their highest since 2021. Global sentiment remains fractured—Latin America shows resilience, but Europe is faltering, with Hungary’s confidence index at just 35.1, the second-lowest globally. As historical analogues to the 1970s gain relevance, quants must recalibrate risk models: mean-reversion assumptions are breaking down, risk-parity allocations are vulnerable to shifting cross-asset correlations, and vol-control strategies must adapt to higher macro uncertainty. Defensive positioning in gold, long-duration bonds, and inflation-linked assets is already reflecting this shift.
Navigational Nudges
Most signals don’t die overnight—they fade, decay, and get arbitraged away. The challenge isn’t just identifying alpha but knowing when it’s losing its edge. A high Sharpe ratio today doesn’t guarantee persistence tomorrow. Markets adapt, inefficiencies shrink, and staying ahead means recognising when a signal is deteriorating before it becomes a liability.
Here's how:
1. Eigenvalue Shrinkage and Signal Decay
Signals often blend into noise before completely dying. Instead of just tracking performance, decompose signals via PCA and monitor eigenvalue shrinkage in the covariance or correlation matrix. If a factor’s eigenvalue steadily declines relative to others, it’s losing explanatory power. A rising condition number suggests increasing multicollinearity—useful context but not a direct death signal. The real concern is whether your signal's variance contribution is fading into statistical noise.
2. Hidden Markov Models for Regime Shifts
Alpha isn’t static—it’s regime-dependent. Instead of relying on rolling Sharpe ratios, fit an HMM to track probability shifts between signal states. A spike in transition probability to a low-performance regime suggests structural decay, but there’s no universal “90% threshold” that confirms it. Hierarchical HMMs can model multi-layered shifts, but properly linking them to signal persistence is non-trivial.
3. Persistent Homology for Structural Breaks
Market inefficiencies have structure, and topological data analysis (TDA) can detect fragmentation in feature spaces. If Betti numbers increase or persistence diagrams destabilise in a well-defined representation, it could indicate a weakening signal. That said, applying TDA meaningfully to time series is complex—it’s an interesting research area but far from a standard approach for tracking signal decay.
4. Granger Causality and Signal Arbitrage
Alpha doesn’t “transfer” so much as become less exploitable. If your signal becomes Granger-caused by liquidity metrics like order book imbalance, it may be getting front-run. But Granger causality only implies statistical precedence, not true causation—there are plenty of confounding factors, from correlated signals to market structure effects. Alone, it’s a weak argument for signal obsolescence.
5. Multi-Armed Bandits for Dynamic Allocation
Deciding when to kill a signal is tough—automating it helps. Multi-armed bandits (MABs) can reallocate weight dynamically, but framing them as a “keep vs. kill” mechanism is an oversimplification. Most implementations gradually adjust exposure rather than outright cutting signals, and defining the right reward structure is far from trivial.
6. Bayesian Change Point Detection (BOCPD) as an Early Warning
Markets leave statistical fingerprints when signals degrade. BOCPD can flag shifts by monitoring a signal’s rolling t-stat, but interpreting a hazard function spike as “the market learned my signal” is speculative. It just tells you something has changed—whether that’s due to a competitor, macro factors, or liquidity shifts needs deeper analysis. GP priors help smooth noise but can also mask real breakpoints.
7. Quantum Annealing for Portfolio Optimisation
For multi-signal portfolios, classical optimisation struggles. Quantum annealing (or simulated annealing) can help solve combinatorial selection problems, but calling it a natural way for signals to “fall out” is a stretch. QUBO formulations require explicit obsolescence penalties—there’s no magic self-pruning mechanism. Plus, quantum methods are still experimental in finance, with real-world applications lagging behind theory.
Alpha isn’t forever, but adaptation is. Markets evolve, inefficiencies fade, and signals lose their edge. Tracking structural decay keeps strategies sharp and ensures you’re always making room for what’s next.
