It would take a while to go through the whole thing, and I'm trying to learn some shit I'll wind up actually use. But to look through that first argument, he basically is looking at the differences in vote values between observations for the Biden team (ie, how much it increased each time). He compiles a set of ALLLLLLLLLL the various vote increases, and then adds together random samples of all of these various increments together 10,000 different times by adding together... his 'frequency' breakdown for the number of vote distributions times ninety.
Basically, he does 10,000 runs of taking 90 different values and adding them together from Figure 2, which is measuring the number of increments of the total which fell into the ranges (0-10k, 10k-20k, etc). This completely disregards the individual vote total increments that happened - it instead rounds them, presumably up (IE a vote increase in an interval that -was- 500 is now weighted at 10k). The frequency with which he selects each of the values is based on their frequency in the total, so the 10k one is theoretically likely to be picked 60-some percent of each 'pick' in the set of 90.
Based on those 10,000 runs, he maps out the distribution of sum totals - IE, very few of the runs resulted in just 400k increases for Joe, and very many of them resulted in somewhere slightly below 900k increases. He uses this to conclude that the chance that Joe would land within 1% of Trump's total would be very, very unlikely.
Now, this is a retarded take for many reasons. First off, Joe ultimately beats Trump over this period by somewhere above 574,724 votes. I say 'somewhere above' because this paper doesn't say how much Trump's count increased over the course of it. Adding 541k (his original deficit) and 34k (his finishing margin) doesn't account for trump's own increase.
His own model suggests that Biden should have gotten somewhere under 900k additional votes from his initial deficit in Navarro's data, ~300k more than Biden actually did.
He basically tries to conclude that by not doing as well as his model says that he should have, this is very suspicious.
Of course, as I said earlier, he would round up an interval of 100 extra votes to 10k extra votes, so that under-900k value is grossly inflated and as such the 'center' of his data is grossly inflated.
His conclusion is also flawed. It's basically "given the delta of these intervals, rounded up, and completely ignoring the fact of time (IE, late-coming batches of ballots causing very small increments and increments even of 0 getting rounded up to 10k further inflating it), it is very unlikely that he would ONLY get this much, because it is close to Trump."
But if he fucking LOST to Trump, by this model, it would be EVEN MORE UNLIKELY.
Any value that Biden landed at would in reality be very unlikely because IT'S AN ELECTION. This guy tries to fixate on the fact that he got so close to Trump, which surely must be abnormal! But then, what's normal...?
By this nimrod's metric, Biden's win would have been more "normal" if PA's voter turnout had been even higher. Assuming he'd need that extra 300k to be "normal," that means he'd need a 2% increased turnout of ALL PENNSYLVANIANS (not just VEP) in an election where PA already broke its record for voter turnout.
It looks like the final tally was within about 100K: https://www.electionreturns.pa.gov/.
It looks like the rounding could potentially be an issue, although admittedly, I don't perfectly understand the methodology or how much the rounding could potentially affect the analysis.
In terms of the conclusion, isn't the idea that according to the data on hand it would be expected that Biden should be much more likely to win by a larger margin, and that he didn't could be a sign of irregularities, such as additional fraudulent ballots stopping once sufficient votes are acquired, or am I misunderstanding something?
His model says that if Biden won some kind of state-wide sweep, it should have been higher. But it was only higher in the "problematic" counties (Philadelphia and Alleghany). He shows they're anomalous pretty clearly: all the other counties followed historical trends, but those counties had surges out of line with trends from previous years and this year.
His conclusion is fraud, of course. The alternative argument is that there's something super special about Philadelphia and Alleghany counties who just love them some Sleepy Joe. Considering how they ignored court orders and fucked with observers, they don't get the benefit of the doubt.
You're leaving out a big point here. From the report:
As above, the problem isn't that there was some big surge of mail-in votes for Biden. That was expected. The problem is that the surge in those specific counties is way out of step with the overall surge everyone else saw.
