Hi CobaltBlue; Re that winnowing process... and may indeed be skill.
That is a great way to find people who have tossed 10 heads in a row... Heck, why not start with 100 million people, you ought to be able to find one who gets heads a couple dozen times in a row. Odds of that are something like 10^-8, I suppose. Not very likely, but it is going to happen to somebody.
But the trader who has a couple dozen good trades are not the ones that impress me. Some of the hotshots do around 100 coin tosses per day, 200 days per year. That's 20,000 coin tosses, and of course they're not all heads.
The better high speed (i.e. 100 trade per day) daytraders average around 6 cents per share, before commissions. Typical standard deviations on their returns are something no more than about 15 cents per share. So what's the probability that they could keep up that kind of record for a year based on luck?
It's been a while, but if 15 cents per share is the S.D., then the S.D. on 20,000 trades is about $.15/sqrt(20,000) = $0.001, and with a return of 6 cents per share, they are something like 60 S.D. away from the null hypothesis.
That's a boat load more lucky than that guy who flipped the 24 heads in a row. The chances of a good daytrader being that way by chance is something like 1/e^60 = 10^-26.
Of course it's a bit unfair to compare the guys return to random chance, he also has to beat commissions. For a good producer, commissions are about 3 cents per share (round trip), sot that means that he is trading 2 or 3 cents over commissions. Even using 2 cents, the odds of getting that lucky is something like 1/e^20 = 10^-9, which is still pretty unlikely. When you take into account that several of them have been doing it for years, the probabilities get much more small.
-- Carl |
