Guides
Plain-language explainers on telling a real edge from a lucky backtest — the methods behind the calculators, written for working systematic and algorithmic traders.
Where Backtest and Live Part Ways — Six Defects
Three data, two code, one config defect from one intraday system, each with a guard test, plus a ten-point checklist. Paper log figures only — no profitability claim.
Read the guide →Spurious Predictability — Falsification Audit
Induced-null workflow falsification, then ΔZ and effective multiplicity. An audit gate and inflation ledger — not a deployment recipe. Paper figures only.
Read the guide →AlgoXpert IS–WFA–OOS
Plateau selection, purged rolling WFA with majority-pass + catastrophic veto, parameter lock, strict OOS — Deploy / Reject / Refactor gates. Related: WFO Python, Nested CV, DSR.
Read the guide →Monte Carlo Trade-Sequence Stress Test
Bootstrap closed-trade P&L paths to see median vs stress drawdown, VaR, and ruin probability — attributed teaching method from an MQL5 essay, rewritten in Python, with no invented Quant for Free backtest numbers.
Read the guide →Walk-Forward Optimization in Pure Python
Rolling and anchored walk-forward without a vendor platform: train/test windows, parameter stability, and why a full-sample grid Sharpe is not an out-of-sample edge — with attributed teaching numbers and rewritten code.
Read the guide →Backtest Overfitting
How a worthless strategy can post a great Sharpe by chance, three real failure cases from hands-on research, and the validation toolkit — Deflated Sharpe Ratio, PBO, purged walk-forward and CPCV — that separates a genuine edge from the best of many tries.
Read the guide →Guides vs. the course
These guides are the long-form, deep-dive companions to our free Learn course, where the same ideas are taught from scratch, in order, with diagrams and worked examples. If you're new to systematic trading, start with the course; come here when you want the fuller treatment of a single topic.
The flagship guide above takes apart backtest overfitting — why a worthless strategy can post a brilliant Sharpe ratio by chance, and the toolkit that catches it: the Deflated Sharpe Ratio, out-of-sample testing, and walk-forward validation. You can see that toolkit applied to real published strategies in our strategy teardowns, and the same ground covered step by step in Module 5 of the course. More guides are on the way.