A.4.6 Advanced Extensions

Basic Prerequisites: 03 · Screening Metrics and 05 · Failure Modes.

---

1. Intuition & Practical Objective

This page is the launchpad from single screens into quantamental systems - where the discretionary criteria of the Graham school become mechanical, backtestable factors. Three ideas:

1. The Piotroski F-score (2000) turns “is this cheap stock actually good?” into nine binary signals summed to a 0–9 score. Applied within high-book-to-market firms, it separates the winners from the value traps - the single most-cited bridge from statement analysis to a mechanical screen, and the direct antidote to the trap of 05.
2. Combining value with quality is where the edge lives. Value alone is cheap-and-maybe-dying; quality alone is expensive. The intersection - cheap and high-quality - is the quantamental core, formalized in Greenblatt’s “magic formula” (rank on return-on-capital + earnings yield) and Gray & Carlisle’s Quantitative Value.
3. The factor-model seat. Fama–French’s five-factor model (2015) added profitability (RMW) and investment (CMA) to market, size, and value - the official recognition that accounting fundamentals (gross profitability, asset growth) are priced factors. Novy-Marx’s gross profitability and the q-factor model’s ROE complete the bridge from screens to asset pricing.

The through-line: a fundamental screen is a factor portfolio in disguise. Once you standardise its signals and test it cross-sectionally and out-of-sample, it becomes a factor - and it must survive the same scrutiny as any other.

---

2. Mathematical Ground Truth & Derivations

The Piotroski F-score (2000). Nine binary signals, summed:

$$\text{F}=\sum _{i=1}^9{s}_i\in \{0,\dots ,9\},\begin{array}{rl}& \text{Profitability: }[ROA>0],[CFO>0],[\Delta ROA>0],[CFO>ROA]\\ & \text{Leverage/liquidity/funding: }[\Delta LTD<0],[\Delta CR>0],[\text{no equity issue}]\\ & \text{Operating efficiency: }[\Delta GM>0],[\Delta ATO>0]\end{array}$$

where $ROA=NI/\text{avg TA}$, $CFO$ operating cash flow / avg TA, $CR=CA/CL$, $GM=(S-\text{COGS})/S$, $ATO=S/\text{avg TA}$. Piotroski’s result: within the top book-to-market tercile, high-F (8–9) firms substantially outperform low-F (0–1) firms - quality conditions the value signal.

Standardising signals for combination. Metrics on different scales are combined via z-scores. For a characteristic $x$ with cross-sectional mean $\mu$ and standard deviation $\sigma$:

$$z(x)=\frac{x-\mu }{\sigma },\text{composite}=w_{v}z(\text{value})+w_{q}z(\text{quality}).$$

The “magic formula” is the equal-weight special case with value $=E/P+B/P$ and quality $=ROIC$.

The five-factor model (Fama–French 2015). The expected return decomposes as

$$E[R_i]-R_f={\beta }_i^{\text{MKT}}{\lambda }_{\text{MKT}}+{\beta }_i^{\text{SMB}}{\lambda }_{\text{SMB}}+{\beta }_i^{\text{HML}}{\lambda }_{\text{HML}}+{\beta }_i^{\text{RMW}}{\lambda }_{\text{RMW}}+{\beta }_i^{\text{CMA}}{\lambda }_{\text{CMA}},$$

where RMW (robust-minus-weak profitability) and CMA (conservative-minus-aggressive investment) are built directly from accounting fundamentals - operating profitability and asset growth. This is where screens become factors.

---

3. Computational Implementation - the quantamental composite screen

Runs on the standard library only. It builds a value+quality composite from z-scores (the F-score as the quality anchor), ranks a universe, and applies a falling-knife momentum veto - the mechanical synthesis of the whole folder.

Read the two layers: the composite ranks Zeta Steel #2 purely because of its enormous value score (E/P 0.150, B/P 1.50) - the same deep-value name that 05 flagged as a trap. The momentum veto then removes it, leaving Sigma Tools and Alpha Machine - names that are both cheap and high-quality and not in free-fall. That is the whole folder in one script: rank on fundamentals, then let a risk overlay veto the falling knives. (In production you would also control industry, size, and standardise point-in-time - the factor-model discipline above.)

---

4. Failure Modes & First-Principles Breakdowns

1. The composite can be dominated by one characteristic. Zeta’s value score swamped its negative quality; without the veto the screen would have bought the trap. Guard: winsorize each z-score and require both value and quality buckets to be positive, not just the sum.
2. F-score is a distressed-value tool, not a universal quality score. It is designed for high book-to-market firms; applied to healthy growth firms the signals are ambiguous (any improvement scores, even from a terrible base). Condition it on value (Piotroski’s own design) as done here.
3. Factor crowding and decay. Once a screen is published and widely run (the magic formula, the F-score), its edge can compress. Guard: test out-of-sample across decades, expect lower live returns, and diversify across several economically-motivated signals.
4. Backtest illusions. Overfitting, survivorship, look-ahead, and transaction costs turn a beautiful backtest into a losing strategy. Guard: few parameters, point-in-time data, realistic turnover/costs, and replication on independent samples (Backtesting & Validation).
5. The factor-model mismatch. A raw screen’s return is often a repackaging of known factors (value, size, profitability, investment). Guard: regress the screen on the five factors - if the alpha vanishes, you have re-discovered a factor, not a new edge (Fama–French 2015; Green–Hand–Zhang 2017).

---

5. References

• Piotroski, Joseph D.: “Value Investing: The Use of Historical Financial Statement Information to Separate Winners from Losers” (JAR, 2000)
• Greenblatt, Joel: The Little Book That Beats the Market (Wiley, 2005; Revisited 2010)
• Gray, Wesley R. & Carlisle, Tobias E.: Quantitative Value (Wiley, 2013)
• Fama, Eugene & French, Kenneth: “A Five-Factor Asset Pricing Model” (JFE, 2015)
• Novy-Marx, Robert: “The Other Side of Value: The Gross Profitability Premium” (JFE, 2013)
• Hou, Xue & Zhang: “Digesting Anomalies: An Investment Approach” (RFS, 2015)
• Green, Hand & Zhang: “The Characteristics That Provide Independent Information About Average U.S. Monthly Stock Returns” (RFS, 2017)
• O’Shaughnessy, James P.: What Works on Wall Street (McGraw-Hill, 4th ed.)

---

6. Connected Graph Bridges

• Back: 05 · Failure Modes · 03 · Screening Metrics · Index Hub
• Ratio ancestry: Core Financial Ratios - Advanced Extensions (Altman Z, the F-score defined there)
• Systematic layer: Quantitative Fundamental Investing · Fundamental Multi-Factor Models · Backtesting & Validation
• Defense & data: Accounting Quality & Red Flags · Data Sources & Corporate Data
• Sibling: Equity Valuation - DCF, Comps & Value Logic