MQL5 EA Audit: Going Beyond Backtest Results
Backtest results tell us what happened. Code auditing helps us understand why.
I’m currently developing YV EA Audit, a browser-based tool for auditing MQL5 Expert Advisors.
The focus is on:
Analysing EA source code and trading componentsVerifying indicator implementations and calculationsTracing how each component reaches the actual trading logicIdentifying unused components and potential implementation issuesProviding evidence-based audit findings rather than judging profitabilityI’ve completed the initial engine development and have now moved into real-world validation, testing the audit engine against actual MQL5 EAs.
The current validation batch has successfully passed 5/5 real EA tests.
The goal is simple:Not to prove that an EA is profitable, but to verify whether the code is actually doing what the strategy claims it does.
I’d be interested in hearing from experienced MQL5/EA developers and traders:
What would you consider essential when auditing an EA beyond its backtest results?
Especially interested in feedback on indicator calculations, execution flow, hidden logic issues, and risk-management implementation.