The Knowledge Buffet
by Mark Flemming-Williams
If you’re working with alternative data in your investments, The Alternative Data Podcast should be on your radar. Mark knows the space inside out—his experience at CFM and Exabel gives him real insight to exactly where alt data adds value. He asks the right questions pushing guests to show how their data can genuinely improve predictive models making it one of the most useful resources out there. Don’t say we didn’t tell you—expect to leave every episode with insights that stick and ideas you’ll actually use.
The Closing Bell
"What is most important isn’t knowing the future — it is knowing how to react appropriately to the information available at each point in time."
- Ray Dalio
The Kickoff
February may be short, but the markets didn’t hold back—signals aged, fixed-income strategies gained momentum, and shifts in data trends kept quants on their toes. As trends settle and cycles play out, the search for alpha continues. In this edition, we explore what’s holding firm, what’s fading, and what’s next as we step into March.
The Compass
Here's a rundown of what you can find in this edition:
Newest partner additions to the Quanted data lake.
From conferences to a crisp new website—what we’ve been up to
Why fixed-income strategies are gaining traction among hedge funds
Insights from our chat with Conrad Poulson, former CEO of Huq Industries
Round-up of the latest news impacting the markets
How to quantify the life-cycle of a quant signal before it becomes a liability
Why we love Mark Fleming-Williams' "The Alternative Data Podcast"
A resonating quote from Ray Dalio to take into March
Insider Trading
We’ve been making moves this month, both online and offline. Our new website is live—built to make it clearer and easier for both quants and data vendors to see exactly how Quanted fits into their workflows.
Meanwhile, we’ve been on the ground at BattleFin in Miami and Eagle Alpha in New York, catching up with partners and meeting new faces. One theme kept coming up in conversations: having data isn’t enough anymore. The real challenge (and opportunity) is how fast funds can find and test new data. The usual approach was agreed to be slow, expensive, and often a shot in the dark when it comes to whether the data will actually improve a model. Good thing there are alternatives out there ;)
Overall this month has been about building momentum—expanding our reach, refining our website, and continuing to deepen our data coverage. More updates are in the works, and we’re looking forward to keeping up the pace and delivering even more value to quants and data providers.
On the Radar
On the partnership front, we’ve added four new data partnerships to the lake, signing agreements with EDI, ETF Global, Intrinio, and SIX. This expands our range of data sources across market, fundamental, alternative, and economic data—bringing more breadth and depth to the features available for quants to test and identify which have the most impact on their models.
Provides corporate actions and pricing data with global coverage, offering accurate insights into dividends, mergers, and stock splits—critical for quants who require clean, timely data for model development.
Provides ETF analytics with proprietary risk and sentiment indicators, offering deeper insights into fund flows, factor exposures, and market trends beyond standard ETF datasets.
Offers a comprehensive suite of financial data APIs, including real-time and historical market data, fundamentals, and options pricing, facilitating seamless integration into financial models and applications.
Sources and consolidates market data from all over the world. From wealth management to media organisations, customers have been using SIX to gain a comprehensive view of global markets for over 90 years.
The Tradewinds
Expert Exchange
We recently sat down with Conrad Poulson, former CEO of Huq Industries, where he led the company in transforming mobility data into actionable intelligence for government, finance, retail, and real estate. Over nearly a decade, Conrad helped scale Huq into a leading provider of mobility insights, serving 300+ local councils, investors, and major corporations.
With a career spanning leadership roles at Orange, startup mentorship at Techstars, and pioneering mobile innovations, Conrad brings deep expertise in data-driven decision-making and high-growth leadership. As he moves on from Huq, we discussed his reflections on the industry, the evolution of mobility data, and what’s next in his journey.
Over your nearly 10 years leading Huq, how have you seen the alternative data industry evolve? What’s been the most defining or memorable moment for you?
