Quant Model Center
5-category factor library · 6-stage AI quant pipeline · Full scientific metrics (IC/ICIR/Sharpe/Sortino/Calmar/Alpha) · Free daily optimization for users · $10,000/mo developer subscription for full build.
Quant Model Center
Users get free daily log replay and one-per…
Users get free daily log replay and one-per-day factor optimization. Developers unlock AI-assisted strategy generation, portfolio construction, backtesting, and publishing with a $10,000/month subscription.
User Lane (Free)
Log replay → factor opt…
Log replay → factor optimization preview → confirm → apply in real time.
- One free optimization per strategy / portfolio per day
- Preview impact before any parameter update
- Full IC / ICIR / Sharpe / Sortino / Alpha metrics suite
Developer Lane ($10,000/mo)
Subscription unlocks AI…
Subscription unlocks AI-assisted strategy build, portfolio assembly, full backtesting, and marketplace publishing.
- AI-assisted single & multi-strategy portfolio generation
- Full scientific backtest: Sharpe / Sortino / Calmar / Alpha / IC / ICIR
- Publish to marketplace and earn profit share after threshold checks
Factor Library
Factor Library
Five factor categories covering the compl…
Five factor categories covering the complete quantitative research chain from signal generation to risk governance.
Price Momentum
IC 0.03–0.1212-1 month cross-sectional momentum score
Earnings Momentum
IC 0.04–0.15EPS revision direction and magnitude
Short-Term Reversal
IC -0.02–0.081-month short-term price reversal
Earnings Stability
IC 0.02–0.10Inverse EPS volatility over 5 years
Return on Equity
IC 0.03–0.11Net income / shareholders' equity
Balance Sheet Quality
IC 0.02–0.09Accruals ratio (low accruals = high quality)
Realized Volatility
IC -0.04–0.0620-day realized vol — low-vol premium
IV Skew
IC 0.01–0.08Put/call IV spread — options market signal
Earnings Surprise
IC 0.05–0.18Actual EPS vs. consensus forecast deviation
News Sentiment
IC 0.02–0.12NLP-scored sentiment from company news flow
Order Flow Imbalance
IC 0.03–0.14Net buy/sell order flow imbalance ratio
Liquidity Score
IC 0.01–0.07Amihud illiquidity ratio (lower = more liquid)
Quant Model Pipeline
Quant Model Pipeline
Data ingestion → feature engineering → si…
Data ingestion → feature engineering → signal generation → portfolio construction → risk management → execution, with AI modules at each stage.
Data Ingestion
OHLCV + alt-data + macro signals — real-time ingestion and cleaning
AI
Anomaly detection + auto imputation
Feature Engineering
Factor normalization, cross-sectional z-scoring, sector neutralization
AI
Automated feature selection and combination
Signal Generation
Multi-model ensemble inference — ranked opportunity queue
AI
RL dynamic weighting + ensemble voting
Portfolio Construction
Mean-variance / risk parity / Black-Litterman — selectable modes
AI
AI auto-selects optimal optimizer regime
Risk Management
VaR / CVaR computation, drawdown caps, position limit enforcement
AI
Real-time risk factor monitoring + auto-hedge trigger
Execution
TWAP / VWAP smart routing, market impact and slippage modeling
AI
Optimal execution timing prediction
Scientific Evaluation Framework
Scientific Evaluation Framework
Industrial-grade performance metrics powe…
Industrial-grade performance metrics powering model training and selection.
Metric
Formula
Good Threshold
Usage
Information Coefficient (IC)
Spearman(Factor, Fwd Return)
> 0.05 credible, > 0.10 significant
Core factor efficacy signal
IC Information Ratio (ICIR)
Mean(IC) / Std(IC)
> 0.5 stable, > 1.0 high-quality
Signal consistency over time
Sharpe Ratio
(Rp − Rf) / σ
> 1.0 good, > 2.0 excellent
Risk-adjusted return assessment
Sortino Ratio
(Rp − Rf) / σ_downside
> 1.2 good, > 2.0 excellent
Penalizes only downside vol — better for asymmetric risk prefs
Calmar Ratio
Ann. Return / Max Drawdown
> 1.0 good, > 2.0 excellent
Annualized return per unit of drawdown risk
Alpha (α)
Rp − (Rf + β·(Rm−Rf))
> 0% beats benchmark, higher is better
Excess return above benchmark — skill attribution
Beta (β)
Cov(Rp, Rm) / Var(Rm)
0.3–0.8 low-vol, ≈1 = market-neutral
Portfolio sensitivity to systematic market risk
Max Drawdown
max(Peak − Trough) / Peak
< 5% excellent, < 15% acceptable
Worst-case capital loss scenario for the portfolio
AI Training Directions
AI Training Directions
Continuous multi-paradigm AI training pip…
Continuous multi-paradigm AI training pipelines anchored to the factor library and scientific metrics framework.
Supervised Factor Scoring
Train gradient-boosting and deep-learning models on N-day forward returns to score and rank factor sets for relative return prediction.
RL Portfolio Rebalancing
Use Sharpe / Calmar as reward signal to train a policy network that learns optimal rebalancing timing and position sizing across market regimes.
Transfer Learning Cross-Market
Pre-train on liquid markets (e.g., US equities) and transfer to data-scarce venues (prediction markets, small-caps) to boost signal quality.
Meta-Learning Regime Adaptation
Train meta-models that adapt within hours of market regime shifts, addressing the stale-model problem in traditional quant approaches.
NLP News & Sentiment Signals
LLM-based real-time sentiment scoring of company announcements, analyst reports, and news flow — feeds high-frequency event factors into the signal layer.
Automated Factor Mining
Genetic programming and symbolic regression to auto-discover new alpha factor expressions, filtered by IC / ICIR backtests with overfitting penalties.
Capability Matrix
Capability Matrix
One interface, role-gated capability dept…
One interface, role-gated capability depth and workflow guardrails.
Capability
User (Free)
Developer ($10,000/mo)
Copy-Trade Log Replay
Full replay with Agent event stream
Includes attribution and comparison
Factor Optimization
1 free optimization per strategy/portfolio per day
Unlimited, all-factor-dimension tuning
Impact Preview
Preview return / drawdown delta first
Full Sharpe / Alpha / IC impact preview
Strategy Blueprint Generation
—
AI-assisted single & portfolio generation
Scientific Backtest Metrics
—
Full Sharpe / Sortino / Calmar / Alpha / Beta / IC / ICIR suite
Marketplace Publishing
Personal use only
Publish to marketplace after eligibility check
Smooth Workflows
User: Log Replay + Daily Optimization
Smooth Workflows