Methodology

1. Graham Number

√(22.5 × EPS × Book Value Per Share)

A conservative ceiling on what a defensive investor should pay, from the last full year's earnings and the latest book value. The constant 22.5 caps Graham's preferred maximums of P/E ≤ 15 and P/B ≤ 1.5 multiplied together.

2. Graham 1962 Formula

EPS × (8.5 + 2g)

Values a stock from current earnings plus an estimate of future growth (g, as a whole number percentage). 8.5 is Graham's assumed fair P/E for a no-growth company.

4. Earnings Power Value (EPV)

Two versions, both shown and labeled:

• Version A (EPS ÷ WACC): labeled “non-Graham” in the app — Graham's own methods were pure arithmetic with no discounting; a required-return input (WACC) is a modern addition he never used.
• Version B (average 7-year EPS ÷ AAA bond yield): Bruce Greenwald's textbook formulation, closer to Graham's original “earning power” concept.

5. Net-Net / NCAV

(Current Assets − Total Liabilities) ÷ Shares Outstanding

What's left for shareholders if the company liquidated today — Graham's deepest bargain test. Flagged as a classic buy signal when price is below 67% of NCAV per share. Genuine net-nets are rare outside of small caps and market bottoms.

PEG Ratio (Lynch)

P/E ÷ Annual EPS Growth Rate (%)

Peter Lynch's growth-adjusted valuation check from One Up on Wall Street. Below 1.0 reads as potentially undervalued relative to its own growth; above 2.0 as expensive. A ratio, not a dollar value.

Owner Earnings price ceiling (Buffett)

Net Income + D&A − CapEx − Change in Working Capital

From Warren Buffett's 1986 Berkshire Hathaway shareholder letter, shown as a rough ~10× price ceiling. Total capital expenditure is used as a stand-in for “maintenance” capex (the two aren't separately reported), which tends to understate owner earnings for capital-light, fast-growing companies. A single year's working-capital swing can also move this figure more than the underlying business really changed.

DCF intrinsic value (Damodaran, simplified)

FCFF × (1 + g) / (WACC − g)

A single-stage Gordon Growth discounted cash flow, simplified from Aswath Damodaran's actual multi-stage methodology (high growth fading to a stable terminal rate). Numerically unstable when WACC is close to the growth rate, so the app won't compute it when the spread is under one percentage point. Free cash flow is also unreliable for banks and other financial institutions, whose deposit- and loan-driven cash flow doesn't map onto this formula. The value is shown with a range: the same formula with WACC and the growth rate each one point lower and higher, and only its low end when a more hopeful combination brings growth within a point of WACC.

Dividend discount model (Gordon growth and two-stage)

Dividend × (1 + g) / (required return − g)

The oldest discounted-cash-flow model: a share is worth the dividends it will pay.

• Dividend: the latest regular payment times payments a year, or the last two payments added together for a company that pays twice a year. A recent raise or cut counts; a special dividend doesn't.
• Growth: the dividend's own growth over the last five full calendar years, with each year counted at its usual number of payments, so a skipped payment doesn't read as growth. For the part that lasts forever, it is held between 0% and 4%.
• Two-stage value: faster growth runs for five years first, no faster than the app's earnings growth estimate, never above 15% a year, and not once the dividend takes 90% of earnings.
• Range: like the DCF's, with the required return and growth each a point lower and higher.
• Shown for companies that have paid a dividend in each of the last six full calendar years, earn more than they pay out and yield under 10%. Not shown for asset managers, funds, split-share corporations or mortgage REITs.

Justified price-to-book (financials)

Book value per share × average ROE / required return

For banks and insurers, whose value rests on what they earn on their book: the average return on equity over up to seven years, divided by the return an owner requires, is the price-to-book multiple that book has earned. No growth is assumed in the value; the range beside it uses (ROE − g) / (r − g) with the required return a point lower and higher and growth from −1% to +1%. Not shown for asset managers, funds, exchanges or insurance brokers, or where the book value or the yearly return on equity is too erratic to trust.

Price to NAV and FFO (Canadian REITs)

Unit price / NAV per unit

For the Canadian REITs whose quarterly results the app keeps. A REIT that carries its properties at fair value has something close to a net asset value on its balance sheet.

