Of course I want Trump to win but I think this guy didn’t do things right.
When he separated the two groups:
Individual candidate voter
Straight party voter
and then the difference in percentage between the groups, he didn’t normalize the percentage from the total. If I can find the raw data, I can do it myself. Anyone knows where I can download the raw data.
I haven’t recreated the analysis from the voting data, but I wasn’t convinced that their claim that the distribution patterns they calculated couldn’t be due to something other than algorithmic cheating.
So I did some simple calculations in a spreadsheet that assumes that a uniform percentage of non-straight-line voting was done in order to crossover on the vote for president. The result is, for any level of R crossovers voting for Biden, the graph will show the linear fall of in vote advantage for Trump.
So, there is an argument for another cause for their result. That doesn’t mean there is no fraud, but it may mean that their analysis doesn’t prove there was fraud.
If anyone wants a copy of my simple spreadsheet, drop me a freep mail with an email address. I would appreciate someone double checking it.