Possibility, built into probability.
From possibility to probability
At Quant Wave Capital we start from the data and let probability decide. Our process turns intuition into evidence through disciplined, repeatable analysis of U.S. equities.
Mid-frequency statistical arbitrage, built differently
Quant Wave Capital's approach is closest to statistical arbitrage — positions typically held days to weeks, built around mispriced individual stocks.
| Type | Holding period | Where the edge comes from |
|---|---|---|
| High-frequency / market making | Seconds | Speed and execution |
| Statistical arbitrage | Days to weeks | Mispriced individual stocks |
| Factor / smart beta | Months | Persistent risk premia |
| Managed futures (CTA) | Weeks to months | Trends across asset classes |
| Systematic macro | Months | Views on rates, currencies, countries |
Quant Wave Capital's approach is closest to statistical arbitrage, highlighted above. Holding periods are typical, not strict.
Small enough for the MidCap 400
We can trade mid-caps where very large managers struggle to build or exit positions without moving the price.
Concentrated, high conviction
A focused book of high-conviction positions rather than thousands of small bets.
Long-biased, no leverage
Many statistical arbitrage approaches run leveraged, market-neutral books. Ours is long-biased and does not use leverage to invest.
A tail-risk layer
Extreme value theory and a cascade generator help us estimate how losses could spread across positions under stress.
Five steps from raw data to a deployed rule
Identify variables
ML models — Random Forest, decision trees — identify the variables with real predictive power.
Detect the signal
Detect when price enters a relevant signal zone and measure the size of the deviation.
Historical analysis
Assess the signal's rarity and frequency, and the probability of a return across different horizons.
Optimize zones
A variable matrix optimizes the zones with the highest probability of success.
Stress testing
Historical and AI-generated synthetic scenarios. Models that fail these tests are not deployed.
"Most of the models we build never make it. Only the ones that hold up out of sample and across market regimes remain."
Overfitting
Models that memorize history instead of learning from it.
Narratives
Story-driven reasoning with no probabilistic basis.
Hunches
Decisions with no basis in data or reproducible logic.
Regime breaks
Models that fail under different market conditions.
Built to survive the bad days
Many quant strategies that failed in past crises did not have bad ideas. They had borrowed money, capital that could leave overnight, and models that assumed a really bad day was impossible. We designed our approach around those lessons. No approach eliminates risk.
We do not use leverage to invest
Our strategy does not rely on borrowed money to amplify positions. In the August 2007 quant unwind, highly leveraged strategies were forced to sell into falling markets.
Liquidity matched to the strategy
We believe investor liquidity should match the holding period of the strategy, to reduce the risk of having to sell positions at the worst moment.
We assume things fall together
Many risk models assume losses arrive one at a time. Ours is built to model them arriving together, as they often do in a crisis.
The process, not the person
Every model is documented, every change is version-controlled, and risk is reviewed independently of the signals. Nothing important lives only in one person's head.
Research, management and partners
Quant Wave Capital draws on the research platform of Quantum Wave S.A. in Santiago, Chile.
Luis Felipe Galleguillos
MSc in Economics and Business and a Diploma in AI from PUC Chile. Serves on investment fund supervisory committees. More than 16 years of experience at Grupo Security and Vida Security across trading, finance, and equity research.
Hans Lembach
Mathematical civil engineer from UTFSM with a Diploma in Machine Learning from PUC Chile. Former Staff Data Scientist at EY, with prior advanced analytics roles at Vida Security and Lemonpot.
Clemente Ferrer
Mathematical engineer and MSc in Mathematics from UTFSM; PhD candidate in Statistics at PUC Chile on an ANID scholarship. Publishes on extreme value theory and neural models.
Diego Rojas
Civil engineer from Universidad de Chile with an MSc in Data Science. Background across a range of machine learning projects.
Rodrigo Guzmán
MSc in Economics and Business from PUC Chile. Over thirty years as a CFO in the financial industry, including Head of Finance at Grupo Security; former director of Vida Security.
Aaron Dujovne
Business management and finance degree from Purdue University. Real estate asset manager in New York and co-founder of Casa Lotos Sotol. Leads business development.
Our edge is discipline.
Freedom to test any idea
No approach is ruled out: every thesis is welcome and the data decides what enters the portfolio.
We don't pick one trade: we measure them all
We measure and rank the alternatives, and act on the evidence rather than the hunch.
Years of being wrong on purpose
We didn't find what works by luck: we discarded everything that doesn't.
Obsessed with alpha
We don't want to be the biggest. We want to be the most precise.
Preparing for the next critical event
We measure tail risk position by position and run the portfolio through simulated crisis scenarios.
Get in touch
For general inquiries about Quant Wave Capital, please contact us directly.