Plain-English explainers on the moves the framework was designed to prevent. Not strategy. Not stock picks. The boring math underneath every trade that determines whether you survive long enough to compound.
Three tiers, twelve lessons each. Each lesson ships as an article (with an interactive widget or visual aid where the math directly benefits) plus a downloadable slide deck. Sequenced so the math floor lands before the chart-reading. Built one tier at a time — beginner first.
Most traders treat 2:1 risk/reward as a rule of thumb. It's not — it's the math floor where a 50% win rate stops being a coin flip. Below 2:1 you need to be right more often than you actually are. Why most retail traders overestimate their hit rate, why R:R isn't something you choose (it's something the chart hands you), and how Swing Deck's structural R:R replaces a hardcoded 1.5 with real per-ticker math.
Read article →CEO sells $25M of stock — bearish, right? Often no. Rule 10b5-1 plans let insiders pre-commit to sale dates months in advance, while not in possession of material non-public information. The Form 4 will say SALE; the AI verification on AAPL caught that all 9 sales were calendar-driven. How to read the chip's color, what the new SCHEDULED_SELLING label means, and when insider sells actually do signal something.
Read article →The stop you set on Friday isn't the stop you get on Monday's open. Why a 2× ATR stop becomes a 5× ATR stop after earnings, the structural difference between a stop that protects capital and a stop that defines your position size, and the math behind why the dashboard's chandelier exit + raise-stop automation exists.
Read article →The framework's whale_sentiment score used to be a closed loop — it would compute its own value, then narrate that value back. v6.3 added a verification layer that runs web_search against SEC filings + dark-pool reports + analyst sentiment and says CONFIRMED, CONFLICT, or WEAK SIGNAL. How to read the agreement %, when a CONTRADICTED verdict is more useful than the underlying score, and the PLTR case where the verification caught $435M of insider selling.
Most "AI trading apps" in 2026 ship a chat box. We deliberately don't. Eight per-ticker AI surfaces, each fired by a specific framework moment, each producing structured output. The discipline trade-off: a chat box lets you talk yourself into trades; surface-bound AI talks you out of them. The architectural commitment, the prompt-design rules, and the per-call cost numbers that actually result.
Read article →Cadence: roughly one article per release plus topical pieces when something interesting comes up. Subscribe via the blog RSS feed — Learn articles are syndicated alongside release posts.