RSS

Category Archives: Uncategorized

Election Fraud: Response to Joshua Holland

Election Fraud:  Response to Joshua Holland

Richard Charnin
April 29,2016

Last week,  actor and activist Tim Robbins tweeted on the exit poll discrepancies . And the media presstitutes went after him with a vengeance.

Joshua Holland wrote a response in The Nation. He followed  with another piece in Raw Story: On Tim Robbins, election fraud and how nonsense spreads around the Internet

Monday, actor Tim Robbins caused a stir when he tweeted out a Facebook meme, charging that CNN and The New York Times are blind to a massive conspiracy going on right beneath their noses. It had close to 1,000 retweets when Robbins apparently deleted it.

A quick glance is enough to know that there are problems with the meme. The exit poll numbers are wrong. In Massachusetts, for example, CNN reported that exit polls showed Clinton winning by 2 points, which is very close to her 1.4 percent margin in the final results. In Alabama, CNN reports the exit polls showing Clinton with a 57-point margin, the Facebook meme claims it was 44.7 points, and the final result was 60.4 percent.

But where did Tim Robbins come up with these numbers? I decided to do a bit of reporting, and I ended up chasing this Facebook meme down a rabbit-hole of misinformation and conspiracism. It offers a pretty good case-study of how bullshit can come to dominate our online discourse.

The meme was created by Lee Camp, a political comedian who hosts a weekly show on RT, the Russian foreign news network. It has over 2,000 shares on Facebook as of this writing. Via email, I asked Camp for his source, and he pointed me to a post on Reddit by a user who goes by the handle “turn-trout.” Turn-trout, who didn’t respond to a message seeking comment, claims that these are unadjusted exit polls, and links to a spreadsheet purportedly showing wide discrepancies between the raw data and the final results.

The spreadsheet was created by Richard Charnin, who writes a blog devoted to “JFK conspiracy and systemic election fraud analysis.” Charnin’s spreadsheet appears to be the basis of a broad swath of viral Internet content alleging widespread election theft during the 2016 primaries, including the work of Free Press editors Harvey Wasserman and Bob Fitrakis Charnin seems to think that exit polls can reveal that virtually all our elections have been rigged, writing, “in the 1988-2008 presidential elections, the Democrats won the exit polls by 52-42%; they won the recorded vote by just 48-46%, an 8% discrepancy.”

I exchanged some… interesting correspondence with Charnin. After calling me “very biased and misleading” for a recent piece, he claimed that “ALL exit polls are forced to match the recorded vote.” I asked him whether the exit poll data in his spreadsheet were unadjusted, and he said that they were the data released by major media organizations. He then told me that “the mainstream media won’t dare touch the Third Rail – ELECTION FRAUD,” but it’s cool because “Tim Robbins just talked about it.” Finally, the truth emerges.

Virtually all of these claims are based on the idea that exit polls are a telltale sign of fraud. In a follow-up tweet, Tim Robbins explained that, “exit polls are historically pretty accurate,” and “are a heads-up on vote tampering.” Turn-trout agrees, writing, “Exit polls have historically and throughout the world been used as a check against, and rough indicator of, the degree of election fraud.”

This is also the basis of claims by Wasserman and Fitrakis – who point to the precision of German exit polling to emphasize the point – and Steven Freeman, a Penn State psychologist who authored the book, Was the 2004 Presidential Election Stolen?: Exit Polls, Election Fraud, and the Official Count.

(Note:  Steve Freeman is not a psychologist. He holds a PhD from MIT’s Sloan School of Management where he teaches research methods and survey design).

So there you have it. They say a lie can travel halfway around the world before the truth can get its boots on, and that’s especially true of the internet. Here we have an example of an actor citing a comedian who picked up a claim from an an anonymous Reddit user citing preliminary exit poll data put together by a JFK conspiracy theorist. Bringing it all full circle is The Hill, which ran a story titled, “Actor Tim Robbins blames Sanders losses on ‘voter fraud,’” which will no doubt be shared thousands of times on Facebook and Twitter.

Last week, I attempted to debunk allegations of widespread election fraud by the Clinton campaign that have been swirling around on social media. My argument was an appeal to common sense: If Hillary Clinton entered the race with a very large lead in the national polls and an enormous amount of support from Democratic Party activists and elected officials, as she did, and then quickly built up a significant lead in pledged delegates, as she did, then at no time since the start of the race, regardless of how unscrupulous her campaign might be, would there be any rational motive for risking infamy by rigging the vote. You don’t need to cheat when you’re winning.

