How we calculate

The sector composite from 0 to 100 is a weighted average of category scores. The weights are Quality 25% · Valuation 25% · Growth 20% · Outlook 15% · Health 15%. If a company has no outlook score, its share is spread across the other categories in proportion to their weights. Missing categories are excluded and the available weights are rescaled. The F-Score is separate and does not enter this average. A score of 80 does not mean the company outranks 80% of peers or has an 80% chance of profit. The Valuation axis uses valuation multiples alone, rather than the overall composite. The reader draws the conclusion.

Read the live AAPL page

• Academic basis

  Piotroski's original research studied firms with high book-to-market ratios, with the strongest benefits concentrated among smaller companies. A high ratio alone does not establish undervaluation. That research does not validate our adapted method for every company or investment horizon.

• Signals from the past

  The sector composite rests on historical filings. The optional outlook layer adds a forward dimension, but none of its signals is a price prediction.

• The financial sector

  The classic valuation multiples and debt metrics work differently for banks, insurers, and brokers. The page flags when a ticker belongs to Financials or Real Estate.

• Survivorship in the sector medians

  Sector medians are computed from the current S&P 500 slice. Delisted or acquired companies do not re-enter the sample, a known limit on backward-looking comparisons.

• One ticker is not statistics

  The classification timeline shows how the method read a company in a few past quarters. It illustrates the logic; it is not a full validation, because the data is limited and transaction costs are left out.

• Partial data

  Margins, returns, growth and the F-Score of the company and its peers use the latest completed fiscal year (FY). Multiples that compare the price with the accounts use the latest data: P/E and FCF yield on the trailing twelve months (the annual report rolled forward with the quarters since), P/B on the latest balance sheet; EV/EBITDA stays on FY, because we resolve debt only at fiscal year-ends. If SEC's aggregated data does not yet include the latest filed quarterly report, that company's multiples use FY. Valuation uses an explicitly identified raw price and reported instant shares aligned using verified split history. Missing coverage, weighted-average shares, or a split between the share date and its filing leave price-based metrics unavailable. Later buybacks, issuance, ADR-ratio changes and multiple share classes still require separate reconciliation. P/E uses equity market cap divided by trailing-twelve-month net income available to common shareholders. Debt means short- and long-term borrowings, including finance leases when the company reports them together; operating leases are not included. Simplified enterprise value is equity market cap plus debt less cash, plus preferred stock and minority interests at carrying value; if either of those claims is unknown, EV/EBITDA is unavailable. Missing debt, cash, tax inputs or comparable sector data leave the affected metric or comparison unavailable.

• Analyst optimism bias

  When the outlook is on, some sub-signals come from analyst estimates, which the literature puts at roughly 5–15% too optimistic. That is why realized EPS growth from SEC filings serves as a separate cross-check sub-signal.

• Sample for the outlook axis

  Outlook percentile ranks need enough valid peers in the sector. For thin sectors the rank shows '—' even after a full data refresh.

• Classification timeline without the outlook axis

  The quarterly timeline is built on three dimensions: fundamental quality, valuation, and growth. The outlook dimension is absent from the archive: it would need analyst consensus for the past quarter, and we do not store retrospective estimates. That is an honest limit of the archive, not a bug.

• Backtest check

  We test the method with a backtest: does it select stocks with better future returns than the market? As of the last run, no statistically significant edge in selection showed up; in some runs the relationship was even negative. Details and figures are in the section 'What the evidence shows'. So the verdict stays a classification. We will add a claim about predictable returns only once testing shows a durable edge.

P/E

Equity market capitalization divided by net income available to common shareholders over the trailing twelve months (the latest annual report rolled forward with the quarters since); the fiscal-year figure when SEC data lags the latest filing. The company and its peers are valued by the same rules.

FCF

The cash a company keeps after running the business and paying for equipment and other long-term assets.

capex

Money spent to buy or improve long-lived assets such as plants, equipment, or property.

EBITDA

Earnings before interest, taxes, depreciation, and amortization: a rough proxy for operating profit before financing and accounting effects.

EV

Market capitalization plus net debt, preferred stock and minority interests: the total price to buy the whole business, not just its shares.

