Every dossier's Track Record is built the same way, from public price data — so a verdict is a rule, not an opinion.
For each dated call we take daily closing prices from CoinGecko for the relevant asset (or the ETH/BTC ratio), from the day of the video to today. The chart marks three points: when he spoke, the lowest close after the call, and now. “Drawdown” is the deepest close-to-close move against the call after it was made.
The thesis or price target was reached and held — with no drawdown deeper than ~10% along the way.
In profit now, but only after a meaningful drawdown (roughly 15%+) that would have tested conviction.
Part of the call worked and part didn't (e.g. a floor held but a paired trade went the other way).
A directional idea that hasn't been confirmed inside the window — still open as the cycle plays out.
Price violated the level he stated — e.g. traded more than ~15% below a called “bottom” — at some point, even if it later recovered.
Some traders post their calls as pictures: a chart with the expected path drawn on it, or a TradingView long/short box. Those are scored on Binance hourly candles (spot, or the perpetual when the chart is the perpetual), with one fixed rule set — the verdict is computed by a script, not chosen:
The call's window is where the drawing ends on the time axis, or the stated timeframe (“2-3 weeks”), or 90 days if no time is given. A level reached later, up to twice the window plus 7 days, still counts and is marked late — hand-drawn timing is approximate, price is what the call is about.
Levels read off a drawing, or rounded numbers in text (“28”, “103”), count within 1.5% (never more than a quarter of the move to them). Prices printed by a position box are exact. When a level counts only under the tolerance, the card says so and gives the date the exact level traded.
Levels are checked in the drawn order, each from the previous one. A level is void if price first moves against the call by more than the distance to it (the implied stop). A turning point fails if price overshoots it by more than 10% before the next move. Played out = every level reached in order. Half right = some reached, or the final level traded out of order, or price got at least half way. Open = the late window is not over and nothing is void. Otherwise went against.
The entry must trade first. Then whichever of stop or target trades first decides: target = played out, stop = went against (even if the target comes later — the card says so). If stop and target fall inside the same hourly candle, we score the stop. Never filled = half right (no trade happened).
Played out if the best move in the called direction is at least twice the worst move against it and at least 5%; went against in the mirror case; otherwise half right.
Self-reported results (exchange PnL screenshots, “21% from my signal” posted afterwards) are reported as claims and never enter the score. A post the author deletes after we captured it stays on the record, marked as deleted.
It's a transparent read of public statements against public prices — not a complete audit of a career, and not financial advice. Calls marked open are revisited as the cycle plays out. Spotted a mistake? Every dossier has a correction link.