Money Checkup analyzes U.S.-listed stocks through a value + quality lens: how cheap a stock is relative to its earnings/book value/sales, and how efficiently and safely the underlying business is run. It is a research tool, not a trading bot or a buy/sell recommendation engine — it organizes public filings and translates them into plain language.
The score is a relative ranking within the current universe, not an absolute judgment — a top score means "better than most peers right now," not "guaranteed good." Safety is currently based on the debt ratio alone; other measures like interest coverage aren't in SEC XBRL data in a consistently machine-readable form yet, so they aren't factored in. Share count used for market cap is a fixed snapshot, not updated for buybacks or issuance between refreshes.
Most home-made backtests fail in the same three ways. This one tries to control for two of them and is upfront about the third it hasn't solved:
Each simulated rebalance only uses a fiscal year whose 10-K had actually been filed with the SEC before that date, using the real filing date from EDGAR — not an assumed lag.
The backtest's universe is today's largest ~3,000 SEC filers. Companies that went bankrupt or were delisted between 2017 and 2024 aren't in that list and couldn't be included — the official Russell 3000 historical membership data that would fix this blocks automated downloads, and no free equivalent exists yet. This almost certainly inflates the absolute return numbers. Full explanation here.
A 0.15% round-trip cost is deducted (slippage only — U.S. retail brokers have largely dropped commissions). Dividends are added pre-tax.
Under this methodology, the value+quality top-20 portfolio beat the equal-weighted universe by an average of about +27 percentage points a year over 2017-2024 — but given the survivorship bias above, treat that margin as an optimistic-leaning signal, not a settled fact. Full year-by-year numbers and caveats →
SEC filings don't tag standalone-quarter revenue and net income directly — a 10-Q's income statement is tagged as a year-to-date cumulative figure (the Q2 filing reports the first half's total, the Q3 filing reports the first nine months' total). This site subtracts the prior cumulative period to recover each standalone quarter: Q2 = H1 total − Q1, Q3 = 9-month total − H1, and Q4 = the annual 10-K's full-year total − the 9-month total. Balance-sheet items (assets, equity, debt) don't need this — they're already point-in-time snapshots. "Beat"/"Miss" labels on the earnings report are a simple ±20% year-over-year net income threshold, not a comparison against analyst estimates (this site doesn't have consensus-estimate data).
Financials, dividends, audit/going-concern signal, and recent filings: SEC EDGAR (free, public, no API key). Prices and index data: Yahoo Finance's public chart endpoint. Sector and industry classification: Nasdaq's own exchange-wide screener data. Each source's own terms apply; this site summarizes and links back to the original filing wherever possible.
The common starting idea for a project like this is "catch the next big mover automatically." Backtesting that idea honestly first — the same discipline this site's own backtest page tries to model — tends to show its edge is smaller than trading costs, and that any apparent improvement from removing risk controls is often just survivorship bias in disguise: bankrupt or delisted stocks quietly missing from the test data. That's a well-documented pattern, not a one-off finding here. The more durable, if less exciting, direction the evidence tends to point to is a low-turnover value+quality approach — which is what this site is built around, together with a straight account of where its own checking still falls short.
Everything on this site is provided for informational and educational purposes only and is not a recommendation to buy or sell any security, nor investment advice. Data may contain errors, delays, or omissions. Any investment decision and its consequences are entirely your own responsibility. Past performance does not guarantee future returns, and investing carries the risk of loss of principal. The operator is not liable for any damages arising from use of this information.
This site does not require account registration and does not currently collect any personally identifying information — there is no email signup, newsletter, or login. If that changes (for example, an email newsletter or analytics), this page will be updated to describe exactly what's collected, why, and how to opt out, before it's added.
Questions: jgl2002@naver.com