Here's the forensic report from Allied Security Operations Group. Michigan politicians did everything in their power to keep this hidden, but a judge ultimately released it. This story was blackballed by MSM and censored on social media.

https://depernolaw.com/uploads/2/7/0/2/27029178/antrim_michi...

It's too difficult to decide credibility with the limited information we have and so the best thing to do is to look at opposing arguments to see which thing makes the most sense logically. In that respect, here is (I believe) the report which shows the numbers put out by the machine are not reliable, at least, and also gives evidence that 2020 log files were deleted whereas 2016 log files were not: https://www.9and10news.com/content/uploads/2020/12/Antrim_Mi...

Next, we know the state's mouthpiece will try to refute that handily, so we take a look at such an article, from Michigan itself:

https://www.detroitnews.com/story/news/politics/2020/12/14/m...

In particular, pay attention to the confounding of two separate issues to discredit the above forensic report:

> The group previously claimed there were six precincts with more than 120% turnout in the state's election. But its data was incorrect, according to publicly available turnout numbers.

"The group" did not make such claims or at least I could not find any, it was the affiant who did and he is management at the company. I could not actually find the source material on this so I am trusting the article.

Anyway, what is important is what was NOT refuted in the article.

Because, quite possibly, there is where efforts should be focused. These pieces of presumably factual evidence should be refuted by an adversarial party but are not and are just left to be memory holed.

I have not yet found anything specifically refuting the forensics report above. Simply news articles confounding things in what appears to be an effort to discredit findings. Not that I've looked too hard, mind you.

So I'm still wondering... why are 2020 logs missing while 2016 logs remain? Why do the machines give different numbers for the same votes?

These are important questions. Do not think we'll get an answer though.

And for the record, I always found Bush's Diebold election fishy.

>We reviewed the Tabulation logs in their entirety for 11/6/2020. The election logs for Antrim County consist of 15,676 total lines or events. Of the 15,676 there were a total of 10,667 critical errors/warnings or a 68.05% error rate.

>The allowable election error rate established by the Federal Election Commission guidelines is of 1 in 250,000 ballots (.0008%). We observed an error rate of 68.05%. This demonstrated a significant and fatal error in security and election integrity.

The first one sounds like they're looking at warning logs the program outputs. Those could be startup logs completely unrelated to ballots. The percent of your program's log lines that are warning logs cannot be in any way compared to the Federal Election Commission guidelines on how often a ballot can be miscounted.

Yes, when I heard that there was a report of "68%" error rate from the machines, I was a little surprised. Then when I heard it came from comedian Russell Ramsland, I thought: "Oh, I bet he did something like count all the messages in a log file and divide that by the number of ballots counted". And sure enough...

It's like the telephone scammers that try to dupe old people into giving them access to their PC by convincing them that they have a virus. Their proof? Get the trusting boomer to open the Windows Event Viewer, which on pretty much anybody's PC is replete with trivial warnings and errors. Then pretend that means something is terribly, terribly wrong.