The Zacks Rank has a 40-year brand, but its methodology is proprietary, its inputs are one-dimensional (earnings-estimate revisions), and the useful tools cost ~$249/year. StoQuant publishes every dimension of the Q-Score, all 149 ML features, and the walk-forward validation.
The Zacks Rank is built on one insight from the 1980s: earnings-estimate revisions predict short-term returns. That edge is real — and it is exactly why analyst revisions are one of the nine dimensions in StoQuant's Q-Score. But a rank driven mostly by revisions is one-dimensional, updated in coarse 1-5 buckets, and its exact formula is proprietary. StoQuant blends revisions with momentum, sentiment, insider activity, technical structure, sector relative strength, macro regime, and 141+ more features in a 149-feature ML ensemble with Venn-ABERS calibration. The full methodology is published at /learn/q-score-methodology, and the out-of-sample results are on /proof.
| Feature | StoQuant | Zacks |
|---|---|---|
| Price | Free core tier; Power plan for full API + MCP access | Free; Premium plan for full access |
| Free tier limits | Full Q-Score, hidden gems, and public research pages | Zacks Rank #1-5 |
| Stocks covered | 3,500+ US stocks scored weekly | US stocks |
| Core scoring / key differentiators | 149-feature ML ensemble with Venn-ABERS calibration | Zacks Rank (earnings estimate revisions) |
| ML / walk-forward proof | Published daily on /proof | No |
| AI-agent / MCP support | stoquant-mcp for Claude / AI agents | No |
| Portfolio optimization | Black-Litterman optimizer | No |
| Data sources | 14 alt-data feeds (insider, social, macro, SEC, app ranks) | Earnings estimates, surprise data |
Choose Zacks if you specifically want to trade earnings-estimate revision momentum with a four-decade brand behind it, or if you value their equity research reports and earnings surprise data. Choose StoQuant if you want a transparent, multi-dimensional, calibrated score you can audit line by line — with alternative-data confirmation, value screening, and a published out-of-sample track record — without paying $249/year to unlock the tools.
Read the full Q-Score methodology at /learn/q-score-methodology, verify the walk-forward results on /proof, and try the free hidden-gems screener at /gems.
Yes. The Q-Score leaderboard, hidden-gem screener, and insider tracking are free. Zacks shows the current Rank list free but locks its screeners, research, and focus lists behind Premium at roughly $249/year.
The Zacks Rank is a proprietary 1-5 bucket driven mainly by earnings-estimate revisions. Q-Score is a 0-100 calibrated probability from a 149-feature ML ensemble whose full 9-dimension methodology is published and validated out-of-sample.
Yes — revisions are one of the nine Q-Score dimensions, because the underlying effect is real. The ML ensemble combines them with 140+ other features and weights each input by its measured out-of-sample predictive power.
Yes. /learn/q-score-methodology documents all nine dimensions and the feature set, /methodology covers the pipeline, and /proof publishes walk-forward validation with hit rates, quintile spreads, and out-of-sample AUC.
3,500+ US stocks, each rescored within a week. We deliberately exclude micro-caps below $250M to maintain data quality and liquidity.
Weekly Q-Score signals and insider buying activity on our highest-conviction stock picks. Email only, no account needed. See a sample digest.