Factor Lab Daily Brief 2026-10-02

Data Cutoff

Price cache as of 2026-10-01 (Wednesday), run at Beijing Time 2026-10-02 06:30. Fundamental data from daily TTM earnings panel (combined_panel.pkl), update frequency depends on earnings release schedule.

Sample & Evaluation Methodology

• Universe: S&P 500 constituents, ~502 valid stocks (474 for ROE)
• Factor definitions: EP (TTM net income / market cap), BP (book value / market cap), FCF Yield (TTM FCF / market cap), ROE (TTM net income / shareholders’ equity), MOM (Price[t-21]/Price[t-273]−1, ~10-month price momentum), VOL (60-day annualized volatility), SIZE (log market cap)
• Evaluation window: 60 cross-sectional observations (fundamental factors) or 39 (price factors); MOM/VOL/SIZE use 21-day forward returns — the most recent 21 trading days cannot yet be fully evaluated
• IC: Spearman rank correlation, cross-period mean; NOT daily IC, NOT future return, NOT downside probability
• ICIR: IC mean / IC std, NOT strategy Sharpe
• p-value: Ordinary t-test, not corrected for overlapping-sample autocorrelation
• Long-short return: Q1 − Q5 (low factor group minus high factor group), gross (not net of costs)

Key Changes This Period

Only one new trading day (10-01) added since last run (10-01). Fundamental factor EP/BP/FCF Yield/ROE IC values are completely unchanged (identical across 5 consecutive runs since 09-26). MOM/VOL/SIZE adjusted slightly due to price update but remain within the range established over the past two weeks.

Complete 7-Factor Table

⭐ = p < 0.05

Cross-Section Comparison: Last 5 Distinct Run Dates

Observation: Four fundamental factors (EP/BP/FCF/ROE) have been identical across 5 consecutive runs since 09-26, indicating the TTM financial panel in the cache did not update after that date. MOM narrowed from −0.076 to −0.019, still negative but with reduced magnitude. VOL stable around +0.117. SIZE gradually moved from negative toward zero but remains statistically insignificant.

Sector Momentum Table (MOM by Sector)

⭐ = p < 0.05

Anomaly Alert Interpretation

System-flagged IC sign “flip” (🔴) items this period:

Key note: ic_history.json contains the May–Jun 2026 period when momentum was broadly positive. Momentum regime flipped in late August; anomaly detection will keep flagging “positive-to-negative” — this is a continuation marker of the existing regime, not a daily突变. Must cross-reference with ic_history to confirm continuation vs new breakout.

Sector Momentum Chart

Caption: Run 2026-10-02, prices as of 2026-10-01. IT sector momentum significantly negative (IC=−0.177, p=0.019), Health Care significantly positive (IC=+0.108, p=0.030). Utilities significantly negative with ICIR=−1.0.

Factor IC Time Series Chart

Caption: Run 2026-10-02. VOL persistently elevated (+0.117); EP/FCF/ROE significantly negative; MOM negative but narrowing.

Strategic Implications

1. Volatility (VOL) is the strongest and most robust signal: IC=+0.117, ICIR=+1.05, p<0.0001. Low-vol组合显著跑赢高-vol组合 (Q1−Q5 = −5.07%), consistent with the low-vol anomaly.

2. Momentum regime shifted negative but reversal not confirmed: Overall MOM IC=−0.019 (insignificant), but IT sector at −0.177 (significant). The “strong gets stronger” dynamic within IT has broken down — low-momentum IT stocks relative outperformed high-momentum ones. Health Care is the only sector with significantly positive momentum.

3. Fundamental factors broadly negative: EP (−0.048), FCF Yield (−0.035), ROE (−0.025) all significantly negative. High-EP / high-FCF / high-ROE groups underperformed low groups over the past 21 days. This does not mean value/quality factors are失效 — the cross-sectional 21-day window captures relative rank changes, not absolute return direction.

4. SIZE mildly positive: Moved from −0.022 toward +0.012, direction shifted from large-cap bias to small-cap preference, but p=0.595 is completely insignificant — insufficient to support any conclusion.

Conclusions NOT Supported

• ❌ MOM negative ≠ market must decline, nor does it mean laggards are worth buying. It only means the “winner portfolio” underperformed the “loser portfolio” within the sample.
• ❌ SIZE positive ≠ small-cap rally; p=0.595.
• ❌ EP/FCF/ROE negative ≠ value rotation confirmed. These factors measure cross-sectional rank correlation with 21-day returns; a negative sign only means high-factor groups underperformed low-factor groups.
• ❌ Sector alert red count ≠ independent risk event. Multiple sectors’ momentum flips are different facets of the same regime change.
• ❌ Gross long-short returns (uncost-adjusted) ≠ realizable strategy returns.

Statistical & Data Limitations

• Price cutoff: 2026-10-01 (latest available at Beijing Time 10-02 06:30 run). Fundamental factor ICs unchanged since 09-26 — the TTM financial panel cache had no new earnings-driven updates after that date.
• p-values: Ordinary t-test, not corrected for overlapping-sample autocorrelation. Overlapping returns understate standard errors; p-values may be optimistically biased.
• MOM holding period: 21-day forward returns means the most recent 21 trading days (09-08 to 10-01) cannot yet form a complete evaluation. MOM valid observations (39) fewer than fundamental factors (60).
• ICIR is IC mean / IC std, NOT strategy Sharpe ratio or annualized return.
• Quantile direction: Long-short return is Q1 (low factor group) − Q5 (high factor group). Positive value means low-factor group outperformed; does NOT mean high-factor group made money.
• Sector coverage: Sector decomposition covers 7 GICS一级 sectors (Consumer Discretionary/Staples, Financials, Health Care, Industrials, Information Technology, Utilities). Energy, Materials, Real Estate, Communication Services are NOT covered. Current count of negative sectors does not equal number of new reversals this period.