What is the TW Quant Score?
The TW Quant Score is TradeWell’s proprietary research rating built to answer a fundamental investing reality: no single stock score makes sense across every investment timeframe. TradeWell evaluates 2,800+ US equities across ten discrete time horizons, providing ten distinct scores, expected returns, and model-implied price targets refreshed daily.
1. Ten Horizons, Not One Blended Number
Traditional equity screeners typically force a single composite score onto a company, flattening short-term momentum and five-year compounding into an arbitrary average. TradeWell scores every covered stock independently across 10 discrete forecast windows:
- Tactical short-term: 1 Day (1D), 3 Days (3D), 1 Week (1W), 2 Weeks (2W)
- Intermediate swing: 1 Month (1M), 3 Months (3M), 6 Months (6M)
- Strategic long-term: 1 Year (1Y), 3 Years (3Y), 5 Years (5Y)
Crucially, scores are not normalized or averaged across horizons. A company displaying exceptional short-term volume breakout characteristics might rank at the top of the 1W board, while carrying a moderate 3Y score due to valuation multiples. Each horizon answers an isolated question for that specific timeframe. Explore our stock screener by time horizon to see how investors filter by horizon.
2. 100+ Daily Composite Signals
Every TW Quant Score synthesizes over 100 proprietary composite features. Rather than relying on isolated technical indicators, our production scoring pipelines process:
- Price & Momentum: Multi-timeframe trend alignment, moving average ribbons, RSI dynamics, and relative strength versus sector benchmarks.
- Volume & Money Flow: Accumulation/distribution patterns, institutional volume pressure, and liquidity metrics.
- Fundamentals & Quality: Earnings revisions, return on capital, margin trends, debt coverage, and balance sheet durability.
- Insider & Institutional Activity: Form 4 insider transactions, cluster purchases, and institutional holdings shifts.
- Market Sentiment & News: Scored financial news flow and natural language catalyst detection.
- Macro & Seasonality: Historical calendar tendencies, rate regime sensitivity, and sector rotation tailwinds.
Signals and scores refresh every trading day before market open on fresh market data and daily bar histories.
3. Production Architecture & Accuracy: NDCG@30
How do we know if rankings are working? In systematic quantitative finance, evaluating a ranked list requires ranking metrics — not noisy binary win/loss labels.
Measured by NDCG@30
Live production rankers (gradient-boosted trees, linear baselines, and Elite deep neural rankers) are evaluated using NDCG@30 (Normalized Discounted Cumulative Gain at rank 30). NDCG@30 measures whether the top 30 ranked stocks on each horizon generate higher realized returns than lower-ranked names, heavily weighting top-tier positioning. When the evaluated list is shorter than 30 names, the product reports NDCG@20. NDCG@10 remains a secondary diagnostic and is not the product headline.
Research architectures and candidate challenger models undergo rigorous walk-forward out-of-sample (OOS) testing across rolling market cycles before they can ever be considered for promotion to live production boards. All model promotions are strictly operator-gated — we never auto-promote research models to production. Read our detailed breakdown on how TradeWell accuracy is measured.
4. What We Refuse to Invent (No Fake Metrics)
Unlike marketing sites that boast simulated "94% win rates" or fabricated Sharpe ratios, TradeWell enforces strict honesty standards:
- We never publish fabricated win rates or hit rates as marketing claims. Financial markets are probabilistic, and ranking quality is measured through rigorous ranking metrics.
- We never publish fake zeros for missing data. If an indicator or metric lacks coverage, it is honestly labeled as missing or unavailable, never filled with a deceptive zero.
- We do not claim research architectures are live until they actively serve the production board. Learn more on our methodology page.
5. See Predictions in Action
Curious how the TW Quant Score looks for real companies? Browse our programmatic coverage of Nasdaq-100 companies, complete with as-of scoring dates and signal breakdowns:
- Apple Inc. (AAPL) stock prediction
- NVIDIA Corporation (NVDA) stock prediction
- Tesla, Inc. (TSLA) stock prediction
- Full Nasdaq-100 stock prediction directory
Frequently Asked Questions
- What is the TW Quant Score?
- The TW Quant Score is TradeWell's proprietary research rating for an individual stock on a specific investment horizon. Each covered stock receives 10 distinct scores across 10 time horizons (1D to 5Y), paired with expected returns and model-implied price targets.
- How does the TW Quant Score differ from traditional analyst ratings?
- Traditional analyst ratings (Buy/Hold/Sell) are subjective, infrequently updated, and usually lack a defined mathematical time horizon. The TW Quant Score is systematic, refreshed daily from 100+ composite signals, and calculated independently for ten specific timeframes.
- Are TW Quant Scores normalized across different horizons?
- No. Scores are not normalized across windows. A high 1-week tactical score does not imply a high 1-year score. Each horizon evaluates distinct market dynamics.
- How is accuracy validated?
- Production rankers are evaluated on NDCG@30 (NDCG@20 when fewer than 30 names were evaluated), measuring how well top-ranked stocks outperform realized market benchmarks. New models must pass walk-forward out-of-sample backtests.
- Is the TW Quant Score investment advice?
- No. TradeWell is a quantitative research platform, not an investment advisor or broker-dealer. Nothing on TradeWell constitutes a recommendation to buy, sell, or hold any security.
Explore related research: Methodology · Ten prediction horizons · Accuracy Measurement · Horizon Screener · Research Library · Stock Predictions