Over the last 10 years we’ve seen the development of more consistent engagement models and pricing levels however, I still think, we are in the foothills when it comes to exploring and extracting value from alternative data. The defining moment for me was our first 6 figure sale as it was an objective validation of the value and potential of our data.
Throughout your career, what were the 3 key skills you honed that elevated your performance most?
Keeping focused on the big issues — It's easy to get caught up in the day-to-day, but the ability to filter noise and stay laser-focused on what truly moves the needle is essential.
Identifying and attracting great people — The best ideas and strategies mean little without the right people to execute them. Identifying the right people to work with is, I think, a deeply human skill and one that I’m always trying to refine.
Building great culture requires constant effort — I want to work in environments where people feel motivated and empowered. Getting the balance between leading and empowering your team takes practice
Could you provide a couple of technical insights or innovations in recent mobility data that you believe would be particularly interesting to investment professionals?
One of the biggest historical challenges in using mobility data for investment decisions has been entity resolution—accurately mapping raw data points to real-world locations and, in turn, linking those locations to relevant investment entities. Recently, I’ve seen promising advances in AI-driven tagging and classification techniques. Refining these tools is key to improving the accuracy of location attribution, reducing false signals, and allowing investors to gain more reliable insights into foot traffic patterns.
How do you see the growing adoption of alternative data changing the relationship between financial data vendors and investment teams?
AI-driven tools are lowering the barriers to extracting insights from complex datasets, making alternative data more accessible to a broader range of investment professionals. I think this shift will likely lead to greater adoption of alternative data across all tiers of the investment landscape, not just hedge funds with dedicated data science teams. At the same time, as AI simplifies data extraction, there’s a risk of over-reliance on automated insights without proper due diligence. Investment teams will need to be increasingly discerning about how they validate and contextualise the data they use.
Aside from AI, what major trends do you see having the most profound impact on mobility data over the next 5, 10, and 25 years?
In the short to medium term, I expect the most profound developments in mobility data to come from advancements in predictive analytics. There’s growing research exploring the ability to forecast human movement at scale, which could unlock tremendous value across industries, enabling investors to anticipate patterns before they materialise and governments & urban planners better able to predict and manage infrastructure demands.
Now that you’re transitioning from Huq, what is the next major project or initiative you’re working on?
My focus is now on helping businesses identify and protect the unique data pools they are creating—many companies don’t even realise the proprietary value they hold. Beyond that, I work with data owners to maximise the commercial potential of their assets, whether through strategic partnerships, monetisation models, or operational optimisations.
Before we wrap up, is there anything else you’d like to share with our readers?
Feel free to connect with me on LinkedIn if you'd like to continue the conversation.
Numbers & Narratives
Fixed Income Hedge Funds: Why the Surge Makes Sense
Institutional investors are ramping up fixed income hedge fund allocations—62% plan to increase exposure in the next 12 months, per RBC Global Asset Management. The report highlights interest rate policies (58%) and geopolitical risks (60%) as the biggest drivers.
This aligns with what we’re seeing in the data. Central banks are navigating inflation without triggering liquidity crises, leading to sharp repricings across sovereign and corporate bond markets. The volatility in rate expectations creates persistent mispricings in swaps, repo markets, and structured credit—prime ground for hedge funds running systematic relative value strategies.
On the credit side, dispersion is increasing. 45% of investors cite improved market liquidity, but it’s uneven. Higher dispersion in CDS indices vs. cash bonds, along with structural changes in market-making, means there’s more room for arbitrage in credit derivatives and distressed debt. Selectivity is also rising—47% of institutional investors won’t consider funds under $100M AUM. That’s partly a function of execution constraints. Many of these strategies depend on tight spreads and high turnover, meaning scale matters.
The opportunity is clear: Investors are betting on funds that can navigate structural dislocations with precision. Quants with models tuned to evolving rate regimes and liquidity cycles will find this an optimal environment.
