Originally Posted by jman5581
I guess the way I see it, I was at the roulette table playing....so, by your analogy, it would be okay to use as part of the sample. I could very easily say, "Hey, I picked red, I'll try red 19 more times and that'll make 20"....
If I do this, it will be a rather long experiment. I will likely use the most recent picks I made as the first sample because they're there for everyone to see if they wish, so long as that thread doesn't get deleted.
If it's even remotely close at the end of the experiment I suppose I could always do one more round.
This would be about like Grits hitting a $100 exacta and telling randall, "see? let's compare exactas to place bets, but I want to start with my $100 exacta."
jman, you will get into trouble down the line if you don't understand that a test has to be well-defined before the test starts. You will find angles that look good, but are really just flukes of the way you are measuring.