Posted on 08/12/2020 6:58:20 AM PDT by nwrep
If your data/statistics are $h!t, which they are in this case, then your predictions will be $h!t too.
Easily, provably wrong.
Those are random results from 40,000 simulations fed into the statistical model. Thank you.
He’s just factoring in the rampant fraud the Dims are allowed.
The dude actually brags about how far off he was with his FINAL prediction in 2016. Then expects us to take his latest load of manure seriously.
In 2016, I was getting polled AT LEAST once a week,often more.
This year: ZERO NADA NOTHING CRICKETS
What a joke! Trump is going to win in a landslide.
Biden taking FL and LA?........Not gonna happen...............
GIGO.....................
If I read the meat of his article correctly, this is what he is really saying:
https://www.270towin.com/maps/AZbgl
I give this buffoon a 1% chance of being right. He is favorable to the left always. He was touted over 2012 because Romney stunk. He gave Hillary 99% in 2016 and fell flat on his face.
An easily, provably and demonstrably wrong statement. Thank you.
I don't know if its possible to have statistical prediction in the strictest sense- statistics is used for events that have happened, explaining the results - Nate is making a prediction - where one has expected results, for that we use probability.
Statistics tries to explain what happened from the data, Probability tries to predict the future results (expected values) from future trials (voting) from the current data.
I would think statistical analysis would lead to making forecasts based on that data and then applying probablity to that data ... maybe .. I could be full of hot air but one time I was confused about the difference between the two and looked it up this is what I remember.
Nate Silver did a pretty good job forecasting actually in 2016 based on the available polls. It was Nate Cohn at NYT and the idiots at HuffPost, etc that all had 90-99% chance of HRC Winning. Silver actually had trump with a 35% chance a week out and a 29% on election day.
There are also simulations showing Trump winning Oregon and Nevada and Minnesota.
Silver wasn’t wrong. He isn’t actually predicting who will win; he’s modeling who is more likely to win based on currently available polling data. Based on that data in 2016, Clinton was more likely to win.
By analogy, if you flip a coin twice, I can confidently predict that you most likely will not flip heads both times. I can even put a number to my level of confidence, namely 75%. There is only a 1 in 4 chance that if you actually do flip a coin twice, you’ll get heads twice. Now if you do the experiment and actually do get two heads, does it mean I was wrong? No, obviously not.
Additionally I base my coin prediction on the assumption that the coin is equally likely to land on either side. If that’s not true, my prediction is altered. Similarly Silver’s probability is based on the assumption of unbiased and reliable poll data. If it’s not, then his probability will be off as well.
The only change from the actual 2016 is Dim cheating gets them MI, PA, and maybe WI back. It is a near mirror of his Nov 8, 2016 prediction except he gives Trump a better chance in Florida and North Carolina. His GA and LA stuff is in the fluff of the article, but not in the statistical portion.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.