this is a hard problem in computer science (probably the most important problem in CS and all of math)

but it is very approachable to amateurs and it is surprising that it is difficult to solve

many people take a crack at it, and it is a trope in the CS community; people look down on those who attempt to solve this sort of problem, especially if they are amateurs (not professionals in CS)

@DrSammyD It just might be human error. When they manually enter values into the NEP system, they sometimes attribute them to the wrong precincts. In order to mask the errors, they use a debit/credit approach to move the values. Logically, you might think they would send a negative entry on one side and the positive on the other. In this case, they are doing a wholesale replacement of the values.

the reason being that it is very difficult to read through the proofs and disprove them (takes a lot of time) and they never end

sort of like Perpetual Motion machines for physicists

anyway, there is a page on the internet that collects such "proofs" for archival purposes

and in order to learn the underlying difficulty of the problem, I spent a lot of time humoring such proofs and taking the time to disprove them

fun days

That would make sense if it wasn't for like sets of 5 different precincts at a time, all with varying total values but not ratios.

anyway, I've never seen one that involves stacraft

he's laughing at his own simulation!

I still don't understand this scheme

OK so it knows how many votes it needs in the "virtual precincts"

and it hijacks some precincts during a period of time

ohh I get it now

those precincts converge to the virtual ratio

Yep

and hold there until it snaps to another precinct

OK, what was missing is the "converge" part

ok now I can think of how to visualize this

what if

we visualized a 2d chart showing the ratios of every precinct

over time

each precinct gets a curve

yea I think I can work with this

Yeah, I think that's a good visualization.

this worked well! @DrSammyD

Conspiracy: Confirmed!

gave up for tonight lol

https://www.msn.com/en-us/news/politics/gsa-chief-says-she-faced-threats-to-family-staff-and-pets-for-holding-up-biden-transition/ar-BB1bivbn

... Is that all of Philidelphia or the United States?

First thing is to filter out all precincts which never share a ratio

next is to filter out all precincts which only share low denominator ratios

Those naturally occur regardless

I noticed on the Edison Research site that they were promoting their superior accuracy, so they might want to avoid highlighting mistakes. If there were manual updates being corrected, because they were mistakenly attributed to the wrong precinct, they might want to effect the changes in micro batches so as to not bring attention to it. So, they would grab a small set of votes and transfer each set at regular intervals until the error was corrected. A lazy programmer might use the same ratio to accomplish this. It would be interesting to see if these same strange patterns occur for a state that has been recounted. Maybe GA and the counties in WI.

Yeah. Agreed. I'd like to also compare it to a non swing state city.

philly

there is prolly something wrong with my data

but i gave up for the night 🙂