On Monday, March 14, 2016 at 8:01:16 PM UTC-4, duncan smith wrote:probability of 'Republican' is 55%... or something like that.
"An impossible event with probability zero - you think that's absurd? The point is that even if you have mutually exclusive events, if you plug the correct probabilities into Bayes theorem you get the correct answer."
Cardinal confusion: Event of probability 0 (impossibility) equated to ABSURDITY.
Look at the absurdity of applying (attempt) Bayes' Theory to a mutually exclusive event (e.g. coin toss) -
P('Heads' | 'Tails') = P('Heads') * P('Tails') / P('Heads') = (1/2 * 1/2) / 1/2 = 1/2 = 0
The absurdity: 1/2 = 0 (a number greater than 0 is equal to 0). You can't simply say: "Make it 0 by definition". That's insane!
That's why the Bayes Theorem does NOT -- repeat, does NOT -- apply to mutually exclusive events. "If 'Woman' is true, then the probability of 'Man' is 1/2 = 50%"!!! Bayes' Theorem correctly applies as in, for example: "If 'Man' is true, then the
"10% of the cast votes in the US election is a very large sample."say on TV) will not be made even if 60% of the votes were counted. In another extreme case, Florida 2000, wrong 'projections' were made after 90% of the votes were counted!
Only in your crazy mind! The Pennsylvania case in 2012 was extreme. Philadelphia was the key, with a large number of Black voters who decided to vote in very large numbers. They voted Democrat 95%. In most cases, the predictions ('projections', as they
Ion Saliu,
Founder of: Probability Theory of Life, Randomness Politics http://saliu.com/dirty-election.html
Sysop: | Keyop |
---|---|
Location: | Huddersfield, West Yorkshire, UK |
Users: | 307 |
Nodes: | 16 (3 / 13) |
Uptime: | 32:09:17 |
Calls: | 6,908 |
Calls today: | 2 |
Files: | 12,376 |
Messages: | 5,428,085 |
Posted today: | 1 |