[soVeryTired]

Even running a statistical test won't cut it. How could a statistical test account for the fact that an internet celebrity trader might be less likely to discuss his losing trades?