Putting this up now so it's goes up before the election.
Enjoyed the response, but while my prior discussion may have been eloquent from what you were saying, I apparently was far from clear as you both got a little of my comment, but missed most of what I was trying to articulate. Not surprising, I have this problem regularly.
My comment was not so much as to "normalize" Trump's behavior with regards to losing as it was to say it was a natural progression. At the same time, that does indeed argue that it is somewhat "normal" for him to behave that way (although I think massively problematic). Also, while I think it has become a litmus test to assert he lost which creates its own problems and is actually a small symptom of a much bigger problem, that wasn't quite accurate either.
My comment is about the ability to manipulate statistics and the way people think about them. I'll provide you with really good internet connection that is up 99.9% of the time, you think you'll always have it, you don't think that it means it'll potentially be down for one entire work day each year (0.1% of a year is a little less than 9 hours which is about one work day).
Similarly, if I told you that 7 out of 10 times "A" happens its caused by "bad" and the other 3 times it is caused by "just happens" every time it happens you would say it is "more likely" to have been caused by "bad", to the point that if it happens 100 times, you will never say any one time was more likely caused by "just happens" even though it almost certainly was. As John knows, I'm sure, this can create an interesting liability determination problem in law.
This statistical imperfection is also true in elections and vote counting. No electoral (state) win has ever been decided by 1 vote, by 10 votes, or even by 100 votes. In many bigger elections 1,000 votes or 10,000 votes may not matter. Heck, even 100,000 or 1,000,000 (e.g. California in 2016 was decided by a more than 4,000,000 vote difference) may not even matter. It's how we can (safely) declare winners before the counting is complete and why candidates concede.
Thus, if you happen to get the count wrong by a few 1000, who cares, it doesn't matter anyway. If California miscounted in 2016 and gave 1,000,000 votes to Clinton that should have gone to Trump, who cares? It had no effect (heck, they could have 100% fraud, it still wouldn't matter). Further, We all know you're going to get it somewhat wrong every time by a small amount. No one actually believes the vote count in any state is perfectly accurate to a single vote. So what if it's off by 10, 100, 1000 every year. It's still right 99.9% of the time. But when is it not a small amount? Can it regularly be off by 10,000. How about 100,000?
And then it did matter. Florida in 2000 which was decided by (likely) around 500 votes. Suddenly, an error of a few thousand was potentially immense and what we all learned in Florida in 2000 is that errors of a few thousand may be quite common (why would all these people vote for Pat Buchannon, hanging chads, etc. etc.). We basically realized that 99.9% accuracy across 50 states means that one state should get it wrong every 20 elections and we got really concerned about the accuracy of a process that has never been perfectly accurate.
This is what I would assert Trump is the logical conclusion of. Trump argues he is the 0.1% "error" . It works, because it's impossible to prove the negative (particularly hard with what happened in 2000). You can't prove Trump lost (black swan problem) and he won't concede (which is what any person who actually believes in the system should do to avoid the question). Instead, you can only prove it more and more likely that he lost. Thus, what does it mean to say he "lost".
Putting this up now so it's goes up before the election.
Enjoyed the response, but while my prior discussion may have been eloquent from what you were saying, I apparently was far from clear as you both got a little of my comment, but missed most of what I was trying to articulate. Not surprising, I have this problem regularly.
My comment was not so much as to "normalize" Trump's behavior with regards to losing as it was to say it was a natural progression. At the same time, that does indeed argue that it is somewhat "normal" for him to behave that way (although I think massively problematic). Also, while I think it has become a litmus test to assert he lost which creates its own problems and is actually a small symptom of a much bigger problem, that wasn't quite accurate either.
My comment is about the ability to manipulate statistics and the way people think about them. I'll provide you with really good internet connection that is up 99.9% of the time, you think you'll always have it, you don't think that it means it'll potentially be down for one entire work day each year (0.1% of a year is a little less than 9 hours which is about one work day).
Similarly, if I told you that 7 out of 10 times "A" happens its caused by "bad" and the other 3 times it is caused by "just happens" every time it happens you would say it is "more likely" to have been caused by "bad", to the point that if it happens 100 times, you will never say any one time was more likely caused by "just happens" even though it almost certainly was. As John knows, I'm sure, this can create an interesting liability determination problem in law.
This statistical imperfection is also true in elections and vote counting. No electoral (state) win has ever been decided by 1 vote, by 10 votes, or even by 100 votes. In many bigger elections 1,000 votes or 10,000 votes may not matter. Heck, even 100,000 or 1,000,000 (e.g. California in 2016 was decided by a more than 4,000,000 vote difference) may not even matter. It's how we can (safely) declare winners before the counting is complete and why candidates concede.
Thus, if you happen to get the count wrong by a few 1000, who cares, it doesn't matter anyway. If California miscounted in 2016 and gave 1,000,000 votes to Clinton that should have gone to Trump, who cares? It had no effect (heck, they could have 100% fraud, it still wouldn't matter). Further, We all know you're going to get it somewhat wrong every time by a small amount. No one actually believes the vote count in any state is perfectly accurate to a single vote. So what if it's off by 10, 100, 1000 every year. It's still right 99.9% of the time. But when is it not a small amount? Can it regularly be off by 10,000. How about 100,000?
And then it did matter. Florida in 2000 which was decided by (likely) around 500 votes. Suddenly, an error of a few thousand was potentially immense and what we all learned in Florida in 2000 is that errors of a few thousand may be quite common (why would all these people vote for Pat Buchannon, hanging chads, etc. etc.). We basically realized that 99.9% accuracy across 50 states means that one state should get it wrong every 20 elections and we got really concerned about the accuracy of a process that has never been perfectly accurate.
This is what I would assert Trump is the logical conclusion of. Trump argues he is the 0.1% "error" . It works, because it's impossible to prove the negative (particularly hard with what happened in 2000). You can't prove Trump lost (black swan problem) and he won't concede (which is what any person who actually believes in the system should do to avoid the question). Instead, you can only prove it more and more likely that he lost. Thus, what does it mean to say he "lost".