Message from @realz
Discord ID: 776177509834293268
I think the conclusion here is:
1. if the variance in magnitude is low, Benford's Law doesn't say anything should be expected.
2. Though perhaps one could ask question about why the variance in magnitude is low, and how does this compare to previous elections and so on; I imagine simply looking at the political history of the districts involved etc. would answer such questions.
3. But if the variance in magnitude is high, _and_ there are enough samples, then if the shape of the graph does not follow Benford, then there is something to analyze further.
4. Even in the best case, % deviation from the expected Benford graph is not as useful as it seems, unless you factor in the number of samples (the more samples, the less expected deviation).
All that being said, I bet Mark Nigrini knows more than me
I don't think he'd disagree though
IIRC he did use a lot of samples
2 hours đź‘€
YES thank you for this.
@realz I think you've put this Benford's law issue to rest ... I work with numbers every day and have for many decades, generally those statistical methods are good for evaluating measurements of physical processes (temperature, wind velocity, etc.) but when you look at numbers associated with human behavior then you'll find many time statistical methods don't readily apply, just take the polling predictions versus the election results both in 2016 and 2020
mhm
that is yet another issue
The reason that the first digit appears in ratios is because we have a system of numbers based in 10
So 1 should appear more often than other numbers. It would be different than if it was evaluating the last digit that appears.
I imagine that benford's law actually applies in every base
which is interesting
as for the last number!
watch Parker's (Stand-up Maths) video
the last number also has a law
he demonstrates that the last 2 numbers are uniformly distributed in chicago
there are so few trump voters in some, that with just 1,2 or 3 digits, the last 2 digits are not uniformly distributed
The last 2 numbers I would expect to be uniform in distribution
yep he runs that test
and they are, except for this blip I just explained
` there are so few trump voters in some, that with just 1,2 or 3 digits, the last 2 digits are not uniformly distributed`
Yeah but I am sharing that Benfords law isn’t about what’s being measured. Though what’s being measured could bias what you would expect to see.
you are referring back to the base
> https://youtu.be/CMMbZH-H4ks
@Michele411 the problem I have with this sort of thing is that there is no evidence ... it sounds like a conspiracy theory and I cannot differentiate it from something from a crazy person
(not saying he is a crazy person)
I think Benford's law may be invalidated based on what is being measured ... consider for example a particular voting precinct, the voters in that precinct are relatively fixed, most of them consistently vote one way or another and very few change from election to election
I agree Meglide that the base may effect the law
so the variation in the number of possible combination of vote outcomes is limited
I might just add a base switcher; my guess is that the law will hold in all bases and fail in all bases equally
also different regions or precincts will be strongly influenced towards one party or another
(except for small sigma)
As for whether or not it is accurate here I do not know. I think that something probably happened, but it was a conspiracy of convenience not malice. Many people probably thought it was their duty to ensure Trump not win. I, erring on the side of process, would disagree that it’s right to intervene but understand the argument. Like there are cases where I would put the outcome above the process.
@realz Agreed. He is making statements that I feel is somewhat unbecoming for a higher officer.
Montgomery is not a....well...not an A1-source
its like listening to coast-to-coast