Value Edge is a 0 to 100 score that measures whether a stock trades below what its fundamentals justify. Four components are averaged equally: Valuation asks if the stock is cheap against peers, Consistency whether the numbers can be trusted, Quality whether the business is good, and Momentum which direction it is heading. Scores of 65 and above rate Undervalued.
Every input is ranked against industry peers, never against absolute thresholds. Everything comes from reported financial data, with no analyst estimates anywhere in the system. A score of 40 to 64 rates Fair Value, and below 40 rates Overvalued.
This post explains what each component measures and why we built it this way. We publish the reasoning because self-directed investors managing $100K+ portfolios should never trust a black-box score, including ours. What follows is the full design logic, with the supporting research cited where each choice is made and listed in full at the end. The exact weightings and data mechanics inside each component stay proprietary.
What question does Value Edge actually answer?
Whether the market is paying attention. Cheap only means something relative to a peer group, so every ratio in the system is ranked within the company's industry. A P/E of 25 tells you nothing in isolation. In an industry where the median multiple sits near 50, it means the market is offering a 50% discount, and the interesting question becomes whether that discount is a mistake or a warning.
The four components exist to separate those two cases. One asks if the discount exists. The other three ask whether it is deserved.
Ranking against peers is not just a preference. It is what the valuation literature says multiples are for. The classic study on comparable-firm selection showed that a multiple only carries information relative to genuinely similar businesses, and that peers chosen for economic similarity predict a company's future multiples far better than coarse industry groupings do (Bhojraj and Lee, "Who Is My Peer?," Journal of Accounting Research, 2002). That finding shaped how we define a peer group. Our coverage universes are built from the value chain itself, meaning every company that designs, manufactures, equips, or supplies the industry in question, rather than from a statistical classification code. A lithography supplier gets compared against the companies actually competing for the same capital spending, not against whatever else a database filed under the same four-digit industry code.
The four components at a glance
| Component | Question it answers | What it looks at |
|---|---|---|
| Valuation | Cheap versus peers? | A blend of price ratios ranked against the same industry |
| Consistency | Can you trust the numbers? | Revenue steadiness and gross margin stability over multiple years |
| Quality | Good business, or cheap for a reason? | Profitability, capital discipline, balance sheet safety |
| Momentum | Getting better or worse? | Growth in revenue, earnings, and cash flow, plus whether growth is accelerating |
Each contributes equally to the composite. A stock has to be good on several dimensions to rank highly, and one strong reading cannot hide three weak ones.
How does the Valuation component work?
It blends several price ratios rather than leaning on any single one, because every ratio breaks somewhere. P/E fails when earnings go negative, and enterprise-value measures, which price the whole business including its debt, stay comparable across different debt loads where equity-only multiples do not. The blend is chosen so that at least one measure stays valid whenever another breaks, which keeps a company scoreable even when parts of its income statement are ugly. The exact set and their weights are part of the proprietary mechanics.
The blend is weighted toward cash-based measures by design. Accounting earnings can be dressed up. Cash either arrived or it did not. The empirical record backs the choice: cash-based operating profitability predicts returns better than earnings measures that mix in accruals, the non-cash estimates that accounting layers on top of actual cash movements, and the long-documented return anomaly built on accruals disappears entirely once profitability is measured on a cash basis (Ball, Gerakos, Linnainmaa, and Nikolaev, Journal of Financial Economics, 2016). The same authors' companion study approaches it from the measurement side: profit taken at the operating level, where this year's expenses are matched against this year's revenues, tracks future returns better than either bottom-line earnings or gross profit alone (Ball, Gerakos, Linnainmaa, and Nikolaev, "Deflating Profitability," Journal of Financial Economics, 2015). Two papers, one lesson: the closer a number sits to actual cash and actual operations, the more it tells you about future returns. That lesson set the weighting inside the blend.
We deliberately left out price-to-book. Our coverage compares companies across a full value chain, from IP designers to capital-heavy manufacturers, and P/B systematically punishes the asset-light end of that chain. A metric with a built-in bias against half the universe is worse than no metric.