D/E

Total debt divided by shareholder equity: how much the company borrows relative to the owners' capital.

current ratio

Current assets divided by current liabilities: whether short-term resources cover short-term bills.

ROE

Net income available to common shareholders divided by average common equity over the fiscal year: the profit earned on each dollar of the owners' capital.

ROIC

After-tax operating profit divided by invested capital (equity plus debt less cash) at the fiscal year end: how well the company turns its full funding base into profit.

gross margin

Gross profit as a share of revenue: what is left from each sales dollar after the direct cost of the goods sold.

operating margin

Operating profit as a share of revenue: what is left from each sales dollar after running costs, before interest and tax.

dividend yield

A year of dividends divided by the share price: the annual cash return at today's price.

market cap

The share price times the number of shares outstanding: the market's price tag on all of the equity.

revenue growth YoY

The percentage change in revenue against the same period one year earlier.

net income

What is left of revenue after every cost, including interest and tax: the bottom-line profit.

F-Score

A 0-to-9 score from nine pass/fail checks on profitability, leverage, and efficiency (Joseph Piotroski, 2000): a reading of past financial strength, not a forecast.

Piotroski

The accounting researcher who designed the F-Score (Joseph Piotroski, Stanford, 2000). Kept in the Latin spelling.

sector median

The middle value of a metric across companies in the same GICS sector: the yardstick for cheap or expensive within an industry.

sector composite

A weighted average of the category scores, with the weights set out in the methodology. It is not the company's percentile and not a probability of profit.

value trap

A stock that looks cheap on its multiples but carries weak fundamentals (a low F-Score), where the low price reflects a declining business rather than a bargain.

fundamentals

The core financial facts of a business (profitability, debt, efficiency, cash flow) drawn from its filings, as opposed to its share-price behavior.

This is an automated analysis based on public financial data. It is a general, informational brief and does not account for your investment goals, financial situation, time horizon, or risk tolerance. The analysis is not investment advice or a substitute for a personal consultation. It combines a company's historical filings with analyst-consensus signals, but it does not weigh management quality, competitive advantages, regulatory news, or geopolitical risk. The F-Score reflects past financial behavior, not a forecast of the future; consensus signals carry a documented optimistic bias and are used as a stress test. Any investment decision stays with you, and PlainTicker is not responsible for its outcome. If the amount is material to your finances, consult a licensed financial adviser first. We classify each company against its sector by a fixed rule, sorting it into one of three classes: “Passes the quality and price filter”, “Partly passes”, or “Does not pass the filter”. This is not a price forecast or investment advice. As of the last run, the backtest behind the rule found no statistically significant edge in selection, and in some runs the relationship was negative.

The figures on this page come from public sources: SEC EDGAR filings first, with market prices and analyst estimates from licensed data providers. The write-up and the 'What's next' steps are drafted by an LLM (Claude), working only from the numbers this same page already computed, under rules we set, not the model. Every generated passage runs through an automated check on format and language before it can appear; a passage that fails the check is replaced with a short template line instead. The classification (“Passes the quality and price filter”, “Partly passes”, or “Does not pass the filter”) is the output of a fixed rule, not the model's judgment and not advice. The one-line company description under the company name is also written by an LLM, from the company name and its SIC classification at SEC EDGAR, not from the filings themselves.

• What we tested

  The exact rule behind the verdict, run on the historical S&P 600 membership, 2016 to 2025, every quarter. It used as-filed SEC data, and results are net of costs. The test was pre-registered and run on 3 and 4 October 2026. On average about 15 companies per quarter landed in the “Passes the quality and price filter” class.

• Result: no return edge

  Over the following 12 months those companies landed, on average, at the 49th percentile of their sector, which is the middle. At 95% confidence the edge over the sector median is no more than about +2 pp a year, and most likely there is none.

• The order of the classes did not hold

  We expected “Passes the quality and price filter” to do better than “Partly passes”, and that better than “Does not pass the filter”. The data show no such order.

• Other ideas did not show a reliable edge either

  We also tested post-earnings drift (PEAD), a mix of momentum, gross profitability and value, and the Ohlson distress model. None showed a reliable edge.

