AI stock predictor for ten horizons
TradeWell is an AI stock predictor that publishes a TW Quant Score, an expected return, and a model-implied price target for every covered name. Each stock gets its own set on ten horizons, from 1 day to 5 years, updated daily across 2,800+ stocks.
TW Quant Score is research, not investment advice. It is not a recommendation to buy or sell any security. Predictions can be wrong. Past model performance does not guarantee future results. TradeWell does not publish fabricated win rates, hit rates, or IC figures as product claims.
What an AI stock predictor is
An AI stock predictor is research software that ranks stocks from model outputs rather than a single analyst label. Searchers looking for the best AI to predict stocks are usually asking which product publishes a score they can inspect, on a horizon they can name. TradeWell does not claim to be the best app, and it does not claim a win rate or that it beats another product. The horizon-by-horizon comparison lives on the best AI stock picking app page.
What this predictor publishes, for each covered stock and each horizon, is three research figures: a TW Quant Score, an expected return, and a model-implied price target. Free visitors see the top 5 names. Rank 6 and below stay locked until a trial.
How the TW Quant Score works
The TW Quant Score is a horizon-specific research score for one stock on one window. Scores are not normalized across horizons, so a high 1-week score does not imply a high 1-year score. The production score combines daily composite signals — price, momentum, volume, fundamentals, sentiment, news, and macro inputs — refreshed on current market data. The full signal list and the research-versus-live split are on the methodology page and the TW Quant Score explainer.
The score is not a buy or sell rating. A higher rank means the model placed that name above others for that horizon. It is not a promise that the price target will print.
Ten horizons, from 1 day to 5 years
TradeWell does not publish one blended AI score. Each covered stock receives a separate forecast on ten windows. Screen those windows in the stock screener by time horizon. The window-by-window walkthrough is ten prediction horizons.
| Horizon | Window | Separate output |
|---|---|---|
| 1D | 1 day | Score, expected return, price target |
| 3D | 3 days | Score, expected return, price target |
| 1W | 1 week | Score, expected return, price target |
| 2W | 2 weeks | Score, expected return, price target |
| 1M | 1 month | Score, expected return, price target |
| 3M | 3 months | Score, expected return, price target |
| 6M | 6 months | Score, expected return, price target |
| 1Y | 1 year | Score, expected return, price target |
| 3Y | 3 years | Score, expected return, price target |
| 5Y | 5 years | Score, expected return, price target |
How accuracy is measured
NDCG@30 is the primary live ranking metric. It scores how well the top thirty names align with realized outcomes for that horizon, rewarding better names placed higher. When the evaluated list is shorter than 30 names, the floor is NDCG@20. NDCG@10 can appear as a secondary field. It is not the headline metric.
Research models must pass walk-forward out-of-sample testing before any promotion to the public board. Promotion is operator-only. TradeWell does not publish fabricated win rates. Read the definitions, not a slogan:
- Stock ranking accuracy — what NDCG@30 and the NDCG@20 floor mean
- How TradeWell accuracy is measured — live rankers versus research tests
- Walk-forward stock model validation — the research gate before a model can serve
Quant performance verification
Quant performance verification starts with the metric, then the page that shows a date. Use the three explainers above for the definitions. Open the live rankings in the web app or the iOS app and pick one horizon. Then open a per-ticker prediction page and read the as-of date on the baked score. Those pages do not replace the live board.
Examples, not a performance ranking: Apple stock prediction, NVIDIA stock prediction, Microsoft stock prediction, and the prediction page directory. More writeups sit on the research hub. TradeWell does not publish a win-rate percentage to verify, because it does not invent one.
What you get for each stock
- 10 TW Quant Scores (1D, 3D, 1W, 2W, 1M, 3M, 6M, 1Y, 3Y, 5Y)
- 10 expected returns and 10 model-implied price targets
- Filters by sector, market cap, and index (S&P 500, Nasdaq-100, Russell 2000, Dow)
Open the rankings
The current board is in the product. This page explains the predictor. It does not reprint live scores.
AI stock predictor FAQ
- What is an AI stock predictor?
- An AI stock predictor is research software that ranks stocks from model outputs rather than a single analyst label. TradeWell publishes a TW Quant Score, an expected return, and a model-implied price target for each covered stock on each of ten horizons, refreshed daily across 2,800+ stocks.
- How does the TW Quant Score work?
- The TW Quant Score is a horizon-specific research score for one stock on one window. Each window has its own score, expected return, and model-implied price target. Scores are not normalized across horizons. Daily composite signals include price, momentum, volume, fundamentals, sentiment, news, and macro inputs. The score is not a buy or sell rating.
- How do the ten horizons work?
- The ten horizons are 1D, 3D, 1W, 2W, 1M, 3M, 6M, 1Y, 3Y, and 5Y, covering 1 day through 5 years. Each horizon is a separate forecast. A high 1-week score does not imply a high 1-year score.
- How is AI stock predictor accuracy measured?
- NDCG@30 is the primary live ranking metric. It scores how well the top thirty names align with realized outcomes for that horizon, rewarding better names placed higher. When the evaluated list is shorter than 30 names, the floor is NDCG@20. Research models must pass walk-forward out-of-sample testing before promotion. TradeWell does not publish fabricated win rates.
- How do I verify quant performance?
- Read the stock ranking accuracy, accuracy measurement, and walk-forward validation pages for the metric definitions, then open the live rankings and the per-ticker prediction pages, which show an as-of date. TradeWell does not publish a win-rate percentage to verify, because it does not invent one.
- Is an AI stock predictor investment advice?
- No. TradeWell publishes research, not investment advice. Nothing on this page or in the app is a recommendation to buy or sell any security. Predictions can be wrong. Past model performance does not guarantee future results.
Related reading: methodology · TW Quant Score · ten prediction horizons · stock screener by time horizon · stock ranking accuracy · how accuracy is measured · walk-forward validation · research · best AI stock picking app · prediction pages.