"normanstrong" wrote in message
news:vrEHb.673278$HS4.4771970@attbi_s01...

"Sockpuppet Yustabe" wrote in message
...

"normanstrong" wrote in message
news:vilHb.56707$VB2.101814@attbi_s51...

"Scott Gardner" wrote in message
...
Does anyone know the general equation that tells you how many
trials
a
subject must complete successfully to achieve a desired
certainty
that
the results weren't from guessing?

In other words, for "N" trials, what number "M" of them must the
subject complete successfully to be "X" percent sure the results
weren't from guessing?

If the probability of guessing right in a single trial is 50%, and
there are 10 trials, the chance of getting all of them right is 1
in
1024 (1/2^10).

If you allow for 1 mistake, your chances are improved about 10
times
(1.1%)

2 mistakes improve the odds another 5 times (5.5%) This is as far
as
I would go.

Another interesting question is: How many times do you have to do
this test before you have an even up chance of correctly guessing
8
out of 10 tries? The answer is 12 times. IOW, if 12 people try
to
get 8 out of 10 right, half the time at least one will succeed--by
luck alone.

you are ready to dive into a cesspool of fallacy.
let me blow the whistle. GO!

Yes. I'm ready to dive--but I haven't dived yet. The entire issue is
one of probabilities. No matter how many trials you pass
successfully, there's always a finite probability that it was luck.
The best we can do is reduce that probability to a minimum.

Can you predict the outcome of a coin flip? Suppose you flipped a
coin 10 times and guessed right 8 of them. Does that mean you can
actually can predict coin flips? Now add this complication: You
flipped a coin 10 times and you were WRONG 8 of them. What does this
mean? Think about it.


It means that I do a bad impersonation of Madam Flora.




----== Posted via Newsfeed.Com - Unlimited-Uncensored-Secure Usenet News==----
http://www.newsfeed.com The #1 Newsgroup Service in the World! 100,000 Newsgroups
---= 19 East/West-Coast Specialized Servers - Total Privacy via Encryption =---