Consistency is the trust filter
A cheap stock with strong margins sounds perfect until you notice the margins swing 30 points between quarters. Consistency measures whether fundamentals are predictable, built on multi-year trailing windows of reported data. Revenue steadiness captures demand. A monopoly whose customers cannot switch should show it in the revenue line, and wild swings suggest the position is weaker than the market share number implies. Gross margin stability captures pricing power. A bottleneck controller that holds its margin through different market conditions is demonstrating the structural advantage in real time.
This component exists to qualify the other three. When Valuation says cheap and Quality says good, Consistency tells you whether those readings are a durable pattern or a lucky snapshot of a volatile business.
The academic quality research draws the same distinction. The canonical quality definition includes safety alongside profitability and growth, and safety in that framework is measured partly in the fundamentals themselves, through the stability of what the business earns rather than only the volatility of its stock (Asness, Frazzini, and Pedersen, 2019). Consistency is our version of that safety dimension, rebuilt for a supply chain context where the question that matters is whether a structural position keeps showing up in the reported numbers year after year.
We already replaced one component. Here is why.
The original second component was History, which compared a stock to its own five-year valuation range. When we reviewed the first full quarters of output, two problems surfaced. History and Valuation agreed with each other far too often, so the system was effectively counting one signal twice. Worse, History punished exactly the wrong companies, in both directions.
Take a business whose niche product just became critical to a major supply chain. The market re-rates it from 25x to 60x because the business genuinely changed, yet against its own five-year range that new multiple reads as wildly expensive, so History scored the improvement as a fault. The old range priced a business that no longer exists. Now take a cyclical at the top of its cycle. Peak earnings make the P/E look low, often below the company's own average, so History called it cheap at exactly the moment earnings were about to roll over. Punishing the transformation while rewarding the value trap is the opposite of the job.
Consistency replaced it because business predictability is genuinely independent information. We publish this openly because a methodology that has never admitted a design mistake is a methodology nobody has stress-tested.
Quality catches stocks that are cheap for a reason
The most common failure in value investing is buying deterioration at a discount. The Quality component screens for it across profitability, capital discipline, and balance sheet safety, using checks like gross margin, return on equity, interest coverage, and share dilution. Gross profitability anchors the set for a reason: it predicts returns about as powerfully as the classic value measures while carrying information they do not (Novy-Marx, "The Other Side of Value," Journal of Financial Economics, 2013). High margins with efficient capital use and a shrinking share count scores near the top. Thin margins with heavy debt service and constant share issuance scores near the bottom.
The evidence base behind this component is among the deepest in modern finance, and its most useful finding is not the one people usually quote. The Quality Minus Junk study defines quality as the characteristics investors should pay a premium for, sorts them into profitability, growth, and safety, and shows that a portfolio that buys high-quality stocks and bets against junk earns significant risk-adjusted returns in the United States and across 24 countries (Asness, Frazzini, and Pedersen, "Quality Minus Junk," Review of Accounting Studies, 2019). The detail that matters for a valuation methodology is why those returns exist. The market does pay more for quality, but by what the authors call a puzzlingly modest margin, which means quality is persistently underpriced. A component that scores quality, sitting next to a component that scores cheapness, is built to catch exactly that gap: the good business the market is only half-paying for.
The phrase worth paying special attention to is risk-adjusted returns. A risk-adjusted return measures what you earned relative to how much risk you took to earn it. Two portfolios can both return 8% a year, but if one swings twice as hard and falls twice as deep along the way, it delivered a much worse deal for the same headline number. Quality stocks earn their premium on exactly this measure. The claim is not that they beat the market every year. It is that the returns they do earn come with less risk attached, and the QMJ evidence is unusually direct about the downside: quality-minus-junk returns are high during market downturns, and instead of carrying crash risk the factor benefits from the flight to quality when crises hit. In plain terms, owning quality has historically meant falling less when everything falls.
