Explanation and Summary of Dr. Frank’s graphs seen at the symposium:

• The title is the name of the county being examined. X Axis is Voter Age. Y Axis is Number of People.

• Number of people that can vote (18+) is the blue line (also called population).

• Number of people that are registered to vote is the black line (also called registrations).

• Number of ballots counted in 2020 election is the red line (also called ballots).

• Predicted number of ballots counted in 2020 election is the light blue line (not shown in all his graphs, also called predicted ballots).

1. He has claimed that he was able to pick a random county in a state, use the ratios between (population to ballots) or (registrations to ballots) for each individual from age 18 to 100+ to generate a 6th order polynomial that is extremely precise and accurate in predicting all the ratios for all other counties in that state. The registrations to ballots prediction is even more precise than the population to ballots prediction, so that will be the focus.

2. The argument he is making is that the 6th order polynomial prediction for (registrations to ballots) ratio should not be doing such a fantastic and nearly perfect job of predicting all the other counties (registration to ballots) ratio, and so there is (most likely) a nefarious algorithm in place, using the number of registrations as a base point to transform the true number of ballots into a calculated and inflated set, skewed towards (most likely) Biden.

3. I screenshotted his graphs linked below in the gofile of the most 5 populated counties in Ohio. The prediction of ballots cast based upon the registrations was within 99.9% for 3 out of the 5 counties, and 99.8% and 99.7% for the other two counties. GOOD REDPILLS

Devil’s Advocate #1: 99.7%+ correlation of (registration to ballots) ratio between multiple large counties is not an anomaly at all. This is because we are dealing with large numbers. You would expect that when comparing counties, the ratio of registration to ballots cast among a certain age group would converge towards the same ratio the larger the sample size. A few million people is a big enough sample to make a correlation of 99.7-99.9% not that extreme.

There are roughly 1,000,000 people in Franklin County. Each individual age from 18-70 has roughly 10,000 people registered. This means that for each age, this magic polynomial is predicting the EXACT number of votes per age of voter, with an average error of just 10 VOTES per age. And not just for this county, but the prediction works for each of the 5 largest counties, and all other counties, predicting the exact number of votes, for each age, plus or minus the number of fingers on your hand. This seems to be a strange anomaly.

Devil’s Advocate #2: Why wouldn’t it just be an exact match of 100% if the same transformation algorithm based on number of registrations was used in each county? In fact, it should be 100% if the number of registrations was the only variable used to make the transformation to the number of ballots.

Right, but I don’t think the number of registrations was or could even possibly be the only variable used in the algorithm. It would be far too obvious to catch if it could be just a simple transformation of the registrations. There are very complex real world variables that the algorithm would have to contend with as the election was going on. It would have to be constantly re-balancing each county and precincts numbers as they come in, adjusting for the speed of the ballots being counted in each location, and using weights to alter the constantly changing lead or deficit of a candidate. I imagine if an algorithm was used, it would have to be very rigorous to be able to contend with an extremely dynamic and fast paced situation, altering the votes as they are coming in real time.

I think the algorithm didn’t just use the number of registrations as the sole base variable to do the transformations, but rather used the number of registrations as a key static variable to constantly be striving to converge towards, without needing or wanting to hit the ratio exactly, essentially giving itself an upper and lower bound of random error that is allowed. It will likely not hit the mark exactly due to rounding and the dynamic and messy nature of the live vote counting.

https://gofile.io/d/YLlh41