It's my experience that curve fitting is only bad if rules are fitted to a random price characteristic. There's the assumption that whatever is used is being fitting to a non random characteristic. That makes curve fitting a good thing however there's probably a better word to describe the process.
Market Data 
Markets 
AI
