Benford's Law works when your data spans multiple powers of ten (or whatever base you're using). This is not necessarily the case with voting precinct data, since most precincts are of similar size, by design.
This video describes the faults with using Benford's Law in this specific situation: https://www.youtube.com/watch?v=etx0k1nLn78
Sheila Dixon (mentioned above) somehow avoided jail after being tried for like 6 counts of theft. She was only found guilty on ONE misdemeanor charge for stealing like $500 in gift cards and buying fur coats and other necessities. Her plea deal ended up being 4 years of probation, where she was prohibited from running for office, but she remained eligible for her $83,000 pension. Immediately after her probation ended, she's come back into politics, running for office (and barely losing) twice. It boggles the mind how somebody can be caught stealing from the city, and still be competitive in politics. It's not even the "D" next to her name, since she's doing well in primaries where everybody's a Democrat. It's staggering...
In addition to Dixon, the last elected mayor (Pugh) was caught accepting bribes from the University of Maryland Medical Center (they paid $800k for her children's books in exchange for city contracts). You can see a sad summary of the scandal here, if you can look past John Oliver's face in the video preview: https://www.baltimoresun.com/maryland/baltimore-city/bs-md-john-oliver-healthy-holly-20190408-story.html
Honestly, Baltimore is hopeless. I'd love for someone like Klacik to get elected, but I just don't see how it's possible with the city's track record of recent candidates.
Zero, the votes in question hadn't been counted yet. It says that explicitly in the article.
Pretty sure Obama's twitter account was auto-following anyone that followed his a few years back. I remember seeing lots of similar posts about other bizarre "followed" users.
Still funny though...
Thanks. I swear I was looking at those numbers earlier and my brain just didn't register the change for whatever reason.
Not seeing anything fishy here. Can someone explain what I'm supposed to be seeing?
Agreed. Nothing more valuable than a friend who will be honest with you, especially when it's not what you want to hear.
Here's an easy way to see what's going on:
Consider the numbers 1-1999. Over 55% of these numbers start with a 1 (everything between 1000 and 1999, plus 1, 10, 11, 12, 101, 102, etc etc)
For other sets of numbers, you'll see a different ratio of numbers that start with 1. For instance, the set of 1-9 has 11% of the numbers starting with a 1.
This is the root of Benford's Law. If you look at ALL the possible sets of numbers, you'll find some sets with 11% of the numbers starting with 1, and you'll find other sets with as much as ~58% of the numbers starting with 1. The OVERALL average, across all possible data sets, is about 30%.
This chart would only imply what you're suggesting if an equal number of Democrats and Republicans requested absentee ballots, right?
That said, it does seem strange to me that Pennsylvania had such a massive difference compared to a similar state next door (Ohio).
Lots of misinformation about Benford's Law in this forum (probably not intentional). Benford's Law isn't some sort of silver bullet that can identify all types of fraud, nor is there a guarantee that numbers that don't conform to Benford's Law are fraudulent.
Specifically, Benford's Law is used to find "non-random" numbers where you expect to find "random" numbers. It wouldn't show any issues at all with voting software that swapped Trump votes for Biden votes (or vice versa), since both sets of votes should be random in the first place. It also wouldn't show any issues if voting software was adjusting votes on a percentage basis (counting Trump or Biden votes as 0.9 instead of 1, for example).
Finally, the law only works as a strong predictor if your data spans multiple "powers of ten", and if the data set is large enough for the percentages to normalize. I believe this is the case when looking at the results from individual voting precincts in Michigan, but you've got to be careful when using it.
Totally worth paying for a separate ride home.
Maryland's just like most other blue states. Lots of red areas (eastern shore, western md, even around Annapolis), but the state just isn't big enough for those areas to counter the super-blue population centers (Baltimore and DC suburbs).
I do think there's hope for change though. We're getting some new blood in Baltimore politics (Klacik, etc) that seems promising, and Hogan's maintaining decent approval ratings as a Republican, which is almost impossible here. I'm in a pretty mixed area politically, but definitely seeing more Trump signage compared to last election.
I like the idea, personally. Much better than leaving it blank, where anyone with a pen can go through later during the count and fill in your missing votes.
Much more concerned that the ballot is in Spanish. We're encouraging people to vote that haven't even made an effort to learn the de facto official language of the country they're living in.
I don't think it'd be nearly that much of a change. Here's an article I read a while back on the subject:
Of course, that depends on the estimated number of illegal immigrants being correct....so really, who knows?
Really thought this was going to be a picture of his pants hiked up to his nose. Only needs another 6 inches or so from his normal position.
He said 52, not 57.
Including a rather important guy born in Bethlehem...
I guess when all you have is a anti-racism hammer, everything looks like a racist nail.