Sunday, November 08, 2020

When The Math Doesn't Work

... things get stolen, things get smashed.

I'm convinced something was stolen. You can find out more about it here. I'm hoping Ohioan can add some commentary in the comments.

What gets smashed is trust in institutions. Unless you've got the East German Stasi on your side and an unarmed populace, you kind of need that trust if you're going to get anything done.

It's simply amazing to me how almost all of our institutions have managed to vaporize their good will with the public over the last, oh, I don't know, 15 years. Of course, the Catholic Church was at it a good deal earlier, but now it turns out that they were simply a template for everyone else.

I'm guessing that the police and the military have managed to hold on to a decent amount of trust and goodwill, but a whole bunch of other groups have incinerated their reputations.

Bonus Question: Just what was won on Tuesday?

13 comments:

Ohioan@Heart said...

Well that certainly looks awful. If you follow the links you can find a bunch of plots that don't look odd - for all the candidates (including a bunch of the minor ones). However, after thinking about this for awhile I realized that there is a possible reason that such an odd curve could be correct.

One of the underlying assumptions to expect a Benfold's distribution is that the numbers cover a 'reasonable' range of orders of magnitude. If all the numbers are in a small range (because, as one possibility, a candidate gets a very high percentage of all the votes in a set of precincts, which are designed to be of about the same number of total voters), then that would produce numbers with a small range of leading digits.

I ran the numbers for Cuyahoga County (Cleveland, OH) and the plot for the number of total votes is VERY non-Benfold and the Biden votes tend to fall at the high end of those, so it is also very non-Benfold, while the Trump curve is more Benfold, but not very.

Anyways, it is worth looking into. But in and of itself, is not a smoking gun for fraud, just something to look hard at.

K T Cat said...

The aggregate data has gone way past my threshold for being convinced that this was voter fraud. This is a single data point, but there are plenty of others.

Rather than list them here, I'll just add this: If you wanted skeptics to be reassured of the security of the election, you'd have used in-person voting and ID cards. Had we done that, even if we spaced the voting out over a week or so, and Biden won, I'd have been completely fine with it. Fortunes of politics and all that.

To me, the only reason we went to mail-in ballots, sprayed out like pesticide from a crop duster, was to enable fraud. The progressives think they're fighting hate and intolerance. There are no rules when you are fighting pure evil like that.

IMO, this election was stolen and there's no reason to believe that all subsequent ones won't be stolen in the same way by both parties. They unleashed the equivalent biological warfare and now both sides will have to use it. That makes discussing and analyzing politics utterly pointless.

Mostly Nothing said...

When history looks back for the start of the end of the United States, is it this, or is it the groundwork of the Obama admistration.

One Brow said...

What you really need here is not just graphs, but an actual test. Going back to the G News site, looking at Milwaukee county (from which this graph is taken), Brian Caroll's and Dan Blankenship's graphs are also off. If G News ran a chi-square analysis of the data compared to a Benford null assumption, they did not bother to include that analysis.

K T Cat said...

MN, I would argue that nations go through things like this. Closer to me, people of faith go through things like this everywhere and at all times. That's why Christian salvation is personal, not collective. You're mission in life is the same if you're a peasant farmer in Mexico or a bond trader in Chicago.

If you want to pick a spot where things started to go sideways, I'd look for the slide from truth to fantasy. We can spend without limits. Men and women are the same. There's no penalty for easily-available hard drugs on the street. Property crimes are no big deal.

That's on us, not on the politicians.

K T Cat said...

One Brow, thanks for the visit and the comment. Whether or not there's a statistical test that has some confidence level or another is meaningless to me. From the moment the mass, mail-in voting was introduced, I saw it as the ultimate mechanism for fraud on a gigantic scale. Now we have this.

For me, what I thought would happen, what made sense to happen, seems to be what happened. It's hardly worth exploring it because the doubt is now not just there, but insurmountable. It's not the outcome of the election, it's the destruction of trust.

The only reason to use untraceable ballots is to steal an election if you need it and they needed it.

One Brow said...

K T Cat,

The lest reliable determination we make is that some piece of evidence conforms to our pre-conceived expectations. If you expect fraud, you will see fraud. If you deliberately go digging for fraud, you will see it all over the place.

For another comparison point, take a look at the Chicago graph. There is no reason for such fraud in Chicago, because there was no chance that anyone besides Biden would take Illinois.

Ohioan@Heart said...

A few responses more or less in order...

KT - I agree that the aggregate reports are sufficient to pass my threshold for requiring 1) careful determination of if fraud occurred, and 2) a need for immediate tightening of the election process. In bad news on that front, now that San Diego county has (well will soon have) a Dem controlled Board of Supervisors, they are considering making all future elections include a full mailout of Mail-in Ballots. UGH. I hope they add an option for people to opt out. If they do, I will.

MN - I would posit that future historians will point to the establishment of all the SJW university majors (aka Bull $hit majors) and the subsequent population of the public schools with leftist indoctrinators, err teachers, as the two biggest steps in the down fall of the American democracy (but they'll call it something else).

