Qademic is deterministic: the same inputs always produce the same scores, and the construction of every signal is published here. We did not invent these factors — each one stands on decades of peer-reviewed evidence. AI never scores, sizes, or predicts; it only reads documents. This page describes what is live today, not what is planned.
BUSINESS = mean of the sector-neutral percentile ranks of Q3 and Q6. “Versus comparable companies, is this a good business at a fair price?” Moves quarterly.
TIMING = mean of the sector-neutral percentile ranks of Q4, Q5 and Q7. “Versus comparable companies, is now a good entry?” Moves daily.
REGIME = market-wide gate. Scales exposure and gates new entries. It is identical for every stock, so it can never move a stock's rank — blending it in would shift the whole universe together and pollute stock-vs-stock comparison.
Each pillar is ranked sector-neutral (percentile vs comparable companies; thin sectors fall back to a whole-universe rank), then the ranks combine with equal weights (1/N). Ranks, not absolute levels — and weights are never tuned to backtests, because naive equal-weight combination is brutally hard to beat out-of-sample (DeMiguel, Garlappi & Uppal 2009). Grades describe entry quality (Timing, with a Business floor) — never buy/sell instructions.
Five components, fixed conventional thresholds: SPY vs 200dma (trend), % of 11 sector ETFs above own 200dma (breadth), VIX level (volatility), high-yield credit spread level + 3-month change (credit), 10y−2y Treasury spread (curve). Sum ∈ [−5,+5] → risk-on / neutral / risk-off; state flips need 2 consecutive closes. Regime gates exposure — it never adds points to any stock.
EVIDENCE · Faber (2007) — trend-following reduces drawdowns; credit spreads as risk gauge
Relative strength = sector ETF return minus SPY return over 1m / 3m / 6m, computed from our own stored daily prices. Sectors ranked by 3m RS; four-phase read (leading / improving / weakening / lagging) from RS level + 1m pace. Context for theme selection — not a stock-score input.
EVIDENCE · Moskowitz & Grinblatt (1999) — industry momentum persists for months
Profitability (gross profitability, FCF margin), growth (revenue YoY), returns (ROE/ROIC), balance sheet (D/E), valuation context — scored from cited fundamentals, then converted to a sector-neutral percentile rank across the universe before it enters the Business score (ranks, not absolute levels).
EVIDENCE · Novy-Marx (2013) gross profitability; Piotroski (2000) F-score
Price vs 200/50/20dma, momentum over 1–12 month windows, RSI position, volume confirmation — computed nightly from daily OHLCV. No intraday data, by design.
EVIDENCE · Jegadeesh & Titman (1993) — 12-1 momentum, the most replicated anomaly in finance
Insider transactions (SEC Form 4 via aggregator), analyst recommendation trends. Migrating to opportunistic-vs-routine insider classification and rating changes (changes predict; levels do not).
EVIDENCE · Cohen, Malloy & Pomorski (2012) — decoding inside information
ROIC level and trend, debt trajectory, capital allocation quality from reported statements.
EVIDENCE · Sloan (1996) — accruals quality predicts returns
Days to next earnings, beat streak, post-earnings reaction context.
EVIDENCE · Ball & Brown (1968) — post-earnings-announcement drift, replicated for 50+ years
Price-based signals are backtested — prices are honest point-in-time data. Components with no decile spread get removed; weights are never tuned.
Fundamental signals are NOT backtested by us. Vendor fundamentals are restated over time; backtesting them looks rigorous and is quietly a lie. The academic literature above is the backtest — done on clean data, by people better at it than us.
Everything is tracked forward instead. Scores are snapshotted daily to an append-only history. As the record matures, this page will show — live and unfakeable — how stocks we ranked highly actually performed versus stocks we ranked poorly, and versus simply holding SPY.
Qademic is an educational research framework providing general information only. It does not consider your objectives, financial situation or needs, and is not financial advice. Nothing here is a recommendation to buy or sell any security.