In this particular case, there's a mathematical reason for this.
The "votes being lost due to rounding" is affecting third party candidates more than the main candidates. At first glance, that would sound like cheating.
But think of it this way. Let's say a rounding error costs a candidate 500 votes. Now what happens when the vote updates. Maybe the third party candidates get around 100 votes in a typical update, while Trump and Biden get around 1500.
How often is a 500 vote rounding error going to make Trump or Biden votes decrease when they average about 1500 votes per update? Not very often, the rounding errors will still be there but the votes won't go down. They'll just go up 500 points more or less than they should.
But now for third party candidates, let's say they're getting around 100 votes per update. A 500 vote rounding error is huge and could much most easily make their number decrease in the update.
So while there is plenty of blatant fraud this election, unless we have more specific numbers or evidence to go on I think this particular example is just a case of people not realizing these numbers are rounded and how that affects the math. We should focus on better examples of fraud.
Edit: Sigh, come on guys, think this through. I know not everyone is a math person but we have plenty of real evidence to work on. We don't want to waste time on things that are simply wrong.
Let's say Trump has 100,000 votes. You add 700 more, but a rounding error costs 500 votes. It still goes up to 100,200 despite the error.
Okay. Now let's say Jo Jorgensen has 4000 votes. You add 50 more votes, but a rounding error costs 500 votes. Number goes down to 3550 because of the error.
That's what's happening here. It's faulty analysis because the numbers are approximate.
This doesn't explain why the votes kept returning to the 2% number including that one big swing where they removed around 30% of 3rd party votes to bring the share of votes for 3rd party from 3% back down to 2%.
That's a fair point, there is something screwy going on when around 13,500 votes are added and then removed again. Those kinds of errors should absolutely be looked at.
And the lack of precision in the data could be used to make fraud harder to detect.
I'd that were true, the changes would be random and not systematically in one direction
In this particular case, there's a mathematical reason for this.
The "votes being lost due to rounding" is affecting third party candidates more than the main candidates. At first glance, that would sound like cheating.
But think of it this way. Let's say a rounding error costs a candidate 500 votes. Now what happens when the vote updates. Maybe the third party candidates get around 100 votes in a typical update, while Trump and Biden get around 1500.
How often is a 500 vote rounding error going to make Trump or Biden votes decrease when they average about 1500 votes per update? Not very often, the rounding errors will still be there but the votes won't go down. They'll just go up 500 points more or less than they should.
But now for third party candidates, let's say they're getting around 100 votes per update. A 500 vote rounding error is huge and could much most easily make their number decrease in the update.
So while there is plenty of blatant fraud this election, unless we have more specific numbers or evidence to go on I think this particular example is just a case of people not realizing these numbers are rounded and how that affects the math. We should focus on better examples of fraud.
Edit: Sigh, come on guys, think this through. I know not everyone is a math person but we have plenty of real evidence to work on. We don't want to waste time on things that are simply wrong.
Let's say Trump has 100,000 votes. You add 700 more, but a rounding error costs 500 votes. It still goes up to 100,200 despite the error.
Okay. Now let's say Jo Jorgensen has 4000 votes. You add 50 more votes, but a rounding error costs 500 votes. Number goes down to 3550 because of the error.
That's what's happening here. It's faulty analysis because the numbers are approximate.
This doesn't explain why the votes kept returning to the 2% number including that one big swing where they removed around 30% of 3rd party votes to bring the share of votes for 3rd party from 3% back down to 2%.
That's a fair point, there is something screwy going on when around 13,500 votes are added and then removed again. Those kinds of errors should absolutely be looked at.
And the lack of precision in the data could be used to make fraud harder to detect.
Just don't want us focusing on the wrong things.