My response:

Holland called it a “rabbit hole of misinformation and conspiracism”?  When someone uses the conspiracy meme,  I have concluded based on 15 years of posting that the author will proceed with misinformation and factual omission.

For the record, I have two masters degrees in applied mathematics and have followed presidential elections since 1952.

Holland said that the exit poll results are “wrong”.  Really? What was the basis for that statement?  He has obviously not viewed the data downloaded from CNN. Holland called the numbers “bullshit”. That says it all. The articles are biased and misleading and  will not fool those who are aware of the facts  and  the math that proves election fraud.

Edison Research

Holland interviewed  Joe Lenski of Edison Research, the polling firm . But he failed to get Lenski to explain why exit pollsters force a match to the recorded vote – even if the recorded vote is bogus.  Lenski  does not logically explain why ALL exit polls are matched to the recorded vote counts.  He has never provided a rationale for the match. Of course Lenski works for the Corporate Media (the National Election Pool). Holland fails to see the significance of that. He never considers that the recorded vote may be fraudulent. In fact, he never mentions the F-word – nor does anyone else in the corporate media. Holland fails to see the significance of that. 

Holland wrote that  Lenski stressed that “pre-election polls are also adjusted to conform their samples to what pollsters know about the populations they’re trying to measure. The irony of all of this is that the adjusted data are far more accurate than the raw data”.

 Are we supposed to take that comment seriously? Apparently Holland does.  If that is the case why won’t the pollsters show us the raw exit poll data in all the precincts polled? And how did they end up with an exact 0.01% match to the recorded vote in the CT primary? Holland fails to see the significance of that. 

1988-2008 Presidential Exit Polls

In spreadsheets linked from my blog, I provide 1988-2008 historical presidential vote and unadjusted exit poll data.  The Democrats led the exit polls by 52-42% but led the recorded vote by just 48-46%. The probability of the 8% discrepancy is one in trillions. Holland does not have a clue about Probability and Statistics 101. He does not  appreciate or comprehend the magnitude of the discrepancy.  The data is available if he would care to truly investigate. Holland fails to see the significance of that. 

The data shows that 135 of 274 state presidential exit polls from 1988-2008 exceeded the margin of error – and 131 moved in a “redshift” to the GOP. The probability is ZERO: E-116. That’s 116 zeros after the decimal. Holland does not  appreciate or comprehend the magnitude of the discrepancy. He fails to see the significance of that fact.

The primaries

Sanders has won 12 of 13 caucuses, but only 6 of 27 primaries. Election fraud anomalies were apparent in NY MA IL AZ IA NV OH DE WY WI MO DE MI AL TN GA AR TX. Holland fails to see the significance of that.

Sanders led hand-counted precincts in Massachusetts   by 17%, while machine counted precincts went to Clinton by 2%. Holland fails to see the significance of that.

The CNN NY Primary exit poll indicated that  Sanders had 48% (officially he had 42.1%). But he may have done better than 48% since ALL exit polls are adjusted throughout the day  to the recorded vote. And hundreds of thousands of voters were disenfranchised  when their registrations were flipped or dropped altogether. Holland fails to see the significance of that.

There was an 11.8% discrepancy between the NY exit poll  (52-48%) and the recorded vote (57.9-42.1%). The probability of the discrepancy is 1 in 120,000. Holland fails to see the significance of that.

Sanders exit poll share declined from the poll to the vote in 21 of 23 exit polls.
The probability P = 1 in 30,000 = binomdist(2,23,0.5,true).  Holland fails to see the significance of that.

Sanders exit poll share exceeded his recorded share by more than the margin of error in 9 of the 23 primaries. The probability of this result being due to chance:
P= 1 in 90 million = poisson( 9,0.025*23,false). Holland fails to see the significance of that.

As usual, in the recent CT, MD and PA primaries, the exit poll was forced to match the recorded vote. The differences between Clinton’s 2-party adjusted exit poll share and the recorded share were: CT .01%;  MD 0.10%;  PA -.17%. Holland fails to see the significance of that.

The mainstream media (including The Nation) won’t dare touch the Third Rail – ELECTION FRAUD. They never discuss malicious, proprietary voting machines and central tabulators that were built to flip the votes. Holland fails to see the significance of that.

