This article by Andrea Widburg contains a link to a video by Shiva Ayyadurai.
The relation that Ayyadurai presents is spurious. Instead of graphing a “y” variable against an “x” variable, he graphs a (y-x) variable against an “x” variable. This forces a downward-sloping line because as x gets bigger, (y-x) gets made smaller necessarily unless y is growing enough to counteract it. In this case, this growth in y isn’t present in the data. Ayyadurai’s whole “analysis” has no basis in theory in order to understand why he transformed the y variable in this manner.
To show algebraically what’s going on would be too detailed for this blog. The bottom line is that I find his work in that video to be worse than useless. It’s actually propagating a spurious relation and suggesting that an algorithm produced it. However, he has no corroborating evidence at all of the presence of such an algorithm.
His video and further articles that accept his unsupported claim can only damage those cases and claims of fraud or election influence that may be or are truly legitimate.
7:05 am on November 14, 2020
