Signal Bench
Publicly logged · resolved against real candles

Crypto signal performance, in the open

No screenshots of winners. Every Signal Bench signal is logged the moment it fires and resolved against live candles — wins, losses and near-misses.

Two records, never mixed

Good performance reporting keeps simulation and reality apart. Signal Bench does too:

  • Backtested — the exact live engine run over recent history with no lookahead.
  • Live verified — real signals the engine emitted, resolved against real candles.

They're shown on separate pages so you always know which number you're looking at.

Metrics that actually matter

Win rate alone is misleading. Alongside it we publish profit factor, expectancy per trade, average win vs average loss and "Avg TP progress" — because a 40% win rate at 1:2.5 risk-to-reward can still be profitable, and a 70% win rate with tiny wins and huge losses is not.

Frequently asked questions

How do you measure signal performance?

Every signal is logged when it fires and resolved against real candles — win (hit target), loss (hit stop) or how close it got. We show win rate, profit factor, expectancy and average TP progress.

Do you separate backtested from live results?

Yes. Backtested (walk-forward, no lookahead) and live-verified results are shown on separate pages and never combined into one number.

Is the track record cherry-picked?

No. Losses and NO-TRADE outcomes are shown alongside wins. The record is public and immutable — levels are never edited after a signal is logged.

Judge for yourself, then start free

Read the public record, then create a free account with a 3-day Premium trial.

Create a free account