https://www.statisticshowto.com/benfords-law/AccountantAg said:
I don't understand this, can someone do a benford law for dummies explanation?
Why would a 1 be the most common number?
Since more people need to see it, evidence of voter fraud
22,535 Views |
142 Replies |
Last: 4 yr ago by pacecar02
Valid voter ID linked to your DL or state issued ID
Bro you're linking 4chan...
https://datatodisplay.com/blog/politics/benfords-law-elections-1/AccountantAg said:
I don't understand this, can someone do a benford law for dummies explanation?
Why would a 1 be the most common number?
Look at the video in the link.
However, my question as to population #s of the areas/precincts are they order magnitudes apart.
Biden has six ending in "00"; Trump has four
AggieDub14 said:
Bro you're linking 4chan...
Because there is more data there. Holy hell the original tweet is IN THE ****ING POST! The man who did the analysis has a PhD in the field and is a professor.
You people have no leg to stand on at all.
So, precincts/ wards/areas in all those areas upload their vote totals?SLAM said:AggieDub14 said:
Bro you're linking 4chan...
Because there is more data there. Holy hell the original tweet is IN THE ****ING POST! The man who did the analysis has a PhD in the field and is a professor.
You people have no leg to stand on at all.
Or each individual polling place uploads their info?
StrikeWater00 said:
Biden has six ending in "00"; Trump has four
Yes, when people tend to make up numbers badly they tend to do it in all instances where they make up numbers badly.
Thank you for illustrating that point.
F2Aggie said:So, precincts/ wards/areas in all those areas upload their vote totals?SLAM said:AggieDub14 said:
Bro you're linking 4chan...
Because there is more data there. Holy hell the original tweet is IN THE ****ING POST! The man who did the analysis has a PhD in the field and is a professor.
You people have no leg to stand on at all.
Or each individual polling place uploads their info?
That I don't know, we would need to check into that more since it probably varies state to state. I know you can view it county by county though and at least in Texas you can see presinct data.
So, because 1% of possible outcomes end with "00" the number are made up? Is this what you're proposing?Keegan99 said:StrikeWater00 said:
Biden has six ending in "00"; Trump has four
Yes, when people tend to make up numbers badly they tend to do it in all instances where they make up numbers badly.
Thank you for illustrating that point.
No.
StrikeWater00 said:So, because 1% of possible outcomes end with "00" the number are made up? Is this what you're proposing?Keegan99 said:StrikeWater00 said:
Biden has six ending in "00"; Trump has four
Yes, when people tend to make up numbers badly they tend to do it in all instances where they make up numbers badly.
Thank you for illustrating that point.
The Benford law/theory applies to first digits, not last digits.
And sometimes second digits, yes.F2Aggie said:StrikeWater00 said:So, because 1% of possible outcomes end with "00" the number are made up? Is this what you're proposing?Keegan99 said:StrikeWater00 said:
Biden has six ending in "00"; Trump has four
Yes, when people tend to make up numbers badly they tend to do it in all instances where they make up numbers badly.
Thank you for illustrating that point.
The Benford law/theory applies to first digits, not last digits.
This was a response to a previous expert tweet on ending digits.
Here is the link where the code being run in the OP is pulling data from.F2Aggie said:So, precincts/ wards/areas in all those areas upload their vote totals?SLAM said:AggieDub14 said:
Bro you're linking 4chan...
Because there is more data there. Holy hell the original tweet is IN THE ****ING POST! The man who did the analysis has a PhD in the field and is a professor.
You people have no leg to stand on at all.
Or each individual polling place uploads their info?
https://county.milwaukee.gov/EN/County-Clerk/Off-Nav/Election-Results/Election-Results-Fall-2020
I've run it myself and I'm getting the same results.
Trump's vote count fits my hypothetical curve better RRREEEEEEEEEEEEEEEEEEEEEEEE
God I'm going to get banned for this it's too easy
Yea, statistics don't count when analyzing election data, but are totally reliable when forecasting global warming.tdb867 said:
Trump's vote count fits my hypothetical curve better RRREEEEEEEEEEEEEEEEEEEEEEEE
***It's your money, not theIRS! (At least for a little while longer.)
Well with 88 posts, it's not like we will miss you.tdb867 said:
God I'm going to get banned for this it's too easy
taxpreparer said:Yea, statistics don't count when analyzing election data, but are totally reliable when forecasting global warming.tdb867 said:
Trump's vote count fits my hypothetical curve better RRREEEEEEEEEEEEEEEEEEEEEEEE
OUCH!!!
Elections are when people find out what politicians stand for, and politicians find out what people will fall for.
BuddysBud said:taxpreparer said:Yea, statistics don't count when analyzing election data, but are totally reliable when forecasting global warming.tdb867 said:
Trump's vote count fits my hypothetical curve better RRREEEEEEEEEEEEEEEEEEEEEEEE
OUCH!!!
Support clean coal RRREEEEEEEEEEEEEEEEEEEEE
If there really are 200K votes that only selected teh POTUS, that is the easiest list to verify authenticity.
I only hear about Jacob Wohl from Democrats as an example of the Right Wing. I don't know of any conservative or Republican that listens to him or takes him seriously. He's basically another Richard Spencer that they can point to and go "See! See! Right Wingers are .....!" when they have less influence than the ex pimp guy that declared himself Dictator of CHOP.
"The most terrifying words in the English language are: I'm from the government and I'm here to help."