Market Pulse
Tariffs & Tumult
The latest escalation in trade tensions is forcing quants to rethink market structure, volatility regimes, and risk modeling. Trump's tariffs—10% on China, 25% on Canada, Mexico, and a potential 25% on EU imports—have triggered a broad market sell-off, with the Nasdaq dropping 2.8% and the Hang Seng plunging 3.3%, reflecting investor anxiety over global growth. The auto sector is under particular strain, as BMW (-2.6%) and Porsche (-2.4%) slid on fears of retaliatory EU tariffs, which could further disrupt supply chains already weakened by post-pandemic constraints. Meanwhile, Nvidia’s 8.4% decline, despite strong earnings, underscores investors shifting away from high-multiple tech stocks amid tightening financial conditions and rising import costs. The dollar’s 0.8% appreciation signals capital rotation into safe-haven assets, while Brent crude’s 1% decline reflects weakening demand forecasts. This combination of equity stress, currency divergence, and commodity softness creates an environment of volatility clustering, where correlation breakdowns and factor rotations become more unpredictable. For quants, this requires rethinking hedging strategies, especially as trade-driven inflation risks challenge traditional macroeconomic assumptions.
Confidence in Crisis
The resurgence of stagflation risks is forcing quants to rethink macro regime classifications and stress-test portfolios against a structurally different inflation-growth dynamic. U.S. consumer confidence plunged 7 points to 98.3 in February, its lowest in eight months, reflecting rising anxiety over economic stagnation and price pressures. Inflation remains anchored at 3%, but 12-month expectations have surged to 6%, the highest since May 2023, eroding real yields and pressuring rate-sensitive assets. Across the pond, UK consumer sentiment collapsed to a ten-month low as private-sector employment contracted for the third consecutive month, with job losses at their highest since 2021. Global sentiment remains fractured—Latin America shows resilience, but Europe is faltering, with Hungary’s confidence index at just 35.1, the second-lowest globally. As historical analogues to the 1970s gain relevance, quants must recalibrate risk models: mean-reversion assumptions are breaking down, risk-parity allocations are vulnerable to shifting cross-asset correlations, and vol-control strategies must adapt to higher macro uncertainty. Defensive positioning in gold, long-duration bonds, and inflation-linked assets is already reflecting this shift.
Navigational Nudges
Most signals don’t die overnight—they fade, decay, and get arbitraged away. The challenge isn’t just identifying alpha but knowing when it’s losing its edge. A high Sharpe ratio today doesn’t guarantee persistence tomorrow. Markets adapt, inefficiencies shrink, and staying ahead means recognising when a signal is deteriorating before it becomes a liability.
Here's how:
1. Eigenvalue Shrinkage and Signal Decay
Signals often blend into noise before completely dying. Instead of just tracking performance, decompose signals via PCA and monitor eigenvalue shrinkage in the covariance or correlation matrix. If a factor’s eigenvalue steadily declines relative to others, it’s losing explanatory power. A rising condition number suggests increasing multicollinearity—useful context but not a direct death signal. The real concern is whether your signal's variance contribution is fading into statistical noise.
2. Hidden Markov Models for Regime Shifts
Alpha isn’t static—it’s regime-dependent. Instead of relying on rolling Sharpe ratios, fit an HMM to track probability shifts between signal states. A spike in transition probability to a low-performance regime suggests structural decay, but there’s no universal “90% threshold” that confirms it. Hierarchical HMMs can model multi-layered shifts, but properly linking them to signal persistence is non-trivial.
3. Persistent Homology for Structural Breaks
Market inefficiencies have structure, and topological data analysis (TDA) can detect fragmentation in feature spaces. If Betti numbers increase or persistence diagrams destabilise in a well-defined representation, it could indicate a weakening signal. That said, applying TDA meaningfully to time series is complex—it’s an interesting research area but far from a standard approach for tracking signal decay.
