Woosub Shin · Selected work

Systematic research, research infrastructure, and financial econometrics

The portfolio separates platform engineering, retained baseline research, post-selection challengers, and academic work so that each result keeps the evidence label it actually earned.

Program sequence

One research program, five layers

Each system retains its own provenance, research question, methods, result scope, and limitations. Strong historical performance does not erase the distinction between selection, validation, and prospective evidence.

01

Research infrastructure

Flagship research platform

Multi-Asset Research Lab

How should a multi-asset systematic research framework move from data provenance to a retained research system without allowing evidence to silently become execution authority?

An asset-neutral research foundation with immutable contracts, canonical evidence identity, deterministic replay and PnL accounting, frozen search, validation, and explicit historical-to-forward boundaries.

Immutable contractsReplay / PnLFrozen searchDeep validationForward evidence

02

Systematic strategy research

Retained baseline · forward research

BTC Final Research System V1

Can one simple BTC trend-following research system survive frozen search and deep validation, then cross into a prospective forward clock without turning research state into trading permission?

A retained BTCUSDT Daily EMA 50/200 long/flat research system with deterministic replay, funding-adjusted accounting, frozen final-system search, thirteen-gate deep validation, and append-only prospective forward observation.

Daily EMA 50/200Funding-adjusted PnLCost stressDeep validationForward runtime

03

Systematic strategy research

Historical challenger · not independent OOS

BTC Selective Regime Challenger C4

Can a sparse, mutually exclusive regime-conditioned long/flat system improve historical risk-adjusted performance without hiding concentration, cost sensitivity, or post-selection bias?

A post-selection historical stress audit of a sparse BTC regime router combining Momentum90, EMA 50/180, and RSI2/EMA200 sleeves. C4 returned +182.29% with 1.100 Sharpe at 5bp per transition side across the available regime period, with 88 completed trades.

Regime conditioningMomentum90EMA 50/180RSI2 + EMA20020bp cost stressBlock resampling

04

Financial econometrics

Academic foundation

Volatility Regime Filtering in Futures Markets

Can volatility-regime filtering improve the discipline of an intraday NQ, ES, and Crude Oil (CL) futures framework without treating EGARCH as a direction predictor?

An academic comparison across NQ, ES, and Crude Oil (CL) futures using an EGARCH-conditioned framework with otherwise identical intraday logic, treating volatility regime as a risk and admissibility layer.

NQESCrude Oil (CL)EGARCH5-minute dataRobustness

05

Crypto-asset diagnostics

Earlier time-series research

Bitcoin Bubble Detection with GSADF

How can GSADF testing identify statistically explosive Bitcoin price episodes while keeping diagnostic evidence separate from market recommendations?

A compact time-series study applying right-tailed explosive-root diagnostics to identify and interpret periods of explosive Bitcoin price behavior.

BitcoinGSADFExplosive rootsTime series

Common standard

What travels across the program

Information is aligned to when it was knowable.
Model roles stay narrower than strategy claims.
Search, stress testing, and prospective observation remain separate stages.
Provenance, cost assumptions, and limitations travel with every result.