Show Parlay

[Dunbar]

Originally Posted by jman5581

I would be more than happy to start fresh, but what's wrong with me using picks that I posted to this board (in advance of the race itself) for the last contest. I've just posted the payouts and cumulative returns at the top of this thread. As far as I'm concerned that meets the criteria of any pick I would make in the "experiment." I made the picks...it should at least count for the 1st of 20 since I disclosed them in the Showvivor thread.

jman, I don't doubt that you made and posted the picks. But it's bad statistics to decide after the fact that you are going to include some results you already obtained in a longer study that wasn't specified ahead of time.

Let's say I walk by a roulette table, and I see that red has come up 5 times in a row (on the history display). If I see one more red come up, does that make it REALLY likely that red is a good strategy?

If you had said ahead of time, 'guys, I'm going to test a show parlay theory and post all my picks here', then fine. Those first 2 sets would belong. But it's bad statistical practice to decide after-the-fact that certain bets count.

At any rate, if there is any merit in show parlays, you should be able to demonstrate it without resorting to past bets.

--Dunbar