Notes from the strategy validation workbench
Blog
Practical writing on backtesting, walk-forward validation, and the discipline of turning trading ideas into evidence.
Evidence, not vibes: what a useful trading journal actually records
Most trading journals record moods and outcomes. A useful one records decisions: the entry thesis, the invalidation condition, and review conclusions you can actually query later.
From Pine script to proof in 15 minutes
A hands-on walkthrough: import a TradingView-style moving average breakout into TideView, run an AI review, backtest it with fees and slippage, walk-forward test it, and export the report.
Walk-forward analysis, explained without the math degree
What a single backtest cannot tell you, how rolling train and test windows work, how to read the window results, and the signals that say a strategy is not worth more of your time.
Why most backtests lie (and how to catch yours doing it)
Five ways backtests flatter a strategy: look-ahead bias, survivorship bias, missing costs, overfitting, and in-sample-only reporting. Each with a concrete example and a self-check.