4829
Comments (128)
sorted by:
You're viewing a single comment thread. View all comments, or full comment thread.
16
LehighLuke 16 points ago +19 / -3

This, I just watched stand up math's video on this subject. Benfords law requires a dataset that spans multiple orders of magnitude...election votes in precincts are the opposite of that, and are roughly the same size. Benfords is a red herring ...there is an ocean of fraud evidence elsewhere...this is not it!

23
Belleoffreedom 23 points ago +23 / -0

Bedfords Law is an indicator that an anomaly exists, not proof of fraud.

A totally separate issue is vote switching, which is when you can point to flips in the reporting of a specific number of votes at a specific time, in a specific direction. Vote switches are actions that affect the count.

5
Walbort 5 points ago +5 / -0

Second on the Stand Up Maths' video on Youtube. He basically explains how there isn't enough of the magic sauce in that data that makes Benford's Law work, and the anomalies you're seeing are actually just the raw data showing through. He also looks at the data a different way, shows an example that looks like an anomaly in Trump's data, and again shows that it's just highlighting the raw data.

4
spezisapedo 4 points ago +4 / -0

many precincts do span orders of magnitude, and were fine previously for proof. But I do agree it is not conclusive as is.

2
KHandle 2 points ago +2 / -0

Benfords law is almost always obeyed in elections. It is really simple math. It's only in this election that people are gaslighting about it. When it is disobeyed in one county or state, and if it coincides with some other fact like a new electronic voting system is used, that is a big red flag. Ignoring it is like ignoring the sound of someone breaking into your house at night.

In Nebraska a recount of the district in question would probably solve the issue. That is not very hard to request. And I think people will be open to a hand recount.

The real danger is the machine will keep flipping votes in future elections, so this problem won't go away if you ignore it.

0
KHandle 0 points ago +1 / -1

Regarding debates about data spanning several orders of magnitude. Can you also express the numbers in base-5. You then have enough digits for statistics (the law is a little different in base 5). And even 15000 votes is 1000000 in base 5, so your data easily spans 6 orders of magnitude.

1
spezisapedo 1 point ago +1 / -0

Yeah, it is supposed to work in other bases.

However your math is a bit misleading, just because you convert the digits to base-5 you would still have the same number of votes, since each digit representation still only counts as 1 vote and we express numbers for statistical validity in base 10.

Of course you could convert EVERYTHING to base-5 including the statistical validity, but that is unnecessary. Just convert the digits to base-5, but you still have the same votes. Regardless, 15000 votes do not necessarily span many orders of magnitude, since ever precinct could have the same number of votes, say 1000 for example. But its possible that each precinct saw very different numbers of votes, in which case you would observe 10 in one, and maybe 1000 in another and maybe 5000 in another.