Do you believe that it can be profitable to trade a "system" build from rules found by data mining? Assume that everything (data preparation, significance testing, trading assumptions) is executed properly (whatever this means for you).
We trade what we believe, so there is no right or wrong in this. For me, it is first the idea that has to make sense. Data is only that, data. At best, it still only represents a fraction of what is really happening with the subject.
If you buyandhold equity index like SP500 you are still Momentum investing, because they change their holdings frequently. So yes, data mining is everywhere even you do not nothing.
Hello luckyputanski, I do not know what to believe in anymore. If it looks good in back test, in one year, it will look ugly with real money.
It depends on the data you mine. I mean, there are literally tons of data floating on the web every waking second. It's mind boggling just thinking about it. Unless you have the manpower and the resources (eg. $$$), you can forget it about it.