Will Holland look at the evidence which proves that the primaries are being stolen from Sanders and write about it? Perhaps he will learn something and see the significance of these posts:  
https://richardcharnin.wordpress.com/…/1988-2008-unadjuste…/  https://richardcharnin.wordpress.com/2012/11/17/a-reply-to-nate-silvers-ten-reasons-why-you-should-ignore-exit-polls/

Note: I challenge Holland to try to refute  the data and analysis in these books:
Matrix of Deceit: Forcing Pre-Election and Exit Polls to Match Fraudulent Vote Counts
Proving Election Fraud: Phantom Voters, Uncounted Votes and the National Exit Poll
Reclaiming Science: The JFK Conspiracy

Fitrakis and Wasserman

Holland criticized  Bob Fitrakis and Harvey Wasserstein, both well-respected and prominent activists who have written books on Election Fraud. He wrote:

That didn’t sit well with Harvey Wasserman and Bob Fitrakis, whose earlier piece for The Free Press, “Is the 2016 election already being stripped & flipped?,” I had mentioned briefly in the column. They’ve now published a lengthy broadside accusing me, and The Nation, of not being able to handle the truth when it comes to “election theft.” (It’s an odd charge, given that my Nation colleague Ari Berman has done some of the best reporting in the country on vote suppression.)

It’s unfortunate that Wasserman and Fitrakis didn’t engage my argument (or link to my piece so that readers might judge it for themselves). Instead, they waved away the idea that looking at motive is a legitimate way of evaluating the likelihood that a crime has been committed, writing that the argument was “a bit hard to follow.”

Fitrakis and Wasserman join these very serious problems with innuendo and thin evidence to claim that virtually all elections, including the 2016 primaries, are rigged. They dismiss those of us who don’t buy their claims as being incapable of handling the truth. But skepticism goes both ways, and true skeptics require more concrete evidence than Fitrakis and Wasserman are able to offer.

Fitrakis and Wasserman responded: 
 Is Hillary Stealing the Nomination? Will Bernie Birth a Long-Term Movement?
 Bob Fitrakis & Harvey Wasserman, Reader Supported News 4/27/16
http://readersupportednews.org/opinion2/277-75/36568-is-hillary-stealing-the-nomination-will-bernie-birth-a-long-term-movement

In 2016, the first thing to face is the massive disenfranchisement of millions of voters, mostly citizens of color and youth. We are heartened to see Bernie and Hillary joined together in an Arizona lawsuit.But the long lines and urban registration stripping that we saw in Phoenix, Madison, and elsewhere this spring will spell doom for the Democrats if they cannot guarantee their constituencies’ the right to vote in November.

At this point, we’re not optimistic. The efforts at re-enfranchisement are little and late. Among those doing superb work on this stripping of our voter rolls are the great Greg Palast (www.gregpalast.com), Ari Berman of The Nation, and others.But the electronic flipping of the alleged vote count remains a demon black box. The 2000 election was turned from Gore to Bush by electronic manipulations in Volusia County, Florida. The 2004 election was turned from Kerry to Bush in a Chattanooga basement which transformed a 4.2% Democratic lead into a 2.5% GOP victory in 90 dark minutes. All that could happen again in 2016.

Over the years we’ve respected the work of The Nation’s Josh Holland, who’s expressed concern about our reporting on indications of irregularities that seem to favor Hillary over Bernie. But our stated conclusions on them remain far from conclusive. If we thought we had definitive evidence that the Clinton campaign was stealing the nomination from the Sanders campaign, we’d say so in direct, explicit and unmistakable phrases.Simply put: we do NOT at this point believe they rise to the level of provable theft, as we are certain was the case in 2000 and 2004. We understand concerns and welcome the dialogue. But we’d like to avoid the usual circular firing squad. Writing in The Nation, Josh has deemed it important to mention disagreements with our former collaborator Steve Rosenfeld, and our good friend Mark Hertsgaard. Mark’s writing on global warming has been legend. In 2004 he criticized some of our reporting on the Ohio vote count. We disagreed with him then and still do. Nothing in the past 12 years of our research and writing while based in central Ohio has surfaced that would make us change our reporting on how the 2004 election was stolen. Quite the opposite.

