EPSILON v2.0 — Quantitative Decision Lab

EPSILON v2.0 — Quantitative Decision Lab

EPSILON has evolved from a desktop trading simulator into an open-source quantitative decision lab for falsifiable market research.

The new product is organized around one disciplined loop:

Observe → Define → Test → Challenge → Retest

Instead of presenting a backtest number in isolation, EPSILON keeps the question, configuration, metrics, trade evidence, provenance boundaries, and failure condition attached to the same research artifact.

Why v2.0

Backtests can look persuasive after the assumptions that produced them have disappeared from view. EPSILON v2.0 is designed around a narrower question: how can a market idea remain challengeable throughout the research process?

This is a major product and identity transition from the v1 desktop simulator, so the original v1.0.0 release remains preserved as part of the project's history.

What is new

• A unified web product with one primary path: Market → Strategy Lab → Interrogate
• Guest access with no required account for the core experience
• Explicit hypotheses and falsification conditions before testing
• Backtest configuration, metrics, trades, and evidence boundaries shown together
• Stale-evidence handling when the subject or research question changes
• A challenge workflow for inspecting assumptions and refining the next test
• Clear separation between controlled synthetic evidence and claims requiring real historical validation
• A rebuilt public landing page, dashboard, navigation system, and mobile-responsive experience
• Open-source product documentation, architecture notes, screenshots, and a public launch kit

Evidence contract

EPSILON deliberately does not claim live market data, historical validity, profitability, statistical significance, or investment advice unless the underlying evidence supports those claims. Unknown inputs remain labeled as unknown.

Guest-mode results are controlled browser-local simulations. They are useful for demonstrating the research workflow, not for establishing market performance.

Explore

• Live product: https://epsilon-livid.vercel.app/landing
• Open the workspace: https://epsilon-livid.vercel.app/dashboard
• Source and documentation: https://github.com/DresdenGman/EPSILON-trading-simulator
• 30-second product film: attached to this release

From v1.0.0

The original desktop trading simulator remains in this repository and its release history is untouched. v2.0 adds the Decision Lab web experience and reframes the project around reproducible, falsifiable research rather than simulated trading alone.

Responsible use

EPSILON is an educational and research tool. It is not a broker, trading service, or source of investment advice. Verify data, assumptions, execution models, and statistical methods independently before drawing real-world conclusions.

v1.0.0 - Initial Release

v1.0.0 - Initial Release

Features

• ✅ Core trading system (Buy/Sell operations)
• ✅ Portfolio management
• ✅ Equity curve & performance metrics
• ✅ Limit orders & stop-loss/take-profit
• ✅ K-line charts with volume
• ✅ Auto trading rules (stop-loss, scale in/out)

Installationh

pip install -r requirements.txt
python mock.py### Notes

• Supports both real market data (akshare) and mock data
• Requires Python 3.7+