He looked at every change going only one direction! He didn't also look at changes made from Biden to Trump. It was an accident I am sure, but this can't keep spreading. It harms out credibility.
Or what if they got so sloppy that they literally only make changes in one direction? I said to my friend earlier today that I'm sure there are republicans who get propped up with it to have controlled opposition but in this election they did not account for the fact of how much they'd have to cheat.
The script did find the original 'glitch' that flipped the votes in that one MI county, which means that even though the source data json file is imprecise to 3 sigfigs the data is precise enough that the original 'glitch' could be uncovered.
The lost data in the timeseries is a problem but the ratios are still calculated from a more precise source, implying that if a 'glitch' occurs in the data the script he wrote should find it. The Biden to Trump scenario shouldn't be an issue because it should show up in the data as a negative number, presuming of course that the logical tests don't require positive magnitudes in the code.
So in conclusion I assert that there are errors in the data because of the imprecision but those errors shouldn't be systemic if they are teased out with a supplemental test reversing the order of the test columns, replacing Biden variables with Trump variables and vice versa then comparing the two resultant ratios.
He wrote the script right, and my point here is that the point of contention associated with the imprecision might very well the point of this aggregator's function in the first place. What if the timeseries in the Edison data was designed to be imprecise using these stupid ratios instead of the raw integers themselves?
Behind those extra trailing digits lies, when the totals are large enough, thousands of votes that are unaccounted for between each time interval, allowing anyone to skim the fractional percentages off of one candidate and apply it to the other during each step before the aggregator reports the imprecision to the timeseries in the json file.
If you had network communication with Edison on one side, and any unsecured terminal with physical access at to any one of these unsecured voting stations around the country, which we know were on independent networks themselves, anyone perverse enough could hijack the raw data feeds on both ends and modify the Dominion machines and the Edison feed simultaneously to push fractional votes from in the way outlined above.
Afterwards you would have to have people on the ground who know about your operation that would need to duplicate the paper ballots associated with each modification such that they read the new values instead of the old ones.
In conclusion I think buried in the errors lies the bare truth of actual hijacked votes which could be found with very little modification to the code.
He looked at every change going only one direction! He didn't also look at changes made from Biden to Trump. It was an accident I am sure, but this can't keep spreading. It harms out credibility.
Or what if they got so sloppy that they literally only make changes in one direction? I said to my friend earlier today that I'm sure there are republicans who get propped up with it to have controlled opposition but in this election they did not account for the fact of how much they'd have to cheat.
The script did find the original 'glitch' that flipped the votes in that one MI county, which means that even though the source data json file is imprecise to 3 sigfigs the data is precise enough that the original 'glitch' could be uncovered.
The lost data in the timeseries is a problem but the ratios are still calculated from a more precise source, implying that if a 'glitch' occurs in the data the script he wrote should find it. The Biden to Trump scenario shouldn't be an issue because it should show up in the data as a negative number, presuming of course that the logical tests don't require positive magnitudes in the code.
So in conclusion I assert that there are errors in the data because of the imprecision but those errors shouldn't be systemic if they are teased out with a supplemental test reversing the order of the test columns, replacing Biden variables with Trump variables and vice versa then comparing the two resultant ratios.
He wrote the script right, and my point here is that the point of contention associated with the imprecision might very well the point of this aggregator's function in the first place. What if the timeseries in the Edison data was designed to be imprecise using these stupid ratios instead of the raw integers themselves?
Behind those extra trailing digits lies, when the totals are large enough, thousands of votes that are unaccounted for between each time interval, allowing anyone to skim the fractional percentages off of one candidate and apply it to the other during each step before the aggregator reports the imprecision to the timeseries in the json file.
If you had network communication with Edison on one side, and any unsecured terminal with physical access at to any one of these unsecured voting stations around the country, which we know were on independent networks themselves, anyone perverse enough could hijack the raw data feeds on both ends and modify the Dominion machines and the Edison feed simultaneously to push fractional votes from in the way outlined above.
Afterwards you would have to have people on the ground who know about your operation that would need to duplicate the paper ballots associated with each modification such that they read the new values instead of the old ones.
In conclusion I think buried in the errors lies the bare truth of actual hijacked votes which could be found with very little modification to the code.