CreativeQuant Market uses CQ Alpha, a multi-factor model that blends price action, fundamentals, risk, institutional activity, and other market signals into a confidence-aware research score.
CQ Alpha ranks research opportunities. The score is not a prediction or recommendation. Confidence measures data support. Missing inputs are neutralized, not guessed.
A proprietary quantitative research tool that ranks a defined universe of stocks by a model-generated score. It is built to make a disciplined process legible: every score carries its factor categories, data freshness, and confidence.
It is not personalized financial advice, not a recommendation to trade any security, and not a claim of future returns. The model can be wrong, incomplete, or outdated. It is one structured input for your own research — nothing more.
CQ Alpha is a proprietary multi-factor research model. It combines momentum, quality, valuation, growth, risk, institutional activity, sentiment and news, and technical structure. Each factor is normalized across the universe and combined into a single confidence-aware score used to rank the universe. The exact weighting and normalization system are proprietary.
High-level view. The weighting scheme is part of CreativeQuant's research IP.
The public dashboard ranks a focused core universe — a deliberately small set of liquid large-caps. Separately, the model runs a wider expanded research universe of roughly 75 stocks used only for forward-testing and methodology research. The expanded research runs do not change the public dashboard rankings. Keeping them separate lets us study the model at scale without quietly altering what you see live.
A small, liquid large-cap set. This is what the live engine ranks and displays.
A wider set used to forward-test and study the model. It does not change the public rankings.
These nine factors are the complete input set. Some inputs may neutralize when reliable provider data is unavailable; confidence reflects whether each signal was available and fresh.
Blends 1-month, 3-month, and 6-month price returns into one trend reading.
Needs roughly six months of price history; newly listed names read thin.
Composite of profit margins, return on equity, and debt-to-equity.
Missing provider ratios fall back to a neutral reading.
Composite of valuation multiples relative to peers. Cheaper multiples read higher.
Missing valuation ratios fall back to a neutral reading.
Composite of revenue growth and earnings-per-share growth.
Missing revenue or EPS growth inputs fall back to neutral.
The direction analysts are revising their earnings estimates.
Some tickers may neutralize when reliable estimates or rating revision data is unavailable.
Combines historical volatility, drawdown depth, and beta deviation. Steadier reads higher.
Needs sufficient price history and a beta input.
Institutional ownership derived from SEC EDGAR 13F filings, or a bounded proxy.
13F filings are quarterly and lag the market.
Aggregated direction of news flow for a company.
Some tickers may neutralize when reliable provider-backed news sentiment is unavailable.
Reads price relative to its moving averages and their structure.
Needs enough price history to compute the moving averages.
Not every factor counts equally. CQ Alpha applies a fixed weighting and normalization scheme tuned so steadier, higher-conviction signals carry more influence. Those exact weights are part of CreativeQuant's research IP and aren't published — but the full set of factor families is shown here, with nothing hidden about what goes in.
Price history comes from Tiingo, with FMP and Alpha Vantage as fallbacks. Company fundamentals come from FMP. Institutional ownership is derived from SEC EDGAR 13F filings. Every factor on a score carries its own source label and an “as of” timestamp, and the dashboard shows freshness state — fresh, cached, or degraded — so you always know how current the inputs are. We do not scrape websites and we do not store third-party commentary or proprietary estimates.
Every factor reports a confidence level from 0 to 100%. When inputs are partial, confidence drops. When provider data is unavailable for a ticker, that factor is set to a neutral value with reduced confidence and clearly marked — it never guesses a number to fill the gap. Low overall confidence is shown on the score itself, so a thinly-supported ranking can't masquerade as a strong one.
Full data coverage
Partial coverage — flagged low
No feed — neutral, reduced confidence
Each research run is stored as a point-in-time snapshot — the scores and factor breakdowns exactly as they were on that date. Over time we measure what happened after each snapshot across multiple horizons. This forward-test record is still accumulating. It is not yet a performance claim, and we won't present it as one until there's a meaningful, honestly-measured sample. Until then, snapshots are research records, not proof.
CQ Alpha v1 is the current version. The near-term focus is connecting additional data feeds, adding a benchmark for fair comparison, and building out the forward-test record. We'll publish changes here. We are deliberately not changing the scoring model until forward-testing justifies it.
Every ranked row carries the same three reads. They describe the research context, not what action to take.
Where the model ranks this name in the universe — a research rank, not a recommendation.
How much usable, fresh data supported the score. Low confidence is flagged, never hidden.
Which providers fed the score and how current they are — fresh, cached, or degraded.
Open the live engine to view CQ Alpha scores, factor breakdowns, confidence, and freshness on the current universe.
CreativeQuant Market provides educational market research and model-generated watchlists. It is not personalized financial advice, investment advice, or a recommendation to buy or sell any security. Markets involve risk, including loss of principal. Model outputs can be wrong, incomplete, or outdated. Always do your own research.