4. Granger Causality and Signal Arbitrage
Alpha doesn’t “transfer” so much as become less exploitable. If your signal becomes Granger-caused by liquidity metrics like order book imbalance, it may be getting front-run. But Granger causality only implies statistical precedence, not true causation—there are plenty of confounding factors, from correlated signals to market structure effects. Alone, it’s a weak argument for signal obsolescence.
5. Multi-Armed Bandits for Dynamic Allocation
Deciding when to kill a signal is tough—automating it helps. Multi-armed bandits (MABs) can reallocate weight dynamically, but framing them as a “keep vs. kill” mechanism is an oversimplification. Most implementations gradually adjust exposure rather than outright cutting signals, and defining the right reward structure is far from trivial.
6. Bayesian Change Point Detection (BOCPD) as an Early Warning
Markets leave statistical fingerprints when signals degrade. BOCPD can flag shifts by monitoring a signal’s rolling t-stat, but interpreting a hazard function spike as “the market learned my signal” is speculative. It just tells you something has changed—whether that’s due to a competitor, macro factors, or liquidity shifts needs deeper analysis. GP priors help smooth noise but can also mask real breakpoints.
7. Quantum Annealing for Portfolio Optimisation
For multi-signal portfolios, classical optimisation struggles. Quantum annealing (or simulated annealing) can help solve combinatorial selection problems, but calling it a natural way for signals to “fall out” is a stretch. QUBO formulations require explicit obsolescence penalties—there’s no magic self-pruning mechanism. Plus, quantum methods are still experimental in finance, with real-world applications lagging behind theory.
Alpha isn’t forever, but adaptation is. Markets evolve, inefficiencies fade, and signals lose their edge. Tracking structural decay keeps strategies sharp and ensures you’re always making room for what’s next.
The Knowledge Buffet
by Mark Flemming-Williams
If you’re working with alternative data in your investments, The Alternative Data Podcast should be on your radar. Mark knows the space inside out—his experience at CFM and Exabel gives him real insight to exactly where alt data adds value. He asks the right questions pushing guests to show how their data can genuinely improve predictive models making it one of the most useful resources out there. Don’t say we didn’t tell you—expect to leave every episode with insights that stick and ideas you’ll actually use.
The Closing Bell
"What is most important isn’t knowing the future — it is knowing how to react appropriately to the information available at each point in time."
- Ray Dalio
The Kickoff
February may be short, but the markets didn’t hold back—signals aged, fixed-income strategies gained momentum, and shifts in data trends kept quants on their toes. As trends settle and cycles play out, the search for alpha continues. In this edition, we explore what’s holding firm, what’s fading, and what’s next as we step into March.
The Compass
Here's a rundown of what you can find in this edition:
Newest partner additions to the Quanted data lake.
From conferences to a crisp new website—what we’ve been up to
Why fixed-income strategies are gaining traction among hedge funds
Insights from our chat with Conrad Poulson, former CEO of Huq Industries
Round-up of the latest news impacting the markets
How to quantify the life-cycle of a quant signal before it becomes a liability
Why we love Mark Fleming-Williams' "The Alternative Data Podcast"
A resonating quote from Ray Dalio to take into March
Insider Trading
We’ve been making moves this month, both online and offline. Our new website is live—built to make it clearer and easier for both quants and data vendors to see exactly how Quanted fits into their workflows.
Meanwhile, we’ve been on the ground at BattleFin in Miami and Eagle Alpha in New York, catching up with partners and meeting new faces. One theme kept coming up in conversations: having data isn’t enough anymore. The real challenge (and opportunity) is how fast funds can find and test new data. The usual approach was agreed to be slow, expensive, and often a shot in the dark when it comes to whether the data will actually improve a model. Good thing there are alternatives out there ;)
Overall this month has been about building momentum—expanding our reach, refining our website, and continuing to deepen our data coverage. More updates are in the works, and we’re looking forward to keeping up the pace and delivering even more value to quants and data providers.