Falling less matters more than it sounds, because losses and recoveries are not symmetric. A position that drops 50% needs a 100% gain just to get back to even, so the less often a portfolio visits deep losses, the less heroic its recoveries need to be. Long-term compounding is won there, not in the occasional spectacular year. Screening hard for quality is how this methodology leans against ever needing the heroic recovery.
Here too the exclusions were deliberate. Debt-to-equity overlaps with interest coverage, and coverage is the more actionable of the two because it measures the ability to service debt rather than the mere presence of it. Every metric in the system had to carry information the others did not.
Momentum measures direction, not price
A company can score 85 on Quality while revenue quietly shrinks. By the time that shows up in the headline numbers, the stock has usually moved. Momentum tracks whether revenue, earnings, and free cash flow are growing on a rolling basis, and it also tracks acceleration. A company whose growth went from 10% to 30% and one that faded from 50% to 30% print the same growth rate today while heading in opposite directions. The distinction matters more than the level.
One exclusion defines this component. Stock price momentum, a staple of quantitative strategies since Jegadeesh and Titman put the anomaly on the academic map ("Returns to Buying Winners and Selling Losers," Journal of Finance, 1993), is not an input. The substitution has empirical support: earnings momentum largely explains what price momentum captures, so measuring the fundamentals directly keeps the signal and drops the market echo (Novy-Marx, "Fundamentally, Momentum Is Fundamental Momentum," NBER, 2015). Price momentum is a market signal, and Value Edge deliberately separates what the business is doing from what the market thinks. The entire point of the score is to find the places where those two disagree.
Dropping the price version also drops its worst habit. Price momentum earns its returns with a catch: the strategy suffers rare but severe crashes, concentrated in the sharp rebounds that follow market panics, when a portfolio built from past winners is positioned almost perfectly against the recovery (Daniel and Moskowitz, "Momentum Crashes," Journal of Financial Economics, 2016). Momentum measured in revenue, earnings, and free cash flow has no such reversal built into it. A company that kept growing through a panic simply kept growing, and the score keeps saying so.
How do the components interact?
| Scenario | Reading |
|---|---|
| High Valuation, high Consistency | Cheap and predictable. The strongest setup in the system. |
| High Valuation, low Consistency | Cheap but erratic. The discount may be deserved. |
| High Quality, low Momentum | Good business, deteriorating. Watch, do not buy yet. |
| Low Quality, high Momentum | Improving from a weak base. Recovery or value trap. |
Equal weighting is what makes the table honest. A stock scoring 95 on Valuation with 25, 30, and 20 on the other three lands at 42.5, which is Fair Value, not a bargain. We could have weighted the components by conviction, and the research would even offer cover: both quality and momentum have historically earned persistent extra returns (Asness, Frazzini, and Pedersen, "Quality Minus Junk," Review of Accounting Studies, 2019; Jegadeesh and Titman, 1993). We chose not to, and the portfolio literature is on our side: across fourteen optimized weighting models and seven datasets, none consistently beat the naive equal-weight rule out of sample, meaning on data the models had not been tuned on (DeMiguel, Garlappi, and Uppal, Review of Financial Studies, 2009).
The scale of that result deserves numbers. Calibrated to US equity data, the authors estimate that a mean-variance model would need roughly 3,000 months of history, which is 250 years, before its estimated weights reliably beat the naive rule on a 25-asset portfolio, and about 6,000 months for 50 assets. Estimation error eats the theoretical gains, and it keeps eating them for centuries. An entire follow-up literature exists precisely because the naive benchmark is so hard to beat, adding mathematical constraints to the models to keep their estimates in check (DeMiguel, Garlappi, Nogales, and Uppal, Management Science, 2009). Our components are score dimensions rather than portfolio assets, but the statistical lesson transfers directly: a weighting scheme is itself an estimate, and estimates are where error enters a system. Equal weighting states plainly that we do not claim to know which dimension matters most, and it will only change if live results earn the change.
