The iterative process of tuning an algorithm’s parameters to improve performance. After backtesting initial rules, quant developers run optimization loops: repeatedly adjusting parameters (e.g. indicator periods, thresholds) and re-testing to maximize metrics like net profit or Sharpe ratio. This cycle often combines brute-force grid search, genetic algorithms, or walk-forward analysis (rolling test windows) to find robust settings. An optimization loop might involve thousands of simulations: for example, backtesting a strategy with varying stop-loss levels and retaining those that yield the best balance of return vs. risk. Proper optimization reduces overfitting and enhances real-world edge.

Back to Robots Glossary Index