One Brow (1st comment) - I actually ran the Chi-Square test. I did not bother to report it, because as I said, I considered the situation and realized that there was a reasonable explanation for why the data would not be Benfold distributed in those cases where a candidate gets a huge fraction of the votes. Hence, no reason to consider how good a fit it was. However, I would not discount 'graphs' and humans ability to 'read' them. If you want a clear discussion of how important it can be, I can heartily recommend the book, "Gravity's Kiss" by Harry Collins. It is a book written by a Sociologist about the discovery of the first gravitational wave. He spends several pages talking about how many of the scientists were completely convinced that they'd actually seen a wave based on the plots (which are at the top of the box on the right hand side of this wikipage). So even though they were convinced in just a handful of days, they still felt the need to take the months to do the statistics.

One Brow (2nd comment) - Do you really want to hold Chicago up as a place where there was no fraud? If so, I would gently suggest you look at these two links, as a start: this and this. I would also point out that "because there was no chance that anyone besides Biden would take Illinois" is at best weak support for "there was no reason for fraud". This is because it is also possible that 'the result of the election is not in doubt' because there is institutionalized fraud.

Ohioan@Heart said...

A few responses more or less in order...

KT - I agree that the aggregate reports are sufficient to pass my threshold for requiring 1) careful determination of if fraud occurred, and 2) a need for immediate tightening of the election process. In bad news on that front, now that San Diego county has (well will soon have) a Dem controlled Board of Supervisors, they are considering making all future elections include a full mailout of Mail-in Ballots. UGH. I hope they add an option for people to opt out. If they do, I will.

MN - I would posit that future historians will point to the establishment of all the SJW university majors (aka Bull $hit majors) and the subsequent population of the public schools with leftist indoctrinators, err teachers, as the two biggest steps in the down fall of the American democracy (but they'll call it something else).

One Brow (1st comment) - I actually ran the Chi-Square test. I did not bother to report it, because as I said, I considered the situation and realized that there was a reasonable explanation for why the data would not be Benfold distributed in those cases where a candidate gets a huge fraction of the votes. Hence, no reason to consider how good a fit it was. However, I would not discount 'graphs' and humans ability to 'read' them. If you want a clear discussion of how important it can be, I can heartily recommend the book, "Gravity's Kiss" by Harry Collins. It is a book written by a Sociologist about the discovery of the first gravitational wave. He spends several pages talking about how many of the scientists were completely convinced that they'd actually seen a wave based on the plots (which are at the top of the box on the right hand side of this wikipage). So even though they were convinced in just a handful of days, they still felt the need to take the months to do the statistics.

One Brow (2nd comment) - Do you really want to hold Chicago up as a place where there was no fraud? If so, I would gently suggest you look at these two links, as a start: this and this. I would also point out that "because there was no chance that anyone besides Biden would take Illinois" is at best weak support for "there was no reason for fraud". This is because it is also possible that 'the result of the election is not in doubt' because there is institutionalized fraud.

One Brow said...

Ohioan@Heart,

Richard Daley is dead, and you are looking at something that occurred 60 years ago, back when Illinois was a swing state. By the way, this is the same Daley machine that, 8 years later, put protesting progressives on trial. Do you think that machine somehow survived and reversed its conservative outlook?

Committing fraud is still a crime, and people have been convicted of it (like the three in 1962 from your article). Why would anyone take that risk for no gain?

Ohioan@Heart said...

One Brow - I hope to convince you that this discussion between us is pointless... and at the same time hope to convince you that you should not be so quick to dismiss my point of view, as I do not dismiss yours. (Note: I am not trying to convince you I am right, merely that I might be right, just as I concede you might be right.)

You and I are arguing about our different posterior probabilities. But are starting with very different priors. Bayesian logic requires that each of us must bring some sort of prior beliefs into our thought processes. My prior is that because we know that fraud has happened in the past, and since we used a voting system in which it is relatively easy to commit fraud, the prior probability for fraud occurring is reasonable (say 0.01). You have essentially dismissed the probability for fraud (thus setting your prior to something like <= 0.0001). When we then see actions that might be indications of fraud we reach different conclusions. Such suggestions cause my posterior probability to rise fairly quickly to something that forces me say it should be investigated. Yours on the other hand remains very low, and instead of thinking that it is evidence requiring investigations, you dismiss it as very probably being nothing (which then further reduces your posterior to essentially the same as your prior). This is why reasonable people can reach different conclusions, even if they are both thinking logically.

I hope you can see and understand my point of view, just as I see and understand yours. For now, I think we will have to agree to disagree.

By the way, if VP Biden is so sure he won fair and square, and if he truly believes in the rule of law, and that he believes (as he's said) that all votes need to be counted (and that that really means all legal votes), then why wouldn't he say that he also demands investigations into any "irregularities"? Having secure elections in which everyone can trust the results should be something we can all agree on.

One Brow said...

Ohioan@Heart,

I don't recall reading anything where Biden says he opposes investigating genuine irregularities. I didn't find one in a brief internet search. Do you have one? If not, why bring Biden saying something into the discussion? Most states have their own methods for investigating irregularities.

Of course, irregularities are completely different from fraud. There no reason to think irregularities are directed.

When people have actively looked for fraud (and several people have), they find it at a rate of something under 20 votes out of a billion. So, if you were to say there were 10, or even a 100, cases of voter fraud in this election, I see no reason to think this is likely. It's also not going to change any national results, and very few local results. That's why my a priori probability is so low.

With all that aside, what actions have we seen indicating fraud?

IlĂ­on said...

It's the abortion regime (and, prior to that, the "sexual revolution") that is closer to the root of all the ails us.

It's not that "God is going to judge us for [pick some latest moral/rational outrage]", it's that these outrages *are* the judgment.