A close shave. Recently, I found myself working day and night on fine-tuning an exit strategy for one of my day trading systems. Delving into statistical probabilities and portfolio theory can quickly turn into a quagmire of data, models, and losses. One can be tempted to add more and more criteria, rules, and layers to achieve a high probability trading system.This is know in the industry as "curve fitting".  C urve-fitting refers to the process of creating a model that closely fits a specific set of historical data. This can lead to models that perform well on past data but poorly on new, unseen/ future data due to over-fitting. Traders may be misled by profitable back tested results if a model has been curve-fitted. The model may show impressive returns and low risk in back-tests, but these results cannot be replicated in live trading. This is a common mistake that traders make and can be extremely frustrating when a trader curve fits a strategy that produces...