Full transparency on who publishes Top Micro Caps, how the model was built, and what our track record actually means.
Top Micro Caps is a free daily quantitative research publication by Phoenix Capital. It publishes the top 5 US micro cap stock picks every trading day before 08:00 ET, generated by the Phoenix screening model — a Mahalanobis distance algorithm trained against 500+ historical micro cap winners.
The model has produced a 68% win rate on closed positions with an average closed return of +34.2% (30-day trailing as of March 2026). Take-profit at +60%, stop-loss at -20%, maximum 5-day holding window. All picks, outcomes, and methodology are fully public.
Top Micro Caps is published by Phoenix Capital, a quantitative trading research operation. The Phoenix screening model was designed and built by a finance professional with over 10 years of institutional investment research experience, spanning investment banking, private equity, and buy-side equity research in emerging markets.
Prior roles include Head of Sector Research and Managing Director at established investment research firms, with coverage of Georgian and CIS capital markets, real estate, and technology sectors. The quantitative work behind Phoenix draws directly from institutional-grade financial modeling methodology applied to the US micro cap market.
The decision to publish the model's daily output publicly was deliberate: transparency about methodology, performance, and limitations is the only responsible way to operate a stock research publication. Every pick, outcome, and model parameter is disclosed.
The site focuses exclusively on US-listed micro cap stocks — companies with share prices typically between $0.10 and $5.00 and share floats under 25 million. This universe includes stocks listed on Nasdaq, NYSE, and OTC markets. These are also commonly called penny stocks, low float stocks, or nano caps.
The daily picks target momentum and catalyst setups — stocks exhibiting abnormal volume relative to their float, strong recent price momentum, and statistical similarity to historical breakout patterns. Three setup types are categorized: Breakout (price clearing technical resistance on volume), Catalyst (news-driven event moves), and Reversal (oversold recovery patterns).
This is explicitly speculative, short-duration trading research — not long-term investment analysis. The model's holding window is a maximum of 5 trading days. Micro cap stocks are among the most volatile securities in the US market. We cover them because they offer the largest short-term return potential in the small market segment, when approached with strict risk management.
The model was built in Python using end-of-day OHLCV data from Polygon.io. The benchmark cohort was constructed by identifying ~500 historical US micro cap stocks (2020–2025) that produced minimum 60% gains within a 5-trading-day window from a qualifying entry point. These stocks were filtered to remove obvious outliers and data errors.
Five features were engineered and validated: 5-day momentum, volume surge ratio (vs. 20-day average), float turnover, RSI-14, and intraday range ratio. All features are winsorized at the 1st/99th percentile and log-transformed before scoring. The covariance matrix of the benchmark distribution is estimated using Ledoit-Wolf shrinkage to reduce small-sample estimation error.
Scoring uses the Mahalanobis distance formula: d = sqrt((x-μ)ᵀ Σ⁻¹ (x-μ)) where x is the candidate's feature vector, μ is the benchmark mean, and Σ⁻¹ is the inverse of the Ledoit-Wolf covariance matrix. Candidates with the lowest distance are ranked highest. The final 0–10 score is 10 × (1 – percentile rank of raw distance).
Performance is measured from the published entry price (last closing price before publication) to the first of three exit events: (1) take-profit trigger — daily high reaches +60% above entry; (2) stop-loss trigger — daily low reaches -20% below entry; or (3) weekly expiry — position closed at Friday's closing price if neither trigger is hit within 5 trading days.
Win rate is calculated as the percentage of closed positions that hit the take-profit (+60%) level or closed above entry before expiry. Average return is the arithmetic mean of all closed position returns. All metrics are backward-looking and based on model rules — actual trading returns would differ based on execution, slippage, and position sizing.
The full trade log is published publicly at topmicrocaps.com/track-record.html. No cherry-picking, no retroactive exclusions. All model picks since October 2025 are included, wins and losses.