The quantitative intelligence behind every stock page, valuation model, and investment thesis on CirclFi.
CirclFi Deep Alpha is not a person — it is an AI-powered quantitative valuation engine that serves as the analytical backbone of the entire CirclFi platform. Every stock page, blog post, investment thesis, and valuation estimate published on CirclFi is generated by this engine.
Each trading day, Deep Alpha runs 13 independent valuation models across 5,892+ US-listed equities, processing hundreds of thousands of data points from regulatory filings, macroeconomic indicators, and global news sentiment. The result is a comprehensive, multi-dimensional view of every stock's intrinsic value — far beyond what any single model or analyst opinion can provide.
The engine combines classical finance theory (discounted cash flow, earnings power) with modern quantitative techniques (Bayesian inference, Markov regime-switching, neural network topologies) to produce valuations that capture both the fundamental reality and the probabilistic uncertainty of each investment.
Unlike traditional equity research — where a single analyst's conviction drives a price target — Deep Alpha's multi-model consensus approach eliminates individual bias and provides investors with a statistically robust range of fair-value estimates.
Deep Alpha's analytical pipeline draws from three primary institutional-grade data sources, each chosen for reliability, coverage, and regulatory backing:
The foundation of every valuation. We process 10-K (annual), 10-Q (quarterly), and 8-K (current event) filings directly from the SEC's EDGAR database. Each company is parsed for 700+ XBRL-tagged financial data points — revenue, operating income, free cash flow, debt structure, share counts, segment data, and more. This ensures our models work with audited, regulatory-grade financial data, not estimated or scraped values.
The Federal Reserve Economic Data (FRED) platform provides the macroeconomic context our models require: risk-free rates (Treasury yields), VIX volatility index, GDP growth, yield curve spreads, credit spreads (BAA-AAA), inflation expectations, and more. These inputs calibrate discount rates, risk premiums, and regime-switching parameters across all 13 models.
The Global Database of Events, Language, and Tone (GDELT) feeds our sentiment-driven models with real-time global news tone, volume, and event data. This allows models like Sentiment SOTP to incorporate market narrative and news momentum into valuation — capturing the qualitative dimension that pure financial models miss. All data flows through a fully automated pipeline with no manual intervention.
Every stock on CirclFi is evaluated through 13 distinct methodologies. Each model captures a different facet of value, and together they form a multi-model consensus that is more robust than any single approach.
Discounted cash flow with Bayesian posterior updating. Incorporates prior beliefs about growth rates and margins, updated with each new filing to produce probability-weighted fair values with confidence intervals.
Bruce Greenwald's EPV framework: values the company based on its sustainable current earnings power, stripped of growth assumptions. Ideal for identifying quality businesses trading below their no-growth value.
Measures the spread between Economic Return on Invested Capital and WACC. Companies with positive EROIC spreads create shareholder value; negative spreads destroy it. Based on McKinsey's economic profit framework.
Three-scenario (bull, base, bear) probabilistic valuation. Each scenario models different revenue trajectories, margin profiles, and exit multiples, weighted by macro-calibrated probabilities.
Dividend discount model enhanced with Markov regime-switching. Recognizes that companies transition between growth, stable, and distress states — each with distinct dividend policies and discount rates.
Machine Learning Residual Income Valuation. Uses gradient-boosted trees to predict residual income trajectories, capturing non-linear relationships between financial variables that traditional RIV models miss.
Net Asset Value with dynamic adjustments for intangible assets, off-balance-sheet items, and asset quality deterioration. Particularly valuable for financials, REITs, and asset-heavy industrials.
Probability-Weighted Expected Return Method. Models equity as a set of option-like payoffs across multiple scenarios, particularly effective for companies with binary outcomes or significant embedded optionality.
Cross-sectional peer comparison adjusted for macroeconomic regime. Instead of static P/E comparisons, this model adjusts multiples for the current rate environment, credit cycle, and sector momentum.
Sum-of-the-Parts valuation augmented with GDELT news sentiment. Breaks conglomerates into constituent businesses and applies sentiment-adjusted multiples to each segment.
Confidence-Uncertainty-Calibrated Ensemble. Meta-model that dynamically weights the other 12 models based on their historical accuracy for similar company profiles, producing a calibrated consensus valuation.
Feedforward Topological Neural Network that maps high-dimensional financial features to valuation outputs. Trained on historical valuation accuracy data with regularization to prevent overfitting.
