Proof Center

Cumulative % gain across resolved picks, in the order they resolved — wins and losses alike, exactly as logged. Price moves from entry, before fees, slippage, liquidity, or taxes.

Win rate = wins ÷ resolved picks. Avg gain = mean % move from entry across resolved picks. Every rate sits beside its resolved denominator — a loud percentage on a tiny sample is still noise.

Losses are printed next to wins. A track record that hides them isn't one.

1. Trigger = entry. A pick is logged the day a setup reaches its pivot — ticker, pattern, entry price, date. Picks are never backfilled and never deleted.
2. Win = target before stop. Each pick closes as a win when its upside target is reached before its stop, and as a loss when the stop is reached first.
3. Data window: the last 90 days (since Jul 29, 2026). The window does not move to flatter the numbers.
4. Honest limitations: the sample is thin — 9 resolved picks is not enough to judge a system. Figures are price moves from entry, before fees, slippage, liquidity, taxes, or individualized risk. Every logged pick in the window is included — nothing is cherry-picked.

How to read this page: the historical replay and the live 30-day attempt audit are different measurements. Replay rows show target before stop and Avg R for a fixed rule set; live rows show whether a logged attempt held above its pivot under the resolver described below. Neither measurement predicts a future winner.

Generated Oct 27, 2026; source-bar window 2024-11-01–2026-10-26. Rows below show no percentage below n=30. Read the exact definitions.

Historical replay: +20% target before −8% stop within 40 sessions, after an executable post-signal fill; sample and source-bar window are shown above. Avg Return and Avg R include targets, stops, and full-window timeouts. Historical and illustrative, not predictive. Each pattern's page in the Pattern Library also shows stored-bar follow-up observations.

Live attempts logged before pattern metadata was captured appear as "Unclassified" in the splits below. These rows use the last-30-day window shown at the top and display the resolved denominator beside every rate.

This is an audit curve of breakout holds versus failures, not a P&L backtest.

We paper-trade our own scan. Every trading day, the top 5 setups from the day's scan are opened as paper trades at their published entry/stop/target and tracked mechanically to exit. This is the honest answer to "does the scanner actually work" — the numbers below include every loser, because hiding them would make the whole exercise pointless. This section is the unfiltered research ledger: every top-5 pick, no readiness filter, no gates — the raw material the model portfolio filters.

By setup type

The ledger — every paper trade, newest first

The rules (fixed in advance, same for every trade)

• This is the unfiltered research ledger: every top-5 pick from the day's scan, with no readiness filter and no gates — the raw material the model portfolio filters.
• Every trading day, the top 5 scan setups are paper-traded — no more, no fewer, chosen by the site's canonical setup ranking.
• Each trade uses the setup's own published plan: the entry, stop, and target the scans page showed that day.
• Entry is assumed filled at the plan's entry price. No slippage or fees are modeled — real trading would do worse on both.
• An entry is booked only if the plan price actually traded on the entry day (inside that day's range, 1% tolerance). A price that never traded is not a fill: the pick is logged as 'phantom entry' — shown in the ledger, excluded from the stats.
• From the next trading day, a daily low touching the stop exits at the stop (a full -1R loss). Otherwise a daily high touching the target exits at the target.
• If a stock gaps down through its stop, the exit is booked at that day's open — the price that actually traded — not at the stop.
• A single bar touching both stop and target counts as a stop-out — the conservative read.
• No stop and no target after 20 trading days exits at that day's close (time-stop).
• Picks that can't form a valid plan are logged as 'no plan' — shown in the ledger, excluded from the stats.
• Win rate, average R, and expectancy count only resolved trades. Losers are never removed.

Paper trading is hypothetical. These are not real trades, real fills, or real returns — no slippage, no fees, no liquidity limits, no taxes. A limit order at the plan price is assumed filled; real trading fills worse. Past paper performance does not predict future results. For educational research only. Not investment advice.

Sample size matters. A 30-day live audit can be thin, and the historical replay is a fixed-window model of setup rules, not brokerage fills, slippage, taxes, liquidity, or individualized risk. Past scanner behavior does not guarantee future outcomes. For educational research only. Not investment advice.

This page is a record of past scanner picks, published for education only. Past performance does not predict future results, and nothing here is investment advice or a recommendation to buy or sell any security. Trading involves risk, including the possible loss of principal.