The six systems in this module are rule-based and therefore testable. A setup only becomes a system once its rules show a positive expectancy across many trades. Backtesting is the meta-discipline that separates a plausible setup from a viable system.
In-Sample, Out-of-Sample & Walk-Forward
Split the history: on the in-sample period you develop and optimise the rules, on the held-back out-of-sample period you test them on unseen data. The walk-forward analysis rolls this window forward repeatedly (optimise → test → shift) and is the toughest test: it simulates how the system would be re-tuned periodically in reality.
Spotting curve-fitting / overfitting
The most common self-deception: you optimise the rules until the past looks perfect, and then fail in the future. Warning signs:
- Too many parameters or „special rules" for individual market phases.
- A suspiciously smooth, almost linear equity curve.
- Performance collapses out-of-sample.
Evaluation metrics
| Metric | What it measures |
| Profit factor | Gross profit ÷ gross loss. > 1 = profitable, robust from ~1.5. |
| Expectancy (R-multiple) | Average win or loss per trade in multiples of the risk (1R) — the core expected value. |
| Max drawdown | Largest peak-to-trough capital decline — decides whether you can psychologically stick with the system. |
| CAGR / MAR ratio | Annual return, or annual return ÷ max drawdown — return relative to pain. |
| Sharpe / Sortino | Return per unit of risk; Sortino penalises only downside volatility. |
| Win rate × R/R | Win probability times reward-to-risk — a low win rate can be offset by a high reward-to-risk ratio. |
Avoiding biases
| Bias | Trap |
| Survivorship bias | Testing only names that still exist today — bankruptcies and delistings are missing and flatter the result. |
| Look-ahead bias | Using data that was not actually available at trade time (e.g. closing prices for an intraday decision). |
| Ignoring costs | Without realistic slippage + commission every high-frequency system becomes artificially profitable. |
🎯 From backtest to forward test: A backtest remains a hypothesis about the past. The consistent forward test is your real trading journal: with sTraderZ.com you document every trade and evaluate it via strategy tags and dimensions — so you can see whether the edge found in the backtest also holds live.