The thought never occured to me until I started trading the markets earlier this year. (Talk about instant fixation ...)
I now have a database of roughly 60 million records. I run batch jobs nightly which update data across 1500 stocks, three differnt data feeds, looking to add a fourth. Inet, Yahoo, Ashkon, ...
Weekly, Daily, 60 Minute, 30 Minute, 15 Minute, 5 Minute intervals across 5,12,16,24, 52 period increments...
All the popular indicators calculated and queriable using sql across umpteen diffrernt tables. Java/Oracle/Xml..
I think I've finally arrived at a system with a positive expectancy after months of non-stop work. I've arrived at a method to find highly-volatile stocks stuck in a wide (3-5 day) trading range, with a high-level of predicatability. (ie; stock is moving with regularity between peaks/troughs for the past 10-15-30 days.)
I've basically shunned the outside world for months until I can make this work ... God, I hope I can make this work ...
I don't think this makes me an "idiot-savant", but I sure as hell feel like one ! ... the idiot part anyways.