Methodology

How it works

Short answer: applied quantitative finance. Long answer below.

What we draw from

We use the standard quant toolkit — not because we're showing off the textbook, but because these are the tools that actually hold up when real money is on the line:

Time-series econometrics — ARIMA, GARCH for volatility clustering, cointegration for pairs.
Regime detection — Markov chains, hidden Markov models, change-point detection. Markets aren't one thing all the time. Treating them like one is how strategies blow up.
Stochastic processes — Brownian motion, jump diffusion, mean-reversion. The math behind why prices do what they do.
Statistical inference — bootstrap confidence intervals, walk-forward validation, out-of-sample testing. The boring discipline that separates “I have an idea” from “I have an edge.”
Information-theoretic measures — entropy, mutual information, transfer entropy. Where real signals hide.
Portfolio theory — mean-variance, Kelly, risk parity. For the day someone needs sizing solved.

A formula or two, so you know we mean it

Sharpe ratio

S = E[R_p − R_f] / σ_p

Ornstein–Uhlenbeck mean reversion

dX_t = θ (μ − X_t) dt + σ dW_t

Shannon entropy

H(X) = − Σ p(x_i) · log p(x_i)

These are standard. We won't pretend we invented them. What we do invent is how we glue them together for a specific problem — and that part stays between us and the client.

How we work with you

You describe the problem or the idea.
We tell you whether it's tractable, what data it needs, and roughly how long.
If it's a fit, we build it, ship it, and document it. You own the result.

No retainers for nothing. No “discovery phase” that bills six figures. We're engineers; we'd rather be writing code.