What does Value Edge deliberately not do?
It makes no predictions. The score describes what current reported data says about price versus fundamentals, not where the price goes next. It uses no analyst estimates, no price targets, and no consensus figures. It contains no technical analysis. And it excludes banks, insurers, and REITs entirely, because financial companies run on different metrics and forcing them into this framework would produce confident-looking nonsense.
The refusal to predict is not modesty for its own sake. The twelve-month price target is the closest thing equity research has to a crystal ball, and its track record is measurable: across roughly a decade of published analyst targets, fewer than four in ten stocks were trading at or above the target when the twelve months ran out, and the average target implied a return about 15 percentage points higher than what the stock actually delivered (Bradshaw, Brown, and Huang, Review of Accounting Studies, 2013). That is not an indictment of any individual analyst. It is what happens when a methodology promises a number the future has not agreed to yet.
Nor is the optimism a local habit that better rules could fix. A cross-country study of target prices, which starts from the well-documented fact that targets are persistently optimistic, finds that analysts in countries with stronger institutions issue significantly less inflated targets. But less inflated is not uninflated: the bias shrinks, it does not disappear (Bradshaw, Huang, and Tan, Journal of Accounting Research, 2019). Even the quality literature makes the same point in passing, finding that analysts' price targets and earnings forecasts imply systematic quality-related errors in expectations (Asness, Frazzini, and Pedersen, 2019). Forecast optimism is a structural feature of professional equity research, not a defect any single firm can hire its way out of. The clean design response is to ingest no forecasts at all. We would rather describe the present accurately than predict the future confidently, which is why every claim the score makes is checkable today.
A scoring system is defined as much by what it refuses to ingest as by what it measures. Every excluded input above was excluded on purpose, and the reasoning is in this post.
The evidence behind the design
Every structural choice in Value Edge maps to a finding in the published asset-pricing literature. The compressed version:
| Design choice | What the research shows | Source |
|---|---|---|
| Peer-relative ranking | Multiples predict best against economically similar peers, not broad industry codes | Bhojraj and Lee (2002) |
| Cash-weighted valuation blend | Cash-based profitability predicts returns better than accounting-earnings measures | Ball et al. (2015, 2016) |
| Quality anchored on gross profitability | Predicts returns about as powerfully as classic value measures, and quality stays persistently underpriced | Novy-Marx (2013); Asness et al. (2019) |
| Fundamental momentum, not price momentum | Earnings momentum explains what price momentum captures, without the crash-prone tail | Novy-Marx (2015); Daniel and Moskowitz (2016) |
| Equal component weights | Across 14 weighting models and 7 datasets, none consistently beat the naive equal-weight rule out of sample | DeMiguel et al. (2009) |
| No analyst forecasts anywhere | Only 38% of 12-month targets met at the horizon, and the optimism persists across countries | Bradshaw et al. (2013, 2019) |
None of these papers hands you a scoring system. They establish which raw materials carry signal and which carry noise, and the craft is in the assembly. But a methodology whose every major joint rests on replicated, peer-reviewed findings is a different object from a proprietary formula that asks to be taken on faith.
How does the score fit into a report?
Value Edge is the third step of five in how we build industry coverage. We map the value chain first, then identify the chokepoints where substitution is hardest, then score every company in the coverage, then rank them, and only then build picks on the highest-conviction names, each with its thesis, key risks, and the trigger that would change our verdict. The score flags where to look. The report explains what we found there.
When we scored 89 equipment and materials companies for the July 2026 issue, the composite surfaced high-conviction picks we would not have reached through a standard value screen, including monopolies buried inside diversified parents that single-ratio screens systematically misread. That gap between what screens see and what the supply chain position implies is the product.
The same system now scores every industry we cover, from semiconductor equipment to life science tools, without changing a single rule. A methodology that needs re-tuning for each new industry is a methodology fitted to its first one. This one was built peer-relative from the start precisely so it would travel.
