Statistics alone is what needs to be done right now. You don't need to have the murder weapon to prove a murder occurred. Examining the results statistically allows you also allows you to put odds on what occurred. It will also allow you drill down exactly where the fuckery happened to a much smaller area and/or know that it happened past a certain point in the process between the vote being received and it appearing in the count.

All this stuff looking for mailmen or one particular corrupt person is looking for a needle in a giant haystack of fraudulent votes/counts. Sure, it happened, it almost certainly happens in every election, and it probably happened more in this one due to TDS. However, these little things don't explain the results.

My realization that this wasn't just sour grapes was in looking at the bellwether counties. I did an analysis of just those in the really close states and it defied belief. There were 35 bellwether counties on that list in Michigan and Trump won all 35 (often by large margins). That doesn't happen naturally. It is the same kind of story even in states that weren't even close.