Regime-Conditional Monte Carlo with Hidden Markov DCF. Combines Monte Carlo simulation with hidden Markov models to generate regime-aware probability distributions of fair value.
For detailed mathematical frameworks and assumptions behind each model, see our Methodology page.
Deep Alpha maintains rigorous quality standards to ensure every valuation published on CirclFi meets institutional-grade reliability:
All 13 models are recalculated daily after market close. This ensures valuations reflect the latest market prices, newly filed SEC documents, and current macroeconomic conditions. Stale data is the enemy of accurate valuation — Deep Alpha eliminates it.
Every stock receives a proprietary QOC score from 0 to 10, computed from 32 fundamental signals spanning profitability (gross margin, operating margin, net margin, ROE, ROA, ROIC), growth (revenue growth consistency, earnings momentum), balance sheet health (debt-to-equity, interest coverage, current ratio), cash flow quality (FCF yield, operating cash flow margin, capex efficiency), and capital allocation (dividend sustainability, buyback effectiveness). The QOC score distills a company's fundamental quality into a single, actionable metric.
Our proprietary Value Trap detection algorithm identifies stocks that appear statistically cheap but carry hidden fundamental risk. The algorithm cross-references apparent undervaluation (models showing significant upside) against deteriorating financial metrics — declining margins, rising leverage, negative free cash flow trends, and weakening competitive position. Stocks flagged as potential value traps receive a warning indicator, protecting investors from the most common pitfall in value investing.
All content on CirclFi is algorithmically generated. Every stock page, investment thesis, blog post, and valuation estimate is produced by Deep Alpha's quantitative models processing SEC filings, macroeconomic data, and market sentiment. There is no human editorial bias in any content published on the platform.
This algorithmic approach ensures consistency, objectivity, and scalability. The same rigorous analytical framework applied to Apple is applied to a $50 million micro-cap — no favoritism, no selective coverage, no conflicts of interest.
Our methodology is fully transparent. Every model's mathematical framework, data inputs, assumptions, and known limitations are documented on our Methodology page. We show confidence scores for each valuation so investors can assess reliability themselves.
What we don't do: We do not accept payment for coverage, we do not issue buy/sell recommendations, and we do not provide personalized financial advice. CirclFi is a data platform, not an advisory service.
CirclFi provides quantitative valuation data for educational and informational purposes only. Nothing on this website constitutes financial advice, a recommendation to buy or sell any security, or an offer of any financial product. All investing involves risk, including the possible loss of principal.
The valuations displayed are generated by automated quantitative models and may contain errors. Past performance is not indicative of future results. Users should conduct their own due diligence and consult a qualified financial advisor before making investment decisions.
CirclFi Deep Alpha is a quantitative equity valuation engine that runs 13 independent valuation models on 5,892+ US-listed stocks daily. It processes SEC EDGAR filings, FRED macroeconomic data, and GDELT news sentiment to produce fair-value estimates, quality scores, and value trap detection.
CirclFi uses 13 independent valuation models spanning intrinsic (Bayesian DCF, EPV, EROIC Spread), scenario-based (First Chicago, PWERM), regime-switching (Markov DDM, RCMH-DCF), machine learning (ML-RIV, FTNN Topology), relative (Regime Cross-Sectional), sentiment-driven (Sentiment SOTP), and ensemble methods (CUCE Ensemble, Dynamic NAV).
No. CirclFi Deep Alpha is an AI-powered quantitative engine, not a human analyst. All content is algorithmically generated from SEC filings and quantitative models with no human editorial bias. This ensures consistency, objectivity, and coverage across 5,892+ stocks simultaneously.
CirclFi sources data from SEC EDGAR (10-K, 10-Q, 8-K filings with 700+ XBRL tags per company), FRED (Federal Reserve Economic Data for macroeconomic indicators like risk-free rates, VIX, GDP growth), and GDELT (Global Database of Events, Language, and Tone for market sentiment analysis).
All 13 valuation models are recalculated daily after market close. The automated pipeline processes the latest SEC filings, market data, macroeconomic indicators, and sentiment data before publishing updated analysis for 5,892+ stocks.
The QOC score is a proprietary 0–10 metric that evaluates companies across 32 fundamental signals including profitability margins, revenue growth consistency, balance sheet leverage, free cash flow generation, and capital allocation efficiency. It distills fundamental quality into a single, actionable number.