Voter Fraud - The absence of evidence is not evidence of absence.

I “think” Biden won but it sure is opaque.

After hearing claims ranging from “there was massive vote fraud” to “there’s no evidence of voter fraud” I decided to take a shallow look.

Interestingly, or maybe sadly,  there is a lack of publicly, easily accessible data available to really look at this question.

One source of data, being used to claim clear evidence of voter fraud,  is from somebody that analyzed “scraped” data from the NYT website as they updated their results.  However, even though there are interesting questions raised by this data, I do not think the data is sufficient to make any conclusions. 

In the attached charts there are several things visible to note from reported Michigan data:

1. It does appear that there is some signal from mail in voting where, because of the randomness involved in mail handling and delivery, that allows you to actually observe it to some extent.  This is the “straightish” line below.  In fact it is the jumps in this line that some claim is a clear signal of voter fraud.  However, because the data does not include information at a more granular level than state, it is not possible to know if the shifts are some effect of different parts of a state counting mail in votes during different and separate times during the count. 
2. There are some unique data points that appear anomalous that have both a significant number of votes and a very significant skew towards Democrats.  For example one dump was of 149,772 votes with 94.32% for Biden. Once again, without more detail on the source of these dumps within the state, it is not possible to conclude much about these points. 
3. Since the data reported on NYT is actually data they purchase from another company that provides details of election results for a price, there is also the possibility that the data reported on the NYT is not perfectly reflected of actual state data being reported. 

CONCLUSIONS

1. It should be possible to look at actual state voting counts by the discrete vote counting regions in a state over time to identify anomalies, as it appears that there is enough randomness in mail in votes due to handling, delivery, etc. that mail in votes within a discrete voting area should have a constant signal.
2. There are potentially statistically unique counts reported that warrant examination, assuming NYT data is representative of actual count reported.
3. For those claiming there is no evidence of voter fraud.  That might actually be true. However, because there is no publicly available data to examine, it’s also possible that there is no evidence because the data isn’t available and those who like the current purported outcome, have no interest/desire to look.
4. Just because you’re paranoid doesn’t mean that they are not out to get you (Louis Lerner, Steele Dossier,etc.)