It's not enough to say "this was an unprecedented election" and just excuse any weird looking returns. The law of large numbers is a real thing. When the major anomalies occur in limited locations, all in strategic states, all tilting one way, that requires a better explanation than "people are unpredictable". They are not that unpredictable (a fact the study covers as well).
Also, I don't know what that Forbes article is supposed to be, it doesn't address any of the points in the study. It seems to generalize some arguments without reading them, then point to "debunking" articles about completely different accusations. Maybe it was referencing something in his press conference, there isn't a shred of math in there.
Item nine is in the second part. My point pertains purely and entirely to the first part, which is the one that discusses variance on the vote on the whole; the first "time series analysis" is completely self-contained enough to draw its own conclusion paragraph. Generally speaking, if I can't trust your entire initial argument because your methodology clearly cooks the books by counting a delta of 50 as 10k, yeah.
To look at what numbers he's fucking with in detail for claims outside of the first part, it would take a while longer. More broadly, I would say it can easily be overlooked by the fact that people who live in cities are going to vote for the democrat more often both in-person and by-mail, and cities took longer to process the mail-in ballots than did rural areas. I'll look at this one real quick, too.
Regarding his 'summary' assertions at the top:
1)philly and allegheny are fucking cities, so calling them deviant is stupid.
2) We did not have mail-in voting for anyone that wanted it in 2008-2016. Gee, wonder why that might be unusual in places with the highest % of mail-ins.
3) Turns out that younger, cosmopolitan people don't tend to register for political parties. 'Progressives' for example are often just independent, even if our primary is closed.
Item 1 choosing to display numbers over % is pretty suspect, given that Philly has about 12 times the population of Allentown. A 1% increase for Biden in Philly is bigger compared to a 10% increase in Allentown. Elk County has about 30k people. A 1% up in Philly is 50% of Elk County. See why this is him clearly fucking with the presentation of the data?
Item 2's comparing the behavior of Elk County to Bucks county in terms of enthusiasm is retarded. You'd have to be completely clueless about the Philly suburbs and surrounding counties, which is part of why trying to apply statistical analysis to the rural-urban divide is prime "wow this has a lot of numbers must be true" grifting.
Item 4 is again based on the raw number change, not the % change. This is stupid. Elk County getting 6k votes is a fifth of its population; Bucks getting 6k is just barely under 1%.
Item 5 is again basing the difference on TOTAL NUMBERS, NOT PERCENT. "It is curious that Mont and Allegheny would have large declines but also large increases" no, not at all, given your graph is averaging together -the populations of every single county in PA-.
Item 6 compares rural democratic registration to urban democratic registration by explicitly removing the urban registration (all 12 counties that are big) from its "average." This is clear manipulation, again. Elk County is not Allegheny in demography or registration habits.
Item 7 again removes the urban centers which would show you precisely why this analysis is fucking stupid. Independents and republicans voted for Biden in the urban-suburban counties. That "70%" number he throws around also counts independents and republicans in those other counties, by the way; he just hopes you won't notice. He is comparing " ALL biden votes" to "number of registered democrats."
Item 8 assumes, again, that voting behaviors between cities and rural areas would be the same. I really don't think I need to explain to you that Philly and Pennsyltucky are different. Also, fucking lol, he calculates the "normal" behavior of the biden votes:registered democrats by EXCLUDING the 11 counties, and then he performs this linear regression with them INCLUDED. Oh, except Philly and Allegheny; we still left those ones out when we did this. nigga lol
Item 9 compares ELK COUNTY TO ALLEGHENY COUNTY (but not to philly). Why has Philly been left out again and again? Oh, because it saw a -decrease- that was very minor in terms of % of votes, but very massive in terms of overall numbers as stated above, which would throw these equations into the bin. And it's -still- based on "votes obtained per registered voter," which was still calculated in item 6 by EXPLICITLY EXCLUDING philly and the 11 "outlier counties" which would reflect an urban trend.
