PRISM
Alpha
Progressive Risk-Informed Signal Model — a next-generation multi-factor strategy that combines static reliability with adaptive intelligence.
The Investment Concept in 60 Seconds
Our model scores every stock each month on three dimensions, then automatically adjusts which dimensions matter most based on recent market evidence.
Score Every Stock
Each month we score all 50 stocks on Momentum (price trend), Value (cheap vs expensive), and Quality (strong financials).
Weigh What Works
The model checks which scoring method predicted returns best recently. Factors that worked get more weight.
Buy Top / Sell Bottom
Stocks with the highest combined score get bought. Those with the lowest get sold short. Monthly rebalance.
Adapt Over Time
As market conditions change, the model smoothly shifts emphasis. Value-heavy in 2023, then Momentum picks up in 2025.
Think of It Like a Hiring Committee
Imagine three interviewers (Momentum, Value, Quality) evaluating 50 job candidates (stocks) every month. The committee chair (our IC model) tracks which interviewer's picks actually succeed on the job (generate returns). Over time, the chair gives more voting power to the interviewer with the best track record. The result: a team (portfolio) that continuously adapts to find the best candidates.
What We Invest In
We invest in 50 of America's largest and most liquid companies across 10 sectors. These are household names — the backbone of the S&P 500.
Equity Curve - PRISM Alpha vs Benchmark
Backtest period: Jan 2020 - Feb 2025. Universe: 50 US large-cap equities. $1M initial capital, monthly rebalance, 5bps commission + 5bps slippage.
Cumulative Returns (Long-Only)
+93.68% totalMonthly Returns - PRISM Alpha
42.7% win rateWhat We Buy and Sell Each Month
Every month the model rebalances the portfolio: buying stocks that score highest and selling those that fall in ranking.
Trade Heatmap - Positions by Month
Three-Phase Adaptive Engine
PRISM Alpha eliminates the traditional warmup penalty while preserving the power of adaptive factor allocation. It trades from day one.
Static Warmup
During early trading days when insufficient data exists for IC computation, the strategy deploys proven static factor weights (Momentum 40%, Value 30%, Quality 30%).
Smooth Blend Transition
Once IC data becomes available, the system smoothly transitions from static to dynamic weights over 126 trading days using linear interpolation.
Full IC Adaptive
Factors are weighted proportional to their rolling Information Coefficient — the Spearman rank correlation between factor scores and realized forward returns.
Strategy Performance Matrix
Six strategies benchmarked on the same universe and backtest period. PRISM Alpha dominates across both Long-Only and Long/Short variants.
| Strategy | Total Return | Ann. Return | Sharpe | Sortino | Calmar | Max DD | Alpha vs SPY | Win Rate |
|---|---|---|---|---|---|---|---|---|
| PRISM Alpha Long | 93.68% | 13.72% | 0.593 | 0.757 | 0.657 | -20.86% | +1.26% | 42.7% |
| Static Long (40/30/30) | 77.77% | 11.84% | 0.407 | 0.511 | 0.575 | -20.57% | -0.62% | 41.8% |
| Pure Dynamic Long | 57.90% | 9.29% | 0.332 | 0.369 | 0.467 | -19.89% | -3.17% | 32.0% |
| PRISM Alpha L/S | 72.02% | 11.12% | 0.447 | 0.570 | 0.522 | -21.29% | -1.33% | 41.7% |
| Static L/S | 62.46% | 9.90% | 0.362 | 0.459 | 0.463 | -21.39% | -2.56% | 41.3% |
| Pure Dynamic L/S | 43.16% | 7.23% | 0.181 | 0.197 | 0.351 | -20.57% | -5.23% | 30.9% |
Information Coefficient Analysis
The IC measures each factor's predictive power. Value is the dominant alpha source with the highest IC and consistency (IC_IR = 0.35).
Negative IC - weight reduced to minimum 5%. Momentum reversal regime in 2022-2024.
Strongest and most consistent factor. EPS yield + margins from real fundamentals. Weight up to 70%.
Moderate signal using ROE, ROA, Gross Margin, Debt/Equity. Provides diversification benefit.
IC Lookback Sensitivity - PRISM Alpha
252d optimalProduction-Grade Technology Stack
Massive.com API
Enterprise-grade market data. Price history, financial statements, and corporate actions for 50+ tickers.
Point-in-Time Fundamentals
Real EPS, ROE, ROA, Margins — forward-filled from report dates. Zero look-ahead bias.
Event-Driven Backtester
Custom engine with position tracking, trade logging, transaction costs, and dollar-neutral L/S support.
Spearman IC Framework
Rolling rank correlation with constrained weight optimization. Min 5%, Max 70%, 10-iteration convergence.
Ready to Explore Further?
This is PRISM Alpha — our Progressive Risk-Informed Signal Model combining static reliability with adaptive intelligence. We are actively developing Macro Regime Overlay, Sector Neutralization, and Walk-Forward Optimization.