What the original video was saying is that the negative vote count for Trump is likely due to a rounding error, since Edison data only reports the total number of votes and then the % for Trump and % for Biden to THREE digits. Ie. 53.4% would show up as 0.534 in the edison data.
Three digits is not enough to calculate precise vote counts for both candidates since it leaves a wide margin of error due to rounding artifacts. 53.4% of the votes could really be anything between 53.35 to 53.44999999999999 and still be rounded to 53.4%. If you have a batch of 10,000 ballots, that would mean between 5335 and 5345 votes. If you apply this uncertainty to both ends (before and after a batch is counted) it compounds the uncertainty and expands the range of possible votes that will give the same change.
So this youtuber took those numbers and found the minimum and maximum votes for Biden that would still round up or down to match the Edison data.
And in this batch the MINIMUM number of votes that Biden could have gotten to round to those values was 23,012 out of 23,487, or 97.9779% of that batch's votes. So even with maximum allowance for rounding errors, Biden had to win 98% of that vote batch, in a state where the county with the HIGHEST democrat absentee ballot percentage was in the mid-80's. It's a VERY Statistically improbable thing to have happen, but it does NOT NECESSARILY mean that Trump lost any votes (even though it's statistically unlikely that he did lose votes given the range of possible values).
For completeness though, the total range of Trump Vote share from this batch could have been between -14% and +0.82% (which is 193 out of 23,487 votes)... This is more likely to be evidence of one of those batches of "pristine uncreased" ballots that went 100% for Biden than it is for Trump to have lost votes.
Negative votes seems to be suspect.
Nice link, and total sus
https://www.thegatewaypundit.com/2020/11/flashback-lou-dobbs-cnn-2006-smartmatic-based-venezuela-sold-sequoia-top-officials-still-venezuela-sale-dominion/
This channel guy does a TERRIBLE job of summarizing the spreadsheet from the other youtube video that he's looking at.
Watch the original video and not this guy.
https://www.youtube.com/watch?v=mp0u4dw7GH4
What the original video was saying is that the negative vote count for Trump is likely due to a rounding error, since Edison data only reports the total number of votes and then the % for Trump and % for Biden to THREE digits. Ie. 53.4% would show up as 0.534 in the edison data.
Three digits is not enough to calculate precise vote counts for both candidates since it leaves a wide margin of error due to rounding artifacts. 53.4% of the votes could really be anything between 53.35 to 53.44999999999999 and still be rounded to 53.4%. If you have a batch of 10,000 ballots, that would mean between 5335 and 5345 votes. If you apply this uncertainty to both ends (before and after a batch is counted) it compounds the uncertainty and expands the range of possible votes that will give the same change.
So this youtuber took those numbers and found the minimum and maximum votes for Biden that would still round up or down to match the Edison data.
And in this batch the MINIMUM number of votes that Biden could have gotten to round to those values was 23,012 out of 23,487, or 97.9779% of that batch's votes. So even with maximum allowance for rounding errors, Biden had to win 98% of that vote batch, in a state where the county with the HIGHEST democrat absentee ballot percentage was in the mid-80's. It's a VERY Statistically improbable thing to have happen, but it does NOT NECESSARILY mean that Trump lost any votes (even though it's statistically unlikely that he did lose votes given the range of possible values).
For completeness though, the total range of Trump Vote share from this batch could have been between -14% and +0.82% (which is 193 out of 23,487 votes)... This is more likely to be evidence of one of those batches of "pristine uncreased" ballots that went 100% for Biden than it is for Trump to have lost votes.
Oh this? This is fine.