Numbers first. Opinions second.
I am a researcher. I believe in measurement, in robust quantitative analysis, and in showing the work, including the parts that do not flatter it. This platform is where that belief became software.
Why it exists
Most trading results arrive as a screenshot and a single number. You cannot see the losing years or the deepest drawdown, and you cannot re-run the test yourself. I built this platform so every result comes with its bad periods attached. A strategy that has never been shown failing has not really been shown at all.
How it is tested
It uses years of intraday market data, not daily bars and not a favourable two-year window. Every mechanism starts switched off and is turned on deliberately, and the exact configuration behind a published number is kept with it. Before the engine serves a single result, it must reproduce a fixed reference run exactly. If the numbers drift by a decimal, it refuses to run until a person looks.
What “robust” means to me
A good backtest is the start of the work, not the end. Each candidate is re-run from scratch, checked year by year, tested against changes to its own parameters, and examined for how much it leans on a handful of lucky trades. Return alone proves nothing. The experiments are run with AI tools; what counts as evidence is decided by a person.
What I do not claim
A backtest records what a set of rules would have done on data that already happened. It is not a forecast. Markets change, and a strategy that worked for ten years can stop working in the eleventh. Nothing here is investment advice, and you can lose money trading.