The Iridium Alpha-Seeker
Manifesto.
Our objective is to systematically outperform standard market indices by identifying structural Superperformers.
We remove human emotion and rely on an analytical method that fuses Fundamental Growth (why to buy), Technical Trend health (what to buy), and exact Entry Timing (when to buy).
1The 5-Step Funnel (Current Pipeline)
1. Universe Selection
Automated ingestion of US & EU equities via Financial Modeling Prep (FMP). Currently, we have imported 55 tickers with a 2-year price history due to the limitations of the free API key. With the paid Premium tier, the universe will scale to thousands of tickers filtered for institutional-grade liquidity (ADV × Price > $2M).
2. SEPA Filter (Trend Template)
Rigid Phase 2 Uptrend verification per Minervini's rules: Price > SMA 50 > SMA 150 > SMA 200, with SMA 200 rising for at least 1 month. The stock must be within 25% of its 52-week high and at least 30% above its 52-week low. If it fails the template, it's instantly marked with a red X.
3. CAN SLIM Analysis
Focus on 'C' (Current Quarterly EPS > 25% YoY), 'A' (Annual Growth stability), and 'L' (Relative Strength Rating > 85, top 1% of the market). For EU stocks lacking quarterly granularity, we apply a graceful fallback with Sales Growth > 20% and ROE > 15%.
Note: The current demo does not use the FMP Premium subscription. Quarterly earnings data is synthetic and provided for demonstration purposes. Live Q-o-Q earnings ingestion will be activated with the production API tier.
4. VCP Engine (Volatility Contraction)
Proprietary detection of Institutional Footprints on the daily chart. The algorithm scans the last 30 candles, splits them into three 10-day blocks, and verifies that the Coefficient of Variation contracts sequentially (CV₁ > CV₂ > CV₃, with CV₃ < 0.04). It also checks for Volume Dry-up below the 50-day SMA. If confirmed, a Pivot Point and Stop Loss are calculated algorithmically.
5. Iridium AI War Room
The finalists land on the Dashboard ready for human-in-the-loop validation. In the War Room, the user inspects the interactive chart (Lightweight Charts), verifies the KPIs, and the AI Coach assists with setup assessment and position sizing using the strict rule: 1% Account Risk = Stop Loss Distance × Number of Shares.
2Partially Implemented
Features that are active in a simplified form and will be tightened in production.
We tighten the existing SEPA filter: the current price must be within 15% of the 52-week high (close ≥ 52w_high × 0.85). We don't buy bounces from lows—we buy leaders. Computational cost: zero.
We don't ingest news in the backend. When the user opens the War Room for a finalist stock, the AI Coach (LLM) has a web search tool to find the latest Earnings Call or Press Release and will tell the user: "The catalyst driving this VCP is the approval of the new patent."
3To Implement (Future Premium)
Features requiring significant data infrastructure, ideal for a future Premium tier.
Requires ingestion and parsing of complex SEC 13F filings to track institutional buying/selling trends. This step will be unlocked with the paid FMP Premium API subscription; with the current free API key, this section is not available. Ideal to place behind the paywall of a future Premium plan.
Candle-by-candle geometric analysis is extremely heavy. Our current Statistical Contraction engine (CV₁ > CV₂ > CV₃) already captures 85% of the mathematical effectiveness of a VCP. The last-mile 15% will be implemented later with advanced candlestick pattern recognition. More than sufficient for the MVP.