The product is invite-only. The research isn't. This is where we work in the open — how we grade, what we got wrong, and what the tape is telling us this week.
Not all Form 4s are equal. The difference between a cluster buy that matters and routine purchases that don't.
Read noteA 30% drop is not a discount. The four signals that separate a real dislocation from a value trap, with worked examples from this quarter.
Read noteBuffett, Lynch, and Ender don't filter for the same thing. The disagreement between them is the point.
Read noteWe backtested 1,200 sweep alerts. The honest answer about what they do — and don't — forecast.
Read noteA Deep Dive that didn't work. What the model missed, what the analyst missed, and what we changed in the rubric afterward.
Read noteInstitutional buying is a real signal and a lagging one. How to use a 45-day-old filing without trading on stale news.
Read noteReal-time alerts feel like an edge and usually aren't. The case for two deliberate windows over an always-on firehose.
Read note