105: A Dangerous Optimization Assumption

This podcast episode features an interview with Perry Kaufman, discussing the pitfalls of linear parameter testing in trading system optimization. The host introduces the topic by highlighting a common misconception among traders regarding optimization results. Kaufman explains that testing parameters in a linear fashion can lead to inaccurate conclusions, particularly favoring longer-term strategies due to the smaller percentage differences between larger parameter values. He advocates for testing parameters using percentage increments rather than absolute increments to gain a more realistic understanding of the system's performance across different timeframes. The episode concludes with practical advice on how to conduct percentage-based testing, even with limitations of existing testing platforms.