Common questions
Is a Value Edge score of 65+ a buy signal?
No. It flags a stock as statistically cheap relative to peers on multiple dimensions, which makes it worth investigating. Our picks add the supply chain thesis, the key risks, and the conditions that would change our verdict on top of the score. The score narrows the field. It does not make the decision.
How often do scores update?
Each industry refreshes in full every quarter, from a fresh data snapshot dated on the report itself. Score updates within days of each earnings release are on the platform roadmap. A methodology tied to reported financials moves when the financials move.
Why does Value Edge exclude financial companies?
Banks, insurers, and REITs run on metrics like net interest margin and book value that have no clean equivalent in an industrial framework. Scoring them with gross margins and free cash flow ratios would produce numbers that look precise and mean nothing. Exclusion is more honest than false coverage.
How is this different from a single fair-value estimate?
A single-analyst fair value rests on one model and its assumptions, usually a discounted cash flow. Value Edge is peer-relative and multi-dimensional, so a stock only ranks as Undervalued when several independent measurements agree. Neither approach predicts prices. The difference is that ours is built to be auditable: every input comes from reported financial data, so any number behind a score traces back to a filing you can open and a calculation you can repeat. A fair-value estimate moves when its author changes an assumption. A Value Edge score moves only when the company reports new numbers.
References
Every empirical claim in this post traces to peer-reviewed research. The sources, in alphabetical order:
- Asness, C. S., Frazzini, A., and Pedersen, L. H. (2019). "Quality Minus Junk." Review of Accounting Studies, 24(1), 34-112.
- Ball, R., Gerakos, J., Linnainmaa, J. T., and Nikolaev, V. (2015). "Deflating Profitability." Journal of Financial Economics, 117(2), 225-248.
- Ball, R., Gerakos, J., Linnainmaa, J. T., and Nikolaev, V. (2016). "Accruals, Cash Flows, and Operating Profitability in the Cross Section of Stock Returns." Journal of Financial Economics, 121(1), 28-45.
- Bhojraj, S., and Lee, C. M. C. (2002). "Who Is My Peer? A Valuation-Based Approach to the Selection of Comparable Firms." Journal of Accounting Research, 40(2), 407-439.
- Bradshaw, M. T., Brown, L. D., and Huang, K. (2013). "Do Sell-Side Analysts Exhibit Differential Target Price Forecasting Ability?" Review of Accounting Studies, 18(4), 930-955.
- Bradshaw, M. T., Huang, A. G., and Tan, H. (2019). "The Effects of Analyst-Country Institutions on Biased Research: Evidence from Target Prices." Journal of Accounting Research, 57(1), 85-120.
- Daniel, K., and Moskowitz, T. J. (2016). "Momentum Crashes." Journal of Financial Economics, 122(2), 221-247.
- DeMiguel, V., Garlappi, L., Nogales, F. J., and Uppal, R. (2009). "A Generalized Approach to Portfolio Optimization: Improving Performance by Constraining Portfolio Norms." Management Science, 55(5), 798-812.
- DeMiguel, V., Garlappi, L., and Uppal, R. (2009). "Optimal Versus Naive Diversification: How Inefficient Is the 1/N Portfolio Strategy?" Review of Financial Studies, 22(5), 1915-1953.
- Jegadeesh, N., and Titman, S. (1993). "Returns to Buying Winners and Selling Losers: Implications for Stock Market Efficiency." Journal of Finance, 48(1), 65-91.
- Novy-Marx, R. (2013). "The Other Side of Value: The Gross Profitability Premium." Journal of Financial Economics, 108(1), 1-28.
- Novy-Marx, R. (2015). "Fundamentally, Momentum Is Fundamental Momentum." NBER Working Paper No. 20984.
Written by the Stocks & Signals research team: 20+ years of experience across equity markets, algorithmic trading, and supply chain analysis.
Want to see real score cards on real companies? Get the free preview issue, with sample fully-scored ranking entries ungated.