Compare S&P 500 companies across eight investing methods, using dated filings and prices. Start with the candidate shortlists or filter the full table below.
Screen results are research inputs, not recommendations. Company investigations live under Companies; simulated holdings live in the Paper portfolio.
All stocks
About this record
The Claude Trader is a sourced stock-research record: every US-listed S&P 500 operating company scored against eight published rule sets, with each number linked to its SEC filing or price source, and facts and scores for the full eligible universe published alongside. The research record itself holds no positions and publishes no house recommendations or price targets (the separate paper portfolio on the Portfolio tab is an experiment, not advice). The Fair value column is a labelled model estimate from SEC filings, not a price target or recommendation (method and dates) — you (or an assistant you point at it) build and judge any shortlist from the evidence.
Research universe: S&P 500 constituents, present-day (500 companies, membership as of 2026-10-14, present-day research universe: current index members only; not a survivorship-free historical set (companies that left the index are absent)). Constituent source; the full list hash is in /q/stats.json. The local database retains data for all 5,305 eligible companies; this public grid scores 500 in the active scope. 821 in the full eligible universe have no US-GAAP net income in XBRL (mostly IFRS filers); that blocks only Graham and Magic Formula, which report it as insufficient data. Not eligible from the 6,062 listed issuers with a CIK: 116 etf / exchange-traded fund; 25 acquisition company with no revenue (spac); 277 blank check company (sic 6770) with no revenue (spac); 271 fund or commodity trust; 41 not common equity (warrant, right, preferred, note or spac unit); 27 not in the current exchange symbol directory (delisted, renamed or acquired). Definitions in /q/stats.json.
Under each score: the date of the oldest current-period input it depends on. “partial” means a test could not be computed from fresh filings; the missing test numbers are shown, the score is the sum of the rest, and partial rows are hidden unless you untick “Complete scores only”. “stale” is a warning: a current-period input older than 120 days that is still used; inputs older than 550 days are excluded outright and make the test insufficient. “mismatch” marks an input where the two sources disagree. “n/a” means not scored under that philosophy’s rules or insufficient data; the reason is in the stock’s JSON. The table shows the first 500 rows of the current filter; use the filters or the CSV for the rest.