• What does predict: a stock's own price volatility

  A stock's own volatility predicts the chance of a fall of 50% or more within a year (out-of-sample AUC about 0.65). That is about risk, not about which stock will rise.

• How to read this

  The verdict is a transparent description of business quality and price relative to the sector. It is not a forecast that the stock will beat the market. For context, S&P SPIVA finds that 85 to 90% of professional active funds fail to beat their index over 10 to 15 years.

• Limits of the test

  Nine years is a short history. Historical data carry survivorship bias. The test used small and mid-cap S&P 600 companies, while the site mostly shows large caps, so the result may not transfer one to one.

• What we do about it

  We keep a live public log of the rule's decisions. From the end of 2026 the rule and the log are frozen for prospective testing on future data, so the rule cannot be tuned to a result already known.

• Why a separate model

  For a bank, deposits and loans are the business itself. The F-Score, debt to EBITDA, free cash flow and the general sector composite do not read a bank fairly, so we do not apply them to banks.

• The comparison group

  13 US banks, identified by CIK in the registry: JPM, BAC, WFC, C, USB, PNC, TFC, MTB, FITB, KEY, RF, HBAN, CFG. A metric is ranked only when at least 8 banks, the target included, have it. The group is recomputed twice a week; if the comparison is missing or older than 16 days, there is no class.

• Metrics

  Profitability: return on common equity (ROE) and the efficiency ratio (lower is better). Risk: net charge-offs to net revenue (lower is better) and tangible common equity to tangible assets (TCE/TA). Growth: net revenue YoY and earnings per share YoY. Price: price to tangible book (P/TBV) and P/E; a lower multiple gives a higher score. For context only, with no effect on the class: net charge-offs to average loans (their level depends on the loan mix: a card book always charges off more, so risk is measured as charge-offs per unit of revenue), ROTCE, the loan-loss allowance to loans, loan and deposit growth.

• How the class is set

  A category score is the mean percentile of its metrics among the banks in the group. "Business" is the mean of profitability, risk and growth with equal weights; "Price" is the price score. "Passes the quality and price filter": profitability of 50 or more, net revenue and earnings per share up on the year, risk of 30 or more and price of 40 or more (30 or more for a position the paper portfolio already holds). "Does not pass the filter": two or three business criteria fail. "Partly passes": everything else. A missing metric means no class, not a fail.

• Consensus signals

  The forward P/E (the analysts' estimate for next year) is ranked only among the banks in the group, and realized growth is taken from this model's growth category (net revenue and earnings per share over the year). If the company is dearer than most of the group and grows more slowly than most (both scores at 40 or below), the "Passes the quality and price filter" class is lowered to "Partly passes". When fewer than 8 other companies in the group have a forward P/E, there is no check: no downgrade and no upgrade. The check can only lower a class, never add one. The forward P/E already carries the analysts' optimism, so it is not part of the price score.

• What the model leaves out

  CET1 (common equity tier 1 capital over risk-weighted assets) is shown for reference only: we take it from each group bank's latest quarterly or annual report under the standardized approach, which all 13 report. The figure is read automatically every week and accepted only when CET1 capital divided by risk-weighted assets gives the printed ratio. Three banks print no capital amount, so they have no automatic value: we show the last checked one until it is more than 200 days old. It does not affect the class or the scores, because more capital lowers both risk and returns. The net interest margin (NIM) cannot be taken from free SEC data in the same way for the whole group, so the model does not use it. The 50, 30 and 40 thresholds are product policy, not a calibration: the model has not been tested on historical returns. We compute the metrics the same way for every bank from 10-K and 10-Q filings, so they can differ slightly from the figures a bank reports itself.

• Investment banks (GS, JEF): descriptive, no class

  Investment bank metrics are shown descriptively: there are too few comparable companies for a class. Fewer than 8 US broker-dealers have a balance sheet like GS and MS, and advisory boutiques with almost no balance sheet have returns that do not compare. So on the GS and JEF pages the return on common equity, pretax margin, net revenue and EPS growth, assets to equity, P/TBV and P/E stand next to the values of the other companies in the group (GS, MS, JEF), without a rank, a median or a class. JEF combines a broker-dealer with merchant banking and has no deposit-taking bank, so even against GS and MS its comparison is only approximate. JEF's Series B preferred shares are counted as common equity, as in the company's own reporting. In the paper portfolio the freeze rule applies to them until the deadline.

