Verified track record

By grade

What this record does not claim

• Net R subtracts an estimated spread and slippage taken from a static per-instrument table. It is an estimate, not the spread that was quoted at the moment each signal was filed.
• Resolved means the stop or the first target actually printed on closed bars. Signals that timed out are reported on their own line and are never counted as a win or a loss.
• Scans with no directional read are kept in the record for the audit trail and excluded from every figure above.
• Whether a signal was actually traded was not linked to scans before this record existed, so taken-versus-skipped only accumulates from now on.
• Signals filed before 18 September 2026 were published without any check that the entry price was still reachable. Measured on 500 of them, price had already closed past the planned entry on 418, by a median of 0.69 of the trade's risk, and 145 resolved on the very first candle. From that date a signal is refused if price already sits more than 0.5 of its risk past its own entry, and the distance is stored on every signal filed.
• Because of that change, figures from before and after 18 September 2026 are not the same measurement and should not be pooled. Re-scored on entries that were still reachable at filing time, the older signals average -0.15R to +0.16R depending on how strict the reachability test is, which is materially worse than the headline figures for the same period.
• How long a signal may run before it expires is now set per market from its own measured time to resolution, replacing a single clock applied to every instrument. Signals with too small a sample keep the old conservative clock. Same-bar fills are still resolved pessimistically: a candle holding both the stop and the target counts as a stop.
• Only signals produced by our own engine appear here. Signals filed through the inbound endpoint by outside tools are sealed and resolved on identical terms but are never counted in these figures.

Every published signal (791)