Research
We study financial markets with quantitative, statistical and computational methods, and publish selected work with the code and data needed to check it.
- Statistical modeling
- Time-series methods, inference and model validation
- Market structure
- Liquidity, execution, price formation and market behavior
- Algorithmic systems
- Rule-based and data-driven methods for trading decisions
- Computational methods
- Simulation, numerical methods and optimization
- Research infrastructure
- Data pipelines, reproducible workflows and tooling
- Model
- Xₜ = 0.6 Xₜ₋₁ + εₜ
- Theory
- ρ(k) = φᵏ
- Band
- ±1.96/√n = ±0.098
- Result
- 4 of 30 lags outside band
Publications
No publications yet. New work will be listed here with its date, authors, abstract and links, and code will be released on GitHub (opens in a new tab).
FormatsPapers · Preprints · Technical reports · Research notes · Experiments · Methodology · Market studies
How the research is done
Testable claims
A hypothesis is only as good as the test that could have rejected it.
Explicit assumptions
Each model states what it assumes about data, costs and the process that generated them.
Reproducible results
Results are tied to versioned code, fixed data snapshots and recorded parameters.
Out-of-sample validation
Conclusions follow out-of-sample tests, multiple-comparison control and honest error bars.
Some research stays private while it is developed and validated, and some is used in KSAN RELAY’s own trading. Published research is not investment advice.