The standard, not the secret sauce.
The internals stay private, but the standard behind them is worth stating plainly.
Nothing is computed mid-candle.
Signals are evaluated only on closed candles, so what you saw at the time is what you'll find in the history a month later. Nothing gets quietly revised once the move is obvious.
Costs and bad data don't get a pass.
Fees and slippage are priced in from the beginning, and if something upstream isn't healthy, it stops contributing rather than contributing badly.
Outcomes are recorded.
Every signal is stored with what happened next. The record is measured rather than remembered, including the parts we'd rather it didn't include.
What using it actually looks like.
01
A candle closes and a signal is evaluated.
Never mid-bar, so nothing repaints.
02
You're alerted, with the reasoning alongside.
The chart opens with the signal marked on the candle that fired it.
03
The outcome is recorded.
Whatever happens next becomes part of the record automatically.
We could publish a backtest today. Everyone does.
But a backtest of your own design, scored on the same data it was tuned on, isn't evidence, it's a well-dressed opinion. We'll publish performance once a frozen version has survived a validation period on genuinely unseen data, with the method alongside it so you can judge it yourself. Until then: no numbers.