Welcome back to Concretum Pills, your weekly selection of ideas from our research archives.
This week’s edition revolves around the persistence of trend across markets and time horizons. From LLM-enhanced momentum selection to a twenty-year revisit of a classic trend-following study, with a stop in crypto markets along the way, the papers below explore how trends emerge, persist, and occasionally break violently. We also include two complementary readings: one on the economics of tail-risk hedging, and another on the growing evidence that even intraday momentum may be a structural feature of modern futures markets rather than a temporary anomaly.
Catching Crypto Trends: A Tactical Approach for Bitcoin and Altcoins
Zarattini, Pagani, Barbon
“We propose an ensemble approach that aggregates multiple Donchian channel-based trend models, each calibrated with different lookback periods, into a single signal.”
Trend-following has documented track records stretching back decades across traditional futures markets. In this paper, we ask whether that same edge carries over to crypto, applying the framework first to Bitcoin alone and then to a comprehensive, survivorship-bias-free dataset covering all cryptocurrencies traded since 2015.
ChatGPT in Systematic Investing: Enhancing Risk-Adjusted Returns with LLMs
Anic, Barbon, Seiz, Zarattini
“We combine daily U.S. equity returns for S&P 500 constituents with high-frequency news data and use prompt-engineered queries to ChatGPT that inform the model when a stock is about to enter a momentum portfolio.”
If investors process and act upon news gradually, an LLM capable of interpreting it in real time should be able to identify which momentum candidates are most likely to outperform. The article tests exactly this idea: using ChatGPT to filter which stocks deserve a place in a cross-sectional momentum portfolio and which ones are riding stale news.
Does Trend-Following Still Work on Stocks?
Zarattini, Pagani, Wilcox
“Our results confirm a highly skewed profit distribution, with less than 7% of trades driving the cumulative profitability.”
In 2005, Cole Wilcox and Eric Crittenden published an influential paper asking whether trend-following works on stocks. Twenty years later, using a survivorship-bias-free dataset of more than 66,000 simulated trades from 1950 to 2024, we revisit and update their framework. The finding that a tiny fraction of trades accounts for nearly all the profit has deep implications for position sizing, diversification, and the discipline required to let winners run.
Tail Risk Hedging: Contrasting Put and Trend Strategies
Ilmanen, Thapar, Tummala, Villalon — Journal of Systematic Investing, 2021
“The common view that Put costs more but is a more effective tail hedge contains a kernel of truth but does not capture the full story. In the end, the cost advantage favors Trend over Put.”
Buying puts feels like insurance: you pay a known premium and get a defined payout if the market crashes. Trend-following feels less precise, yet over long horizons it has delivered positive total returns while still providing meaningful crisis protection. The paper from AQR walks through the data and shows that the cumulative cost of rolling put protection can be large enough to overwhelm the payoff advantage in all but the sharpest drawdowns.
Hedging Demand and Market Intraday Momentum
Baltussen, Da, Lammers, Martens — Journal of Financial Economics, 2021
“Using intraday returns on over 60 futures on equities, bonds, commodities, and currencies between 1974 and 2020, we find strong market intraday momentum everywhere.”
Before this paper, academia largely treated intraday momentum as a localized curiosity, something mostly confined to U.S. equities during a specific historical period. After this study, that interpretation became much harder to defend. By examining more than 60 futures markets across equities, bonds, commodities, and currencies over nearly five decades, the authors show that the effect appears consistently across asset classes, geographies, and time. When a pattern survives across such different market structures and environments, it starts looking less like a temporary anomaly and more like a structural feature of how modern financial markets absorb information and clear trading imbalances.
That’s it for this week. More next Tuesday!
If you’ve got a paper, a book or an interview you think we should read, reach out at info@concretumgroup.com.
Questions and comments are always welcome.
And if you find our research useful, share our page with a friend.
