Crazy returns mean nothing. People should look at the Sharpe Ratio and the statistical representativeness on the whole.
For example, a system whose return is 1000%, but is working only 2 months (Dec. 25 till Jan 5 is also considered 2 months) and made many trades, but they all are in one direction. How do you know that it's not just one trade split into many? Because if it is then all the above said return is nothing but luck.
"Markets are constantly in a state of uncertainty and flux and money is made by discounting the obvious and betting on the unexpected" - George Soros