A published case, in full

This is everything a research case contains.

One real case, dated, exactly as it was published — the reasoning, the numbers behind it, and the risks recorded at the same time. No account, nothing withheld, nothing summarised. It is a past case shown as a worked example, not a current view and not a recommendation.

$STC
STEWART INFORMATION SERVICES CORP
A High conviction

Published · scored on the filing date, not the date you are reading this.

InsiderApel Thomas G Dir, Director
Reported purchase$851,130
Price when scored$43.54
Scenario upper bound$47.89 10.0% above the scored price
Backtest weight20.0% weighting inside the backtest, not advice on position size
Review window30 days

The reasoning, as recorded

Cluster of 3 unique directors bought ~$851K on Oct 23, one raising his stake 42% and all three buying into the post-earnings washout — stock is 1.9% off its true 52w low after a 23% monthly drawdown with RSI 29.7 (capitulation). Title-insurer is profitable with 13% revenue growth, trades well below analyst fair-value estimates, and buyers stepped in right as the EPS miss on housing/rate headwinds hit. Capped at A because it is director-only (c_suite_count 0) — the hard rule blocks A+ despite the cluster, deep oversold setup, and multi-source confirmation.

What was recorded against it

Director-only, no C-suite buyer. Net flow not checked (net_flow_checked false). Below both MA20 and MA50 — falling knife, but cluster + capitulation RSI justify the entry. Financial sector lagging SPY by ~5.8% over 3mo; Q3 EPS missed on persistent mortgage-rate/housing headwinds expected into next year.

Check it yourself

Every case starts from a filing anyone can read. This one is Apel Thomas G's Form 4 dated 2026-10-23.

Form 4 filings for STC on SEC EDGAR ↗

That link opens the issuer's full Form 4 list rather than this one filing — the archive behind this page does not carry an accession number for it, so pointing straight at the document would mean guessing the URL. The insider's name and the date above are what you need to find it in the list.

InsiderScanner identifies unusual insider purchases for research. Ratings, scenarios and historical statistics are model outputs, not investment recommendations.