OUR OWN TRADES · MARKED IN PUBLIC · WIN OR LOSE
Every guru shows you their wins.
We never delete a loss.
We bet against the crowd in public and mark every file every week. 7 are open right now: some up big, some red, none ever deleted. A win and a loss get the exact same page. How the finish line and the checkpoints work is in the board below.
running numbers as of 2026-08-07 · 4 already past the line that counts
LIVE FILES · ONE AXIS + THE RECORD
SORT7 open · 4 over the bar · 3 running · numbers as of 2026-08-07 · prices: continuous front-month file (roll gaps included), close-to-close, provisional until Friday grades · click a row for its check weeks + file
every file bets against the crowd: crowd long, we track the short · crowd short, we track the long. the bar is the line a move has to clear to count as a real turn: a jump too big to be this market’s normal weekly noise, and it grows the longer the trade runs (young files need less than old ones). the lime line on each row is that bar. the move is graded only at five fixed check weeks after entry (weeks 1 · 2 · 4 · 8 · 13, confirmed by Friday’s pipeline): cleared the line at a check week = OVER THE BAR. a grade is history and it sticks: if the % later slips back, the chip says NOW UNDER next to it. between check weeks a file can sit past the line and still read RUNNING.
the bar on each row = how far the trade has moved toward its finish line · every live mark here is ours, losses included · how often bets like this one paid off here since 2006 is the graded record, and it opens in the trade file · numbers settle each Friday · the exit is the clock: the window simply closes at week 13, win or lose · a measurement, not a recommendation · new to the words? the 60-second glossary →
THE RULE · A MEASUREMENT PROTOCOL, NOT A TRADE RECOMMENDATION
Every extreme is entered blind, on purpose. The crowd hits the top or bottom 5% of its own 3-year range (back-to-back weeks count once). We enter at the first open after Friday’s release, against the crowd. The move is graded against its bar (0.75× this market’s own volatility) inside 13 weeks. No filters, no discretion. This board answers one question: what is the extreme alone worth? The published answer: not much. Most cases do not turn. We tried bolting “improvements” onto the rule; tested on years they had never seen, they failed the bar we set in advance.
If you’d wait for confirmation first (the crowd’s first cut, open interest draining, a close against the trend): good instinct. Those are measured too, in every file under “what marked the turns”. As evidence, not as a promise. Full spec in the methodology →
CLOSED LAST WEEK · WIN OR LOSE, ON THE RECORD
These are our fades. Pro opens every file.
The board is free proof. Pro opens the graded conditions and the odds on every crowded market · or read one market’s full file worked end-to-end, free.