It looks like the final tally was within about 100K: https://www.electionreturns.pa.gov/.
It looks like the rounding could potentially be an issue, although admittedly, I don't perfectly understand the methodology or how much the rounding could potentially affect the analysis.
In terms of the conclusion, isn't the idea that according to the data on hand it would be expected that Biden should be much more likely to win by a larger margin, and that he didn't could be a sign of irregularities, such as additional fraudulent ballots stopping once sufficient votes are acquired, or am I misunderstanding something?
Heh... I'm sure any statistician could give you an answer for that one.
Basically, he has 90 'deltas' which are the changes over this period of time from one measurement to the next. He is only concerned with behavior within the start and end data point, so I guess what I should moreso be saying is that he alleges these numbers regarding what the behavior should have looked like at his end point, the 7th; by his mark, Biden should have been up by ~300k more votes by that end-mark.
However, each of those 90 deltas is put into a bin of 10k. There's 10k, 20k, 30k, 40k, etc. A delta of 100 goes into the 10k bucket; a vote of 10,001 goes into the 20k bucket. There is no rounding down.
Each of these 10k-buckets has a frequency based on the number of the original 90 deltas that falls within them - so, for example, if 60 of the deltas were <=10k, then the frequency of the 0-10k mark would be 66.6%.
With this, he does 10,000 different trials of taking 90 values based on those frequencies. Each of those values is one of the 10k ranges and added to the 'total' for that trial. Normal distribution becomes the 'expected' result where the random polling most closely matches the frequency distribution, which would be the center of his graph at the end that compares all of the different totals from all of the different trials in terms of how many times those totals emerged from the trial.
But, again, because a delta of 1 single vote (or even 0) is counted instead as TEN-THOUSAND VOTES by his frequency thing, his numbers are horrendously inflated. Notwithstanding that this is not a great system by which to judge an election, given that the basic assertion of this "statistical analysis" is "votes should come in timewise in roughly the same amounts," ie that the mail-in ballots should share some similarities to the earlier in-person votes of the day.
So yes, the conclusions suggests he "should" have more, but only if you ignore the inflation and assume that the votes he receives throughout this period "should" be uniform despite the fact that we're excluding the deltas from -all of the other time periods in which votes were tallied-.
Furthermore, you could 'normalize' the 10k increments by instead counting each one as its midpoint (5k, 15k, etc)... but this again assumes regularity that these increments would have no reason to obey; you would be underestimating them if 75% were over 5k, but still overestimating if 75% were below it. If he did that, he would call it out (and probably just use them as separate buckets / frequencies), though.
Retard Louie Gohmert had his lolsuit rejected on appeal by the Fifth Circuit. It took less than a day. On a Saturday. This is the entire ruling, with the court adopting the position of Mike Pence and Trump's DOJ that the lolsuit is without merit
Next stop, getting dunked on by the Supreme Court!
EDIT: This crazy Paddy Byrne tweet fills me with joy. So many good things coming together. It is only missing the Sith Lord
As an Aussie I find it extremely pathetic that Americans simply bends over and accepts the official election results when literally every available piece of data points towards massive voter fraud all across the country. Now, I don't know what the average American thinks of this but you must realize that EVERY. SINGLE. newspaper and newsoutlet here, save for the extreme leftist SJW ones, have long ago looked at numbers which literally screams massive voting fraud and realized that something is completely off here. You are being shagged by a walking corpse and bloody refuse to do a damn thing about it.
Funny.
You Americans have tons of guns and for years sported that one day you will use them to remove dictators and tyrants from your nation but when you actually are faced with just that you are too pussy whipped move even a finger.
No sympathy from here. Biden is what you deserve and Biden is what you get.
I must have imagined all those Leftists tweeting till they're sore comparison pics between Trump's inauguration attendance and Obamas's. Hell, the BBC even did it.
