QUANTITATIVE RESEARCH, WITHOUT THE SALES PITCH

Find the edge.
Then try to break it.

Evolve, test, and interrogate trading strategies in one rigorous workbench. Because a return means nothing until it survives the questions.

Explore the workbench pip install sophostechne-quant
SCROLL TO INTERROGATEENGINE · WORKBENCH · DATA
01 / THE METHOD

A backtest that reports
only a return is a sales pitch.

Quant answers the uncomfortable questions by default. No flags to hide the awkward two. No synthetic history dressed up as conviction.

01

Evolve

Search a population of strategies, scored in-sample and on bars the search never saw.

02

Interrogate

Place observed drawdown inside a Monte Carlo distribution built from the strategy’s own trades.

03

Verify

Walk the entire method forward and compare every survivor against a same-market noise floor.

02 / GLOBAL MARKETS

One market.
Many clocks.

Daily movement across twelve regional equity proxies, mapped to the markets they represent.

SPYS&P 500 proxy
···NEW YORK
READING PUBLISHED BARSIEX · HISTORY ONLY
DRAG TO ROTATE · SCROLL OR PINCH TO ZOOM

Regional ETFs are used as investable market proxies. Bars are published IEX history, not a consolidated real-time tape.

03 / THE WORKBENCH

Every result.
In context.

BTC-USD · 6H · EVOLUTION 20/20ANALYSIS COMPLETE
STRATEGY / EMA + ATR

Validation report

OUT-OF-SAMPLE+18.42%
MAX DRAWDOWN−7.16%
PROFIT FACTOR1.84
NOISE FLOORBEAT
EQUITY CURVE
TRAIN HOLDOUT
UNSEEN BARS →
04 / BUILT ON PRINCIPLE
0fabricated bars

Real history,
or an honest absence.

Every chart names its source. Gaps stay gaps. Costs are charged on every path. Reproducible by seed, composable by JSON, and designed to make overfitting visible.

OPEN SOURCE · MIT LICENSED

Your strategy has a story.
Make the data prove it.