setting type…
After each weekly scan: the new SIGNAL-grade wedges on the S&P 500 and ASX 200 — ticker, direction, entry, stop and targets. Need a scan sooner? Sign up and reply to any email — on-demand runs happen by request. No spam, unsubscribe any time.
The octopus reads seven hundred charts every week so you don't have to.
☕ Fund the press runEvery week the model reads the daily chart of every S&P 500 and ASX 200 stock and hunts for one specific structure: the wedge — price coiling between two converging trendlines, energy compressing until it has to pick a side.
Both lines slope up but support climbs faster — buyers grinding into shrinking room. Typically resolves down.
Both lines slope down but selling pressure fades into a coiling base. Typically resolves up.
1 · Geometry first. Deterministic code draws the two trendlines, counts the touches, and measures convergence — the same wedge is found every time, no black box deciding what a pattern is.
2 · Then a judge. A LightGBM classifier, calibrated with isotonic regression, scores each geometric wedge: the printed confidence is the empirical probability the wedge resolves into a clean breakout move. Above the gate it's stamped SIGNAL; below it stays on WATCH.
That precision was measured on three years of data the model had never seen — and that holdout is now spent: we don't re-test on it, and we don't retrain to chase it. The model you're reading is frozen.
The confidence calibration was fit on US daily data; on other markets (e.g. the ASX 200) the same frozen model is applied unchanged — the geometry is universal, but treat the printed percentages there with an extra grain of salt.
RisingOctopus V1 is a pattern-detection tool, not a P&L-validated trade signal. Confidence is the probability of a clean pattern resolution — not a probability of profit. Wedges carry weak directional base rates. Not financial advice; do your own work.