• Brokers (SCHW, HOOD): descriptive, no class

  Broker metrics are shown descriptively: there are too few comparable companies for a class. Both profiles, the bank-broker (SCHW) and the retail broker (HOOD), share one group of 7 companies: SCHW, HOOD, IBKR, LPLA, RJF, SF, OPY. That is fewer than the 8 a rank needs, and the brokers that serve clients through advisors (LPLA, RJF, SF, OPY) also earn and spend differently from the online platforms SCHW, HOOD and IBKR. SCHW and HOOD are not yet among the companies analysed on the site; when they are, their return on common equity, revenue and EPS growth, assets to equity, P/B and P/E will be shown descriptively, next to the values of the other companies in the group, without a rank, a median or a class. There is no pretax margin: some companies report revenue net of interest expense and others do not, so it has no common base, while the change in revenue over the year compares each company with itself. Client assets, net new assets, net interest margin and CET1 are not in the standard SEC data, and we do not add them by hand. Much of HOOD's revenue is transaction revenue (options, crypto), so its comparison is only approximate. In the paper portfolio the freeze rule applies to them until the deadline.

• Why a separate model

  An insurer collects premiums before it pays claims, and that money works as part of the business. The F-Score, free cash flow, debt to EBITDA and the general sector composite do not read an insurer fairly, so we do not apply them to insurers.

• The comparison group

  11 US property and casualty insurers, identified by CIK in the registry: PGR, CB, TRV, ALL, AIG, HIG, CINF, WRB, ACGL, AFG, EG. A metric is ranked only when at least 8 insurers, the target included, have it. The group is recomputed twice a week; if the comparison is missing or older than 16 days, there is no class.

• Metrics

  Profitability: return on common equity (ROE) and the combined ratio (claims and expenses per dollar of earned premium; lower is better). Risk: prior-year reserve development to earned premium, 3-year average (a negative value means past reserves proved too large and were released; lower is better). Growth: net premiums written YoY and earnings per share YoY. Price: price to book (P/B) and P/E; a lower multiple gives a higher score. For context only, with no effect on the class: the loss ratio.

• How the class is set

  The same way as in the bank model. A category score is the mean percentile of its metrics among the insurers in the group. "Business" is the mean of profitability, risk and growth with equal weights; "Price" is the price score. "Passes the quality and price filter": profitability of 50 or more, premiums and earnings per share up on the year, risk of 30 or more and price of 40 or more (30 or more for a position the paper portfolio already holds). "Does not pass the filter": two or three business criteria fail. "Partly passes": everything else. A missing metric means no class, not a fail.

• Consensus signals

  The forward P/E (the analysts' estimate for next year) is ranked only among the insurers in the group, and realized growth is taken from this model's growth category (net premiums written and earnings per share over the year). If the company is dearer than most of the group and grows more slowly than most (both scores at 40 or below), the "Passes the quality and price filter" class is lowered to "Partly passes". When fewer than 8 other companies in the group have a forward P/E, there is no check: no downgrade and no upgrade. The check can only lower a class, never add one. The forward P/E already carries the analysts' optimism, so it is not part of the price score.

• Where the figures come from, and what the model leaves out

  The combined ratio is not tagged the same way in standard SEC data, so we take it from each insurer's annual report and check it against its components (for HIG and ACGL we weight the printed segment ratios). Everything else is computed the same way for every insurer from 10-K and 10-Q filings. ROE is our calculation on five-quarter average equity; a company that averages equity monthly can print a figure that differs by about a point and a half. Statutory capital and regulatory requirements differ by jurisdiction and cannot be compared, so the model does not use them. The 50, 30 and 40 thresholds are product policy, not a calibration: the model has not been tested on historical returns.

• The move from the general method

  JPM, BAC and PGR had no class before: the general method does not fit them, and their sector models were in preparation. The sector models classify them from rule version v0b.7. In the dry run before the change: JPM "Partly passes", BAC "Partly passes", PGR "Partly passes". The table is published together with the rule change and is not changed afterwards.