im saying there are several different polling methods.
Do you know them well enough to know why you dont like each?
My original post to you was saying we still dont know the results...
...those polls could still be 100% right.
There are things that cant be quantified hence why I said the ghost in the numbers.
So let's take Nate Sliver's methodology of aggregating numbers. Aggregate polls mean that is supposed to make them more robust, but if everyone has flawed methodology your not aggregating robustness your make the aggregation more concentrated.
Second is the weighting of polls in the aggregate system. The more polls that are correlated together that move like to each other again it's not robust it's more concentrated.
Third fat tails. Rare events. Say we get Talfgar polls that look janky with their numbers, but maybe their methodology is correct. So in the aggregate polls, they would be assigned less weight because it seems their numbers are off. Yet it's a fat tail poll a rare event it should be assigned a greater weighting. The problem is how do you decide what weight to assign it. Obscure fat tail events run this planet not the other way around. Ie 9/11
Four maybe the Talfgar group methodology is on but the numbers are off. They look for things that are hard to quantify and try to assume a bias in their polls. Say something out of blue you can't account for like rally effects or can't account for people not wanting people to know who they vote for so they lie. Or something like a trust issue with the people that are doing the actual polls if you believe the whole system is rigged. "Shy voters!" This was outright dismissed yet I have had that same conflicting dilemma when I was younger. I leaned right and would lie not to make people know my political leanings. Is that baked into the polls?
Back to my what, I said there are ghosts in the numbers! But like I said originally I was told I have no idea what I talking about and polling is correct science.
![troll :troll: :troll:](https://www.thecoli.com/styles/default/xenforo/smilies/troll.png)
Last edited: