Then we publish the verdict — with the full method, the data, and the code. Not opinions. Backtests on real data, reads of the seller's own fine print, and the math of the game itself. If a claim has no out-of-sample test, it has no evidence, no matter what it's called.
Get the 10-point checklist →A 50%-win-rate trade list, a mechanized backtest that hugs breakeven, 12 configs with zero edge — and a Fibonacci filter that made the best strategy worse (PF 1.59 vs 1.94).
"This EA cannot profit on live accounts" — its own disclaimer, quoted verbatim. It arbitrages demo prop-firm eval conditions, not markets.
No source, no entry rule, "no martingale/grid" sold as if it were a profit claim. A neural network that "self-trains on each candle" is lookahead bias with extra steps.
Explanatory power isn't predictive power. Liquidity can't be wrong, and an alternate wave count means it's never wrong — so it's never right in any way that matters.
93% of funded accounts blow. The eval fee is the product. The challenge only wins against the edgeless — and that's who the funnel feeds it.
Subscription revenue decorrelated from your results. Survivorship at industrial scale. A shelf that's grids all the way down.
A desk that only buries scams is a hit-piece factory. We also vindicate what survives — and correct our own numbers when they're wrong.
Walk-forward validated out-of-sample. The number everyone quotes (PF 2.21) was selection bias; the honest figure is 1.65 — and we published the correction.
MT5 real-tick PF ~1.1–1.25, Sharpe 2.7. And your Python backtest overstates it 40%. The honest number is the only number we report.