trickybeck wrote:
Yeah, a "guarantee" requires asking 100% of the population. That's not a sample anymore.
You can get a 99% accurate estimate with a sample size of much less than 1/2 of 1% of a population. Which I already showed. Apparently 99% accuracy isn't an "accurate prediction" according to Almalieque.
Edited, Nov 3rd 2012 9:02pm by trickybeck
Did you read any of what I just wrote? You don't need to 100%, How many times do we count every vote before declaring a president? We only care about states that carry enough delegates to make a difference. Those other votes get counted, but no one cares because they wont change the outcome of the race. Just like pulling the 51st blue marble in my example. The remaining marbles are irrelevant. You can count them if you want, but they wont matter.
No matter what proof you might show, it isn't mathematically possible. I'm referring to a guarantee, not estimation. Even your 99% estimation in this case is false. Anything can happen between today and Tuesday to change the outcome. These polls are literally only good for the moment that they are released and then are no longer good anymore. Hence on why they are pointless.
Jophiel wrote:
"Guarantee" is a strawman. No one guarantees polling results.
Read above
Edited, Nov 4th 2012 4:14am by Almalieque