What was the model’s error ? Was extra Biden vote numbers outside the error?
I have no idea.
The witness stated that the actual votes were above and below the model's predictions except in counties (NATIONWIDE) where Dominion systems were used. In these Biden got FIVE PERCENT above the model's predictions.
I don't know, but I do have Math degrees with a heavy emphasis on statistics. I assume that at least 100 counties are involved. If all the other counties are 50-50 (above-below) and the Dominion counties are say 98-2 (above-below) regardless of the percentage (and FIVE PERCENT seems high to me) that it is beyond a reasonable doubt (way beyond) that there was something wrong with the Dominion machines, just on the basis of simple binomial (coin-flip) distribution statistics. (I.e., if you flip a coin and 98 times out of 100 it comes up heads, it isn't an honest coin.)
ML/NJ