Transparency report

Most black-box EA shops label every strategy “consistently profitable” and ask you to pay. This page is the opposite: it is a live count of how many strategies our own audit rejected, how many passed, and under which specific test. The numbers below are read from the live library on every request. Nothing is rounded up and failed checks are not removed.

The numbers, as measured

Strategies audited3794
Flagged by the audit2128
Flag rate56.1%
Pending (checks still running)1656
Pass rate0.3%
Passed the DSR overfitting check1
Passed the significance check513
DSR threshold used0.90
Average DSR across the library0.14
Average PSR across the library0.84

Why the pass rate is not 100%, and why that is the point

A 100% pass rate would mean the audit is not testing anything. Our pipeline rejects strategies that never trade, strategies that always hold a position, strategies whose discovery record cannot be reproduced, and strategies whose out-of-sample behaviour collapses. When a strategy fails a check, the failed check stays visible on its page instead of being quietly deleted.

The two statistical checks are deliberately separated. The Deflated Sharpe Ratio adjusts for how many variants were tried, and the significance test asks whether the result could plausibly have come from chance. A strategy can pass one and fail the other, and we report them separately rather than merging them into a single flattering badge.

How to verify any of this yourself

Interactive detail

The panel below adds live charts and per-check breakdowns. Everything in it comes from the same source as the table above.

机构级诚信报告 · 真实聚合 · 不美化

大多数策略过不了我们的审计,而我们告诉你是哪些

市面上的黑箱 EA 会把每一条策略都标成「稳定盈利」诱你付费。我们反其道而行——每条策略都留检验痕迹,过不了的一律如实标红;而「通过了」到底是哪种通过,我们也不含糊:绿标里真正同时过 DSR 过拟合检验与统计显著性检验的有几条、依据只是走前向验证的有几条,下面分开列,绝不含并。数字直接来自引擎产出,实时聚合,绝不修饰。

严重未通过审计
—
全库策略被检出过拟合 / 统计不显著的比例
仅承认可信—

加载全库诚实统计中…