• NAV: the unit price is set against the REIT's own NAV per unit, with its label, date and page, converted where the REIT reports in US dollars.
• No NAV published: IFRS equity per unit stands in, but only for a REIT that carries its properties at fair value and has no more than a sliver of units exchangeable into trust units. At June 30, 2026, equity per unit came within 3% of the REIT's own NAV for six of the 18 REITs that publish one and have equity per unit in the data feed, and was as much as 61% away where exchangeable units sit in liabilities or minority interests.
• FFO and AFFO: each REIT's own per-unit figures. Most REITs follow REALPAC's guidance, some with exceptions they list, so the figures don't always compare like for like. The multiples use four times the latest quarter.
• Cap rate: the REIT's own. Where the REIT's table shows what a quarter-point move in that same cap rate does to its property values, the card gives the move in dollars a unit.
• Takeovers: where a REIT has agreed to be taken over, the card says so, with the terms, since its units then trade on the offer.

Market-implied growth rate (reverse DCF)

The same Gordon Growth formula above, solved for the growth rate the current price already implies, shown next to the app's own growth estimate for comparison. An implausibly high implied growth rate is itself the signal — “the market is pricing in unsustainable growth” — so it's not suppressed the way the forward DCF is.

Acquirer's Multiple (Carlisle)

Enterprise Value ÷ Operating Income

Tobias Carlisle's deep-value screen from The Acquirer's Multiple. Uses operating income as reported rather than an “adjusted EBITDA” figure, on Carlisle's own stated reasoning that companies use adjustments to flatter results. Also ranked across the full Screener universe (lower is cheaper).

ROIC / ROCE / CROIC

Three capital-efficiency ratios (return on invested capital, return on capital employed, and cash return on invested capital). ROIC here uses net income rather than after-tax operating income (NOPAT) — the simpler, more common practitioner version.

Warning Signs panel

An automated red-flag checklist, separate from the Defensive Investor screen's pass/fail bar. Eight checks: negative free cash flow, negative net income, EPS decline over 3 years, revenue decline over 3 years, book value per share decline over 3 years, any dividend cut in the recent window, negative Buffett owner earnings, and total debt exceeding book equity. This is deliberately not exhaustive — share dilution, auditor changes, and insider selling are real red flags this panel does not currently check.

Piotroski F-Score

A 9-point quality checklist (Joseph Piotroski, 2000) comparing this year against last on profitability, leverage/liquidity, and operating efficiency. Shown on the Screener.

Altman Z-Score

Edward Altman's 1968 bankruptcy-risk model, combining five balance-sheet and income-statement ratios into one number: above 2.99 reads as safe, 1.81–2.99 a gray zone, below 1.81 distress risk. Reads artificially low for banks and insurers, whose deposits and policy liabilities don't fit the model's assumptions — Altman himself published separate variants for non-manufacturers for this reason.

Beneish M-Score

Messod Beneish's earnings-manipulation probability model (1999), combining eight year-over-year financial ratios into one score. Above −1.78 (Beneish's own threshold) reads as an elevated red flag. Excluded for Financial Services tickers, whose balance-sheet structure produces false positives in several of the underlying ratios.

Accruals ratio

(Net income − Operating cash flow) / Average total assets

The part of a year's profit that didn't arrive as cash, sized against the business, from the same two annual statements as the Beneish model. Richard Sloan (1996) found that the accrual part of earnings persists less than the cash part. Below zero is the norm, since depreciation lowers profit without using cash: up to 5% reads low, 5–10% some, above 10% high. Excluded for Financial Services tickers, and not shown when accruals would exceed total assets, which points to a data problem.

Greenblatt's Magic Formula

Joel Greenblatt's The Little Book That Beats the Market ranking: every eligible ticker is ranked by earnings yield and separately by return on capital, the two ranks are summed, and the universe is sorted ascending — #1 is the best combination of cheap and good. Excludes Financial Services and Utilities sectors, whose balance-sheet structure distorts the underlying ratios.

Quality snapshot

A summary that scores five categories from 0 to 5, using only checks already on the research page. A category without enough data isn't scored, and Overall is the average of the categories that are. Not shown for preferred shares.

• Value: the margin of safety against the average intrinsic value. A price above that value scores 0; a margin under 15% scores 1, under 25% scores 2, under 33% scores 3, under 50% scores 4, and 50% or more scores 5.
• Financial health: the share of four checks that pass, among those with data, scaled to 5 and rounded: Graham's current-ratio and debt tests, and the Warning Signs panel's free-cash-flow and debt-versus-equity checks.
• Past performance: the same for six Warning Signs checks: earnings per share, revenue and book value not declining, no dividend cut, and positive net income and owner earnings.
• Growth: the app's historical earnings growth estimate, a look-back rather than a forecast. Below 0% scores 0, under 5% scores 1, under 10% scores 2, under 15% scores 3, under 25% scores 4, and 25% or more scores 5.
• Dividend: no dividend scores 0. A payer starts at 2, adds 1 for a trailing yield of 2% or more (2 for 4% or more), and adds 1 more when the dividend-cut check is clear, up to 5.