Ronald Reagan
Ronald Reagan
I don't find this that unusual depending on the percentage of overall ballots. I think it might be a lot of young people that are just getting in to the election process. A college kid might have been caught up in Trump mania or TDS and decided to vote for President, but have no idea about any of the other candidates and refrain from voting.beerad12man said:Hmm. That may be the single most suspicious thing. Why is someone only voting 1 thing? You may as well vote for all. Or at least the dame senator.aggiehawg said:
FWIW.
Sometimes I myself will skip down ballot votes for some local offices if the candidates are unknown to me. Of course I still vote in all state and local elections.
I'd say its certainly a valid way to vote, but there will be some statistical trend for voters like this over past electionsAmerican Hardwood said:I don't find this that unusual depending on the percentage of overall ballots. I think it might be a lot of young people that are just getting in to the election process. A college kid might have been caught up in Trump mania or TDS and decided to vote for President, but have no idea about any of the other candidates and refrain from voting.beerad12man said:Hmm. That may be the single most suspicious thing. Why is someone only voting 1 thing? You may as well vote for all. Or at least the dame senator.aggiehawg said:
FWIW.
Sometimes I myself will skip down ballot votes for some local offices if the candidates are unknown to me. Of course I still vote in all state and local elections.
a trend well outside that becomes suspicious
If you look at all the wards listed under County Clerk website that is being used as raw data, all of the wards fall between roughly 300 to 2500 registered voters (let alone ballots casted). That's less than one order of magnitude. From the breif amount I've read about this law (thanks to this thread) it seems as if that spread is not sufficient to satisy the requirements.TravelAg2004 said:
Here is the link where the code being run in the OP is pulling data from.
https://county.milwaukee.gov/EN/County-Clerk/Off-Nav/Election-Results/Election-Results-Fall-2020
I've run it myself and I'm getting the same results.
Based on my understanding of the statistic process being used, the code does a few things:
1) Calculate the benford values of a known set of numbers with an equal order of magnitude as the data set.
2) Then calculate the benford values of the data set.
3) Most importantly, calculate how closely the data set matches the "known" set.
It's the "fit" that should be important.
When looking at the results, here is what you get (the closer to zero, the better it fits the "known" distribution):
Biden: 1.5076774999383611e-27
Trump: 0.00048111250713426005
PS - I'm learning a lot about this as well. This is just my understanding based on basic research and reading through the actual code and running it.
1) Calculate the benford values of a known set of numbers with an equal order of magnitude as the data set.
2) Then calculate the benford values of the data set.
3) Most importantly, calculate how closely the data set matches the "known" set.
It's the "fit" that should be important.
When looking at the results, here is what you get (the closer to zero, the better it fits the "known" distribution):
Biden: 1.5076774999383611e-27
Trump: 0.00048111250713426005
PS - I'm learning a lot about this as well. This is just my understanding based on basic research and reading through the actual code and running it.
Keep them coming. Although we know that the amount of evidence will never change these people's minds. there may be hundreds of individual examples by the time this is over.
Bump, since more people need to see this info.
Quote:
AccountantAg said:
I don't understand this, can someone do a benford law for dummies explanation?
Why would a 1 be the most common number?
I use Benford analysis in my job. It is kinda like F=M*A in that it is much harder to explain "why" it works than to just observe that is "does" work. However, I'll try to give an example that helps people somewhat understand it.
Imagine that you are a company and your expenses are comprised primarily of purchase orders. Your purchase orders are typically anywhere from a few hundred dollars up to several thousand dollars depending on what/how much is ordered. Why would the first digit be one more often than any other? Well for every dollar amount that is less than $1,000 ($1-$999) there is the same amount of dollar amounts that start with a "1" ($1,000-$1,999). If you include the "1's" below 1,000 (1, 10-19. 100-199) you get the the number 1 occurs 55.5% of the time if your number set is from 1 to 2,000. As you expand the range, the percentage of "1's" goes down, but doesn't become "equal" until you get to all 9's (9,999). Since data sets typically aren't so defined (i.e values can be "anything") the number "1" as a first digit has an advantage.
This doesn't hold up though if the data set has some sort of non-random bind to it. For example, if you ran this on a company's sales that only sold one product for $499 and sold to retail clients so most people only bought one, the number 4 would probably be the leading number, and it wouldn't be a sign of fraud.
So, while the Benford analysis is quite interesting, it would take some one with a lot (and I mean a lot) of statistical knowledge and governmental knowledge about Wisconsin to determine that wards/precincts are not set up in some way that would lead data sets to generate non-random first integers. Also, verify that the data isn't entered in a way that would cause this (i.e) voter program limits amount of ballots to 500 or 5,000 at once or something.
TLDR: Benford's law is confusing to anyone who doesn't use it/study it extensively (and still confusing to those who do). It doesn't really "prove" fraud, just that there is some reason (potentially fraudulent or not) that is causing a data set to not follow it.
In discussing this with my dad, (trial attorney) he told me he uses benford's law quite often in the cases he tries, so the courts should be familiar with this information...
Reading between the lines: this might be the irregularities the Trump team has been referencing.
Reading between the lines: this might be the irregularities the Trump team has been referencing.
Featured Stories
See All
LIVE: No. 13 Texas A&M at South Carolina
by Luke Evangelist
68:03
8h ago
3.9k
Presenting Texas Aggies United's 'Florida Slam Dunk' Sweepstakes!
by Texas Aggies United
Keys to the Game: No. 13 Texas A&M at South Carolina
by Luke Evangelist