Overfitting check
Drop in your MetaTrader 5 Strategy Tester report. We tell you whether it is evidence of an edge or evidence of a search.
What it looks at
Trying 500 variants and keeping the best inflates the result even when nothing works. We compare your Sharpe with what a search that size produces by luck alone.
If moving one parameter by a single step collapses the result, you fitted the curve. This one needs the optimisation file.
Where your in-sample winner lands in the out-of-sample distribution. Below the median means picking it was worse than picking at random.
Resamples your trades to estimate expected drawdown, probability of ruin and the range of outcomes.
A backtest that charged no commission is not a backtest of anything you can trade. We compute how much extra cost per lot the edge absorbs before it dies.
What this does not do
Monte Carlo resamples the same trades, so a strategy that was fitted into existence gets a beautiful confidence interval around a fantasy. It measures sequence and sample risk. Overfitting is the job of the deflated Sharpe and the optimisation file — different questions, routinely sold as one. And none of it says a strategy works: only whether this backtest is evidence.
Your file is not uploaded
Every calculation happens in your browser. A MetaTrader report contains your broker, your server and your account number; they are discarded as the file is read and never reach the screen or any server.