Retard Louie Gohmert had his lolsuit rejected on appeal by the Fifth Circuit. It took less than a day. On a Saturday. This is the entire ruling, with the court adopting the position of Mike Pence and Trump's DOJ that the lolsuit is without merit View attachment 1825266
So yet again, a court rejects a suit based on standing. Seems to be very difficult to get a court to actually hear you on this. How many of the suits brought have actually been heard?
It looks like the final tally was within about 100K: https://www.electionreturns.pa.gov/.
It looks like the rounding could potentially be an issue, although admittedly, I don't perfectly understand the methodology or how much the rounding could potentially affect the analysis.
In terms of the conclusion, isn't the idea that according to the data on hand it would be expected that Biden should be much more likely to win by a larger margin, and that he didn't could be a sign of irregularities, such as additional fraudulent ballots stopping once sufficient votes are acquired, or am I misunderstanding something?
Heh... I'm sure any statistician could give you an answer for that one.
No, because in most of America, the large cities are who vote Democrat while the more rural areas and smaller cities typically vote more Republican. Philly is a gigantic city that votes Democrat.
You guys really underestimate the difference in political culture between cities and rural areas. California is a great example. Check out LA during the election season, then drive outside LA. It's like two different countries.
Trump's own election commission and AG said there was no fraud. He lost.
So yet again, a court rejects a suit based on standing. Seems to be very difficult to get a court to actually hear you on this. How many of the suits brought have actually been heard?
Maybe it'd be worth a shot to look up how this "standing" thing actually works before going to court and getting your shit dismissed for being incorrect on a fundamental juristic level?
No, because in most of America, the large cities are who vote Democrat while the more rural areas and smaller cities typically vote more Republican. Philly is a gigantic city that votes Democrat.
You guys really underestimate the difference in political culture between cities and rural areas. California is a great example. Check out LA during the election season, then drive outside LA. It's like two different countries.
Trump's own election commission and AG said there was no fraud. He lost.
I think it's not so much about them underestimating the difference in political culture and more of doublethink. They're happy to point out that Trump won last time for going to rural parts of the US and taking advantage of his strong support there, yet now they are completely dumbfounded that he's mainly supported in rural areas and can't fathom that the cities don't really love Trump. Nah. The cities MUST love him, any election result showing otherwise must be fake news.
I think it's not so much about them underestimating the difference in political culture and more of doublethink. They're happy to point out that Trump won last time for going to rural parts of the US and taking advantage of his strong support there, yet now they are completely dumbfounded that he's mainly supported in rural areas and can't fathom that the cities don't really love them. Nah. The cities MUST love him, any election result showing otherwise must be fake news.
Exactly. And a lot of those city people who didn't vote in 2016 because Hillary sucked realized that Biden may not be their ideal candidate, but they gotta vote out Trump.
I must have imagined all those Leftists tweeting till they're sore comparison pics between Trump's inauguration attendance and Obamas's. Hell, the BBC even did it.
So yet again, a court rejects a suit based on standing. Seems to be very difficult to get a court to actually hear you on this. How many of the suits brought have actually been heard?
Thank you for making my point. No one gave a shit about inaugurations until Trump made it out to be such a huge thing. The fact that people are already planning to use Biden’s inauguration turnout, during a pandemic, as some kind of proof of fraud is retarded and frankly embarrassing.
The Democrats will win. Either the GOP base will be too demoralized to show up, in which case the Democrats win, or the Republicans do show up, but there will be a magical 4AM dropoff of 95% Democrat ballots in which case, the Dems still win. No use to discuss it further.
The Democrats will win. Either the GOP base will be too demoralized to show up, in which case the Democrats win, or the Republicans do show up, but there will be a magical 4AM dropoff of 95% Democrat ballots in which case, the Dems still win. No use to discuss it further.
The Democrats will win. Either the GOP base will be too demoralized to show up, in which case the Democrats win, or the Republicans do show up, but there will be a magical 4AM dropoff of 95% Democrat ballots in which case, the Dems still win. No use to discuss it further.
