All Posts (Page 35)

Curve fitting occurs when a trading strategy is optimised so thoroughly on historical data that it captures noise rather than genuine patterns. The result looks compelling in backtesting and fails in live trading. Understanding how curve fitting happens — and why it is so easy to do by accident — is essential for anyone building or evaluating systematic trading models.

The SMA Trail position manager exits a trade by trailing a stop loss against a moving average rather than holding it at a fixed distance from entry. This allows profitable trades to run further during strong trends while still providing a structured exit when momentum fades. In darwintIQ, SMATrail is one of four position manager types that can be paired with any entry logic and regime filter.

Volatility in financial markets is not constant. It moves through cycles of compression and expansion, and those cycles correlate strongly with changes in market regime. A model that performs well in expanding volatility may suffer in quiet, compressed markets, and vice versa. In darwintIQ, the TrendMatrix and regime classification provide real-time context about where a market sits in its current volatility cycle.

Walk-forward validation is a method for testing trading strategies on genuinely out-of-sample data by repeatedly training on one period and testing on the next. It exposes overfitting that standard backtesting conceals. In darwintIQ, continuous rolling evaluation serves a related purpose — ensuring models are assessed against current market conditions rather than optimised history.

Profit Factor is a ratio that compares the total profits of a trading model to its total losses. A value above 1.0 means the model is generating more than it loses. In darwintIQ, Profit Factor is one of several metrics used to assess whether a model has a genuine structural edge rather than a run of good luck.