Before I press on with the next post showing the results from the other states, I feel the need to clarify something.
I received some great feedback for my first post which demonstrates fraudulent audit transactions in Pennsylvania, Virginia, and Minnesota among other state anomalies - and I very much appreciate that.
I also ran into some of the Twitter Troll Farm when I replied to a tweet thread by Erick Erickson an erstwhile blue check Trump supporter who has a dim view that election fraud is real (otherwise Erick seems a pretty solid conservative).
To my reply Erick said:
That'd you'd rely on Edison to make your claim makes your claims unreliable as Edison is not affiliated with states and makes errors in every election on top of the flood of absentee ballots that come in late. But you do you. Hope you make some good cash off it.
Italic emphasis mine.
I addressed Erick’s concerns in line with his thread, but, out of the barrage of ad hominem drivel that proceeded from the farm (all from denizens that clearly didn’t read my post - I know it’s really difficult to, like, actually read something), one idea was raised that I think is related to Erick’s answer and I want to address it as well. That idea is that to make the Edison data synch up with the Secretary of State of each state’s tabulations (which it does since it’s actually produced by the state), that Edison has to “correct” its “mistakes”. That’s how you must explain those pesky “audit transactions”.
“Rely on Edison”
It’s interesting that this should be derogatory, because every network and virtually every news outlet (except for AP since they still apparently run a competing service) and therefore the public, rely on Edison provided data to see tabulation results and forecast winners, etc.
From the Edison Research web site:
A proven source for accurate and timely
nationwide vote tabulation
In an effort to improve quality, streamline data collection, and expand election coverage in 2018, ABC News, CBS News, CNN and NBC News ended their arrangement with the Associated Press for vote tabulation and now partner with Edison Research for these data.
For the 2020 General Election, Edison Research provided the NEP with a fast and accurate vote count throughout the nation, providing data for all statewide races and all House races. Race projections were made in all 50 states for statewide races, ballot measures, as well as all 435 House races.
And further:
Vote Count
The NEP provides media organizations with real-time vote results in all 50 states for statewide races, ballot initiatives and House races. The NEP vote count from Edison Research is the only service continuously updated after Election Day. We track every vote until results are certified by all states. Vote count data is available for statewide results, vote by Congressional District, county vote data breakouts for statewide races, and all U.S. House races.
The fastest and most complete vote results
Multiple sources allow Edison Research to provide faster reporting of vote data than any other organization.
Clearly, quite a few entities rely on Edison. And every single called contest in this election was called by Edison based on the data they transmitted (other news sources may have made calls at different times, but they did so based on Edison data as well.)
“Edison is not affiliated with states”
Edison must have a contractual relationship with each state to provide tabulation data. Sure, they are not part of the states themselves - they are a commercial company, but they perform tabulations for each state and they have access to and provide the same tabulation data to their clients that the states use to inform their own election apparatus.
It seems there’s quite a bit of purposeful disinformation about Edison. They are basically a data provider. Within the data stream that they produce there is poll exit data, news items, some commentary, election race calls, etc. which are proactive data they create and provide to their subscribers, but there’s also time-stamped election tabulation data for the presidential election embedded in their stream that comes from the states. In their own words: “real-time vote results in all 50 states”. You don’t get that by standing outside of the precincts trying to scratch out totals on a chalkboard called to you by a colleague inside the precinct. It’s quite clear from the data streams themselves that this is a fairly automated process.
“Makes errors in every election”
So there are transcription errors in Edison’s data? If so, the states feature those as well, as it’s the same reported data. They simply transfer and propagate the data that they receive from the states. I’ll get into this in more detail shortly, but fundamentally Edison is a data provider - they don’t correct or edit their data at some post date - they just transmit what they receive - and I will prove this shortly. It would actually be far more concerning if the individual tabulation reports were altered over time - for obvious reasons.
“…Top of the flood of absentee ballots that come in late”
This seems to mean that there’s some part of the vote tabulation stream that Edison can’t keep up with. It’s just simply not true. In my observations, Edison updates tabulation data far faster than the Secretary of State sites for the states. For most states, we can readily see from the Edison data if they counted their mail-in ballots early or late. Most importantly, the Edison tabulations match up to the Secretary of State published data (though Edison is earlier with their published data than the States - the Secretary of State sites eventually catch up to Edison.) Note that there are still some states that are counting so the numbers even now are not everywhere static.
Sufficient proof for what I’m saying here can be demonstrated by snapshots of the NYT site files of interest (I’m going to use Pennsylvania over a short time.)
Let’s take what I think is the first release of the Pennsylvania Edison NYT .JSON file on November 4th:
You can see that it just has a few records (or rows in Excel parlance). These are the first few vote tabulations reported by Pennsylvania.
Now, let’s go forward to the main file I’ve been working with for Pennsylvania from November 10th. (Yes, it’s pretty large by that time.)
So, if you compare the first few rows of this group of tabulations with the first one - the time stamps are the same, the vote totals are the same, the percent tabulated, is the same, the decimal division between the candidates is the same - the “intersection” of the first spreadsheet and this one is exactly the same. (Well, I tweaked the name a little bit to reference the file date from the first one to this one - and will for the next as well.)
Now, sorry but an even larger one. The file from November 22nd. They were almost done counting.
And guess what, the intersection between the previous file and this file are identical. Same numbers everywhere (well, if you want to get picky I didn’t format the total votes number so missing commas per thousands) - including the massive audit transactions that support some of the basis of my first post. Incidentally, anyone can reproduce these results by using the references I put in my first post.
And if you don’t have “receipts” yourself - you can get these files from the Wayback Machine. For Pennsylvania there are about 100 snapshots just of that state (the internet is forever you know.)
The fact is that Edison provides the exact data as produced by tabulations from the states. It just “appends” records / rows as it goes along and doesn’t change any data that has a previous timestamp. This is the case for all states and all publications of these .JSON files.
And if we compare the ‘final’ tally in this last file (they continued to count I believe) with the Pennsylvania Secretary of State total, they are within 2,099 votes, which makes the count between them accurate within 99.97%. This means that the massive audit transactions are still in the tabulation stream and that if Pennsylvania certifies this vote that they are certifying these audit transactions that occur at 16% percent of the tabulation.
Now, it’s likely that these numbers will continue to change a bit as the state site and the final tabulations occur. Eventually they should be the exact number. If they are within 0.1%, then we are at the mathematical limit for accuracy as I documented earlier.
Aside: As I was listening to the Michigan hearings today, I heard retired Colonel Phil Waldron - the former military intelligence officer that has been testifying in the state hearings - talking about how the data proceeds from the machines for tabulation and he confirmed (which I didn’t know for sure before) that the machine data itself is in the form that Edison streams - meaning it provides the vote total and the fractional amounts of the total to the two major party candidates to three decimal places for each tabulation. That makes it more disturbing to me as it makes voters ‘fractional’ and really does offer the cover of the 0.1% accuracy resolution for nefarious doings. He also confirmed that tabulations for given precincts occur with some frequency during given sessions - they don’t just wait until all the votes are counted for a particular precinct or precincts. That makes the big spikes in these states even more suspicious. As I’ve been saying, you can’t tell the difference between a vote tabulation and a positive audit transaction. Those transactions should be preserved somewhere and they should be examined.
Conclusion
Edison, like I’ve contended, is the primary election data vendor in the United States that produces the tabulation data for the votes - for all contests in the United States - in a record by record fashion, and for that tabulation data only reports it. It doesn’t modify that tabulation stream, and it reports it verbatim from the tabulation machines.
Anyone that tells you different is brimming with excrement.