On the Radar
On the partnership front, we’ve added four new data partnerships to the lake, signing agreements with EDI, ETF Global, Intrinio, and SIX. This expands our range of data sources across market, fundamental, alternative, and economic data—bringing more breadth and depth to the features available for quants to test and identify which have the most impact on their models.
Provides corporate actions and pricing data with global coverage, offering accurate insights into dividends, mergers, and stock splits—critical for quants who require clean, timely data for model development.
Provides ETF analytics with proprietary risk and sentiment indicators, offering deeper insights into fund flows, factor exposures, and market trends beyond standard ETF datasets.
Offers a comprehensive suite of financial data APIs, including real-time and historical market data, fundamentals, and options pricing, facilitating seamless integration into financial models and applications.
Sources and consolidates market data from all over the world. From wealth management to media organisations, customers have been using SIX to gain a comprehensive view of global markets for over 90 years.
The Tradewinds
Expert Exchange
We recently sat down with Conrad Poulson, former CEO of Huq Industries, where he led the company in transforming mobility data into actionable intelligence for government, finance, retail, and real estate. Over nearly a decade, Conrad helped scale Huq into a leading provider of mobility insights, serving 300+ local councils, investors, and major corporations.
With a career spanning leadership roles at Orange, startup mentorship at Techstars, and pioneering mobile innovations, Conrad brings deep expertise in data-driven decision-making and high-growth leadership. As he moves on from Huq, we discussed his reflections on the industry, the evolution of mobility data, and what’s next in his journey.
Over your nearly 10 years leading Huq, how have you seen the alternative data industry evolve? What’s been the most defining or memorable moment for you?
Over the last 10 years we’ve seen the development of more consistent engagement models and pricing levels however, I still think, we are in the foothills when it comes to exploring and extracting value from alternative data. The defining moment for me was our first 6 figure sale as it was an objective validation of the value and potential of our data.
Throughout your career, what were the 3 key skills you honed that elevated your performance most?
Keeping focused on the big issues — It's easy to get caught up in the day-to-day, but the ability to filter noise and stay laser-focused on what truly moves the needle is essential.
Identifying and attracting great people — The best ideas and strategies mean little without the right people to execute them. Identifying the right people to work with is, I think, a deeply human skill and one that I’m always trying to refine.
Building great culture requires constant effort — I want to work in environments where people feel motivated and empowered. Getting the balance between leading and empowering your team takes practice
Could you provide a couple of technical insights or innovations in recent mobility data that you believe would be particularly interesting to investment professionals?
One of the biggest historical challenges in using mobility data for investment decisions has been entity resolution—accurately mapping raw data points to real-world locations and, in turn, linking those locations to relevant investment entities. Recently, I’ve seen promising advances in AI-driven tagging and classification techniques. Refining these tools is key to improving the accuracy of location attribution, reducing false signals, and allowing investors to gain more reliable insights into foot traffic patterns.
How do you see the growing adoption of alternative data changing the relationship between financial data vendors and investment teams?
AI-driven tools are lowering the barriers to extracting insights from complex datasets, making alternative data more accessible to a broader range of investment professionals. I think this shift will likely lead to greater adoption of alternative data across all tiers of the investment landscape, not just hedge funds with dedicated data science teams. At the same time, as AI simplifies data extraction, there’s a risk of over-reliance on automated insights without proper due diligence. Investment teams will need to be increasingly discerning about how they validate and contextualise the data they use.
Aside from AI, what major trends do you see having the most profound impact on mobility data over the next 5, 10, and 25 years?
In the short to medium term, I expect the most profound developments in mobility data to come from advancements in predictive analytics. There’s growing research exploring the ability to forecast human movement at scale, which could unlock tremendous value across industries, enabling investors to anticipate patterns before they materialise and governments & urban planners better able to predict and manage infrastructure demands.
Now that you’re transitioning from Huq, what is the next major project or initiative you’re working on?
