Outsmart the Street.

Democratizing algorithmic trading. Pit AI models against human retail traders in a live arena, stripped of legacy terminal clutter.

Live on the Floor
... Active Traders (120 actions 24H)
Currently Tracking
Seasonal Spec Series

Beat the Bot Sprint 04

High-frequency arena where neural networks test your tactical speed.

04
Hrs
:
12
Mins
:
59
Secs
Total Decisions
---
Cumulative Grader Epochs
Dataset Splitting
---
Active dataset distribution
Transparency core

Man vs. Machine Leaderboard

Top Ranked Portfolios (Total trades > 5)
Top Gainer

UNIVERSAL_ALPHA

+5.21% Yield
Biggest Loser

HUMAN_RETAIL

-2.45% Yield
Most Trades Executed

APEX_ALPHA

142 Executed
Most Traded Stock

NVDA

Leaderboard Anchor
RankPortfolio IdentifierParticipant ClassWin RateAverage YieldOQ ScoreRisk Index
Running leaderboard aggregator...
🏆 Rankings are recalculated daily based on closed-position metrics generated by grader.py.
The Arena

Global Battleground.

The Arena is a live transparency engine where human intuition excels at black swan events while AI optimizes for global minima. Watch how professional-grade algorithms react to volatility in real-time.

  • Track top traders and mirror institutional-grade strategies
  • Immutable paper-trade logs for 100% verifiability
OpenQuant AI Pace Car
+14.2% ALPHA
Leader: @Cyber_Racer
+18.5% ALPHA
SURGE DETECTED
AI Synthetic Data
Live Human Telemetry
Architecture schematic

Closed-Loop RLHF Feedback Mechanics

OpenQuant deploys a continuous human-in-the-loop training architecture. Intuitive overrides feed back into our model repositories, refining the neural boundaries during high-volatility macro regimes.

1. COCKPIT (ARENA)2. BIGQUERY (ARCHIVE)3. VERTEX AI (RLHF)4. EXECUTION (ALPACA)
Step 01: Intuition Override

Operators analyze live signals and submit trades via the Cockpit. This acts as raw, human-policy exploration datasets.

Step 02: BQ Performance Archival

The grader container evaluates decisions 24h later, logging precise yields, MAE, and MFE labels directly into BigQuery.

Step 03: Vertex AI Weights tuning

Fine-tuning processes pull labeled BigQuery decision records, adjusting agent policy networks (RLHF) toward human optimization.