But other comments on the nature of electronic election theft throw up a HUGE red flag. And here we worry about a dangerous gap in the work from The Nation and the left as a whole. If international election standards were applied to the 2016 primaries, eight states – Georgia, Massachusetts, Alabama, Texas, Mississippi, Ohio, New York, Tennessee – would be investigated for suspected fraudulent election results, because the actual vote deviates so greatly from the exit polls. Also, the exit polls indicated that Sanders won in Illinois, Massachusetts and Missouri. The e bottom line is this: there is no viable method for monitoring or verifying the electronic vote count in 2016. In a close race, which we expect this fall, the outcome could be flipped in key swing states where GOP governors and secretaries of state are running the elections. This includes most notably Ohio, Michigan, Iowa and Arizona, plus North Carolina and Florida (where the situations are slightly different). Steve has called this “a stretch.” He and Josh seem to dismiss the assertion that an election can be electronically stolen as “conspiracy theory,” apparently based on the idea that such thefts would become obvious fodder for an infuriated media and public outrage.This we find this overly trusting and dangerous. Under our current system there is no way to counter-indicate a stolen electronic vote count except by exit polling, for which Josh has expressed contempt. Exit polls in other countries (especially Germany) are highly reliable; here the raw data is too, but can be hard to get. And it’s now standard procedure to have the public numbers “adjusted” to fit official vote counts, fraudulent or otherwise.And even raw data exit polls have no legal standing. —–

Cyber-security expert Stephen Spoonamore told the Free Pressthat the computer configuration was set up to allow a “man in the middle attack” to alter Ohio’s votes.The late night shift in the 2004 electronic vote count in 10 decisive swing states was by all accounts a “virtual statistical impossibility,” with the odds against that happening in the millions. But now we are being told the idea that this could indicate a stolen election is “conspiracy theory.” PLEASE!!! If someone – anyone! – can demonstrate EXACTLY how the electronic vote count will be monitored, verified and made clear to the media in 2016, and then guarantee that the public and the courts will react with enforceable fury, we will be eternally grateful.We hope in the meantime The Nation will add to Ari Berman’s fine reporting on the stripping of voter eligibilities an in-depth investigation into the “other shoe” of election theft – the flipping of the electronic vote count. Rep. Hank Johnson (D-Ga.) raised the “Diebold question” at a Congressional Black Caucus hearing on April 21, 2016. Johnson noted how easy it would be to hack the old voting machines, many that are over 20 years old, and vowed to introduce legislation that would make voting secure.

Finally, we are often asked how, if the 2000 and 2004 elections were stolen, Obama won in 2008 and 2012. We did, after all, write in 2004 that the 2008 election was being rigged. The answer is simple: it was. But Obama won by far too many votes to have that election credibly stolen. And his campaign was not in denial.  We are happy to hear from Steve that our reporting on Ohio 2004 might have enhanced Obama’s scrutiny on the 2008 vote count. But it should be made clear that Obama’s victory could easily have been flipped had the vote count been closer and had fewer states been so definitively won. We believe he actually won by more than 10 million votes in both 2008 and 2012, but was officially credited with far less.

Where, exactly, is the line beyond which an election can’t be stolen? Do the Democrats need to win by 5%… 10%… to get an official victory? And what then happens to the down-ballot races? We prefer not to see those limits tested again.  And we need to have people prepared to take tangible action. In 2012 Bob Fitrakis filed a successful Election Day lawsuit preventing illegal computer patches being rigged into Ohio’s electronic machines. In a closer race, those patches might have made the difference. We believe the expectation that they would work did cause Karl Rove to do his legendary flipped-out double-take on Fox News as he was told Mitt Romney had lost Ohio. We also reported (as did The Nation) that voting machines in key Cincinnati precincts were financially linked to the Romney family. We each wrote separate articles about that and were each blacklisted by Daily Kos for doing so, even though the vast bulk of Harvey’s 150+ previous blogs on that site were about nuclear power and renewable energy. Some publications that aren’t progressive understand the problem.

Twenty-three minutes into the 2012 Election Day, Forbes took the Free Press reporting seriously, and warned voters of the dangers of private, for-profit companies owning and maintaining voting machines. Over the years we’ve been repeatedly told that we should stop reporting on electronic election theft because it might discourage voter turnout. And that the key to a Democratic victory in 2016 will be another massive vote count victory that will be “too big to steal.” Frankly, we don’t see that happening this year. And we find such talk deeply disturbing.