My focus is now on helping businesses identify and protect the unique data pools they are creating—many companies don’t even realise the proprietary value they hold. Beyond that, I work with data owners to maximise the commercial potential of their assets, whether through strategic partnerships, monetisation models, or operational optimisations.
Before we wrap up, is there anything else you’d like to share with our readers?
Feel free to connect with me on LinkedIn if you'd like to continue the conversation.
Numbers & Narratives
Fixed Income Hedge Funds: Why the Surge Makes Sense
Institutional investors are ramping up fixed income hedge fund allocations—62% plan to increase exposure in the next 12 months, per RBC Global Asset Management. The report highlights interest rate policies (58%) and geopolitical risks (60%) as the biggest drivers.
This aligns with what we’re seeing in the data. Central banks are navigating inflation without triggering liquidity crises, leading to sharp repricings across sovereign and corporate bond markets. The volatility in rate expectations creates persistent mispricings in swaps, repo markets, and structured credit—prime ground for hedge funds running systematic relative value strategies.
On the credit side, dispersion is increasing. 45% of investors cite improved market liquidity, but it’s uneven. Higher dispersion in CDS indices vs. cash bonds, along with structural changes in market-making, means there’s more room for arbitrage in credit derivatives and distressed debt. Selectivity is also rising—47% of institutional investors won’t consider funds under $100M AUM. That’s partly a function of execution constraints. Many of these strategies depend on tight spreads and high turnover, meaning scale matters.
The opportunity is clear: Investors are betting on funds that can navigate structural dislocations with precision. Quants with models tuned to evolving rate regimes and liquidity cycles will find this an optimal environment.
Market Pulse
Tariffs & Tumult
The latest escalation in trade tensions is forcing quants to rethink market structure, volatility regimes, and risk modeling. Trump's tariffs—10% on China, 25% on Canada, Mexico, and a potential 25% on EU imports—have triggered a broad market sell-off, with the Nasdaq dropping 2.8% and the Hang Seng plunging 3.3%, reflecting investor anxiety over global growth. The auto sector is under particular strain, as BMW (-2.6%) and Porsche (-2.4%) slid on fears of retaliatory EU tariffs, which could further disrupt supply chains already weakened by post-pandemic constraints. Meanwhile, Nvidia’s 8.4% decline, despite strong earnings, underscores investors shifting away from high-multiple tech stocks amid tightening financial conditions and rising import costs. The dollar’s 0.8% appreciation signals capital rotation into safe-haven assets, while Brent crude’s 1% decline reflects weakening demand forecasts. This combination of equity stress, currency divergence, and commodity softness creates an environment of volatility clustering, where correlation breakdowns and factor rotations become more unpredictable. For quants, this requires rethinking hedging strategies, especially as trade-driven inflation risks challenge traditional macroeconomic assumptions.
Confidence in Crisis
The resurgence of stagflation risks is forcing quants to rethink macro regime classifications and stress-test portfolios against a structurally different inflation-growth dynamic. U.S. consumer confidence plunged 7 points to 98.3 in February, its lowest in eight months, reflecting rising anxiety over economic stagnation and price pressures. Inflation remains anchored at 3%, but 12-month expectations have surged to 6%, the highest since May 2023, eroding real yields and pressuring rate-sensitive assets. Across the pond, UK consumer sentiment collapsed to a ten-month low as private-sector employment contracted for the third consecutive month, with job losses at their highest since 2021. Global sentiment remains fractured—Latin America shows resilience, but Europe is faltering, with Hungary’s confidence index at just 35.1, the second-lowest globally. As historical analogues to the 1970s gain relevance, quants must recalibrate risk models: mean-reversion assumptions are breaking down, risk-parity allocations are vulnerable to shifting cross-asset correlations, and vol-control strategies must adapt to higher macro uncertainty. Defensive positioning in gold, long-duration bonds, and inflation-linked assets is already reflecting this shift.
Navigational Nudges
Most signals don’t die overnight—they fade, decay, and get arbitraged away. The challenge isn’t just identifying alpha but knowing when it’s losing its edge. A high Sharpe ratio today doesn’t guarantee persistence tomorrow. Markets adapt, inefficiencies shrink, and staying ahead means recognising when a signal is deteriorating before it becomes a liability.
Here's how:
1. Eigenvalue Shrinkage and Signal Decay
Signals often blend into noise before completely dying. Instead of just tracking performance, decompose signals via PCA and monitor eigenvalue shrinkage in the covariance or correlation matrix. If a factor’s eigenvalue steadily declines relative to others, it’s losing explanatory power. A rising condition number suggests increasing multicollinearity—useful context but not a direct death signal. The real concern is whether your signal's variance contribution is fading into statistical noise.
2. Hidden Markov Models for Regime Shifts
Alpha isn’t static—it’s regime-dependent. Instead of relying on rolling Sharpe ratios, fit an HMM to track probability shifts between signal states. A spike in transition probability to a low-performance regime suggests structural decay, but there’s no universal “90% threshold” that confirms it. Hierarchical HMMs can model multi-layered shifts, but properly linking them to signal persistence is non-trivial.
3. Persistent Homology for Structural Breaks
Market inefficiencies have structure, and topological data analysis (TDA) can detect fragmentation in feature spaces. If Betti numbers increase or persistence diagrams destabilise in a well-defined representation, it could indicate a weakening signal. That said, applying TDA meaningfully to time series is complex—it’s an interesting research area but far from a standard approach for tracking signal decay.
4. Granger Causality and Signal Arbitrage
Alpha doesn’t “transfer” so much as become less exploitable. If your signal becomes Granger-caused by liquidity metrics like order book imbalance, it may be getting front-run. But Granger causality only implies statistical precedence, not true causation—there are plenty of confounding factors, from correlated signals to market structure effects. Alone, it’s a weak argument for signal obsolescence.
5. Multi-Armed Bandits for Dynamic Allocation
Deciding when to kill a signal is tough—automating it helps. Multi-armed bandits (MABs) can reallocate weight dynamically, but framing them as a “keep vs. kill” mechanism is an oversimplification. Most implementations gradually adjust exposure rather than outright cutting signals, and defining the right reward structure is far from trivial.
6. Bayesian Change Point Detection (BOCPD) as an Early Warning
Markets leave statistical fingerprints when signals degrade. BOCPD can flag shifts by monitoring a signal’s rolling t-stat, but interpreting a hazard function spike as “the market learned my signal” is speculative. It just tells you something has changed—whether that’s due to a competitor, macro factors, or liquidity shifts needs deeper analysis. GP priors help smooth noise but can also mask real breakpoints.
7. Quantum Annealing for Portfolio Optimisation
For multi-signal portfolios, classical optimisation struggles. Quantum annealing (or simulated annealing) can help solve combinatorial selection problems, but calling it a natural way for signals to “fall out” is a stretch. QUBO formulations require explicit obsolescence penalties—there’s no magic self-pruning mechanism. Plus, quantum methods are still experimental in finance, with real-world applications lagging behind theory.
Alpha isn’t forever, but adaptation is. Markets evolve, inefficiencies fade, and signals lose their edge. Tracking structural decay keeps strategies sharp and ensures you’re always making room for what’s next.
The Knowledge Buffet
by Mark Flemming-Williams
If you’re working with alternative data in your investments, The Alternative Data Podcast should be on your radar. Mark knows the space inside out—his experience at CFM and Exabel gives him real insight to exactly where alt data adds value. He asks the right questions pushing guests to show how their data can genuinely improve predictive models making it one of the most useful resources out there. Don’t say we didn’t tell you—expect to leave every episode with insights that stick and ideas you’ll actually use.
The Closing Bell
"What is most important isn’t knowing the future — it is knowing how to react appropriately to the information available at each point in time."
- Ray Dalio