We have no doubt that innumerable US House and Senate races have been stolen over the years, along with governorships, control of state legislatures, referenda and more, all of it producing a deep reinforcement of the corporate control of our government. We’re also reasonably certain that neither Hillary nor Bernie is likely to amass in November a margin of victory over either Ryan or Trump that would be big enough to negate the possibility of massive disenfranchisement and electronic vote flipping in key states like Ohio, Michigan, Iowa or Arizona. And anyway … why the hell are we even thinking about leaving such a problem unsolved?

This disease needs a definitive cure. We look forward to further reasoned and reasonable dialogue. We invite Josh and Ari to join us on our panel at the upcoming Left Forum in New York in May. We welcome a public discussion with Steve and Mark in California.Above all, we hope to see those millions of Bernie supporters joining us at the reactor sites, the banks, the women’s health centers, the shelters, the schools and so many other critical hot spots in our corporate-plagued society, no matter who wins (or how) in November.

Bob Fitrakis and Harvey Wasserman are co-authors of six books on election integrity, including the new Strip & Flip Selection of 2016: Five Jim Crows and Electronic Election Theft(www.freepress.org and www.solartopia.org). Bob’s Fitrakis Files are at www.freepress.org. Harvey’s Organic Spiral of US History is coming soon at www.solartopia.org.

 

 

 
29 Comments

Posted by on April 29, 2016 in 2016 election, Uncategorized

 

Tags: , , , , , , , , , , , ,

Democratic Primaries 4/26: Exit Poll anomalies (continued)

Richard Charnin
April 27, 2016

There were three exit polls yesterday in CT, MD, PA.  Sanders exit poll share declined from the poll to the vote in two of the three elections. As usual, the exit polls were forced to match the recorded vote.

The difference between Clinton’s  adjusted exit poll and recorded share were: 
CT .01%;  MD 0.10%;  PA -.17%

In 21 of 23 primaries, Sander’s exit poll share exceeded his recorded share.
The probability of this being due to chance:
P = 1 in 30,000 = binomdist(2,23,0.5,true)

In 9 of 23 primaries, Sanders exit poll share exceeded his recorded share by more than the margin of error. The probability of this being due to chance:
P = 1 in 90 million = poisson( 9,0.025*23,false)

The pollsters ALWAYS force the unadjusted exit polls to match the recorded vote. Where are the unadjusted exit polls?  The American Association for Public Opinion Research (AAPOR) discusses the matching process in Explaining Exit Polls.

In close races, the projection models also employ actual vote totals, first in sample precincts as it becomes available and then at the county level for all counties in a state as they become available. It is important to note that after the votes have been counted, the exit poll results are adjusted to match the actual election outcomes. It is in this way that the final exit poll data can be used for its primary and most important purpose – to shed light on why the election turned out the way it did.

Note: It’s not just in close races. Unadjusted exit polls are adjusted to match the recorded vote in ALL races. The pollsters need to provide the actual exit poll respondent data for all precincts.  They need to provide the data and then explain the rationale for making the adjustments to match the vote. But they won’t. Ask yourself WHY? 

The AAPOR never mentions election fraud as a likely cause of the discrepancies. In actuality all of the adjusted exit poll crosstabs contaminate the true statistical results and are misleading as they do not reflect the the actual responses of those exit polled.

Exit pollsters at Edison Research should not be making adjustments. But it is standard operating procedure. It is unscientific and hides the actual exit poll results. It serves to cover-up the fraud which is measured by the recorded vote discrepancy .

https://richardcharnin.wordpress.com/2016/03/02/ma-primary-unadjusted-exit-poll-indicates-bernie-won/

Final Adjusted Exit Polls (forced to match the recorded vote)

CT
1234 total respondents
 Pct Clinton Sanders
men 39% 43% 55%
women 61% 57% 41%
Total 51.54% 46.46%
2-party 52.59% 47.41%
Recorded 52.60% 47.40%
Diff 0.01%
MD
1364 total respondents
 Pct Clinton Sanders
Men 39% 55% 40%
women 61% 68% 29%
Total 62.93% 33.29%
2-party 65.40% 34.60%
Recorded 65.50% 34.50%
Diff 0.10%
PA
1425 total respondents
 Pct Clinton Sanders
Men 39% 49% 50%
Women 61% 60% 39%
Total 55.71% 43.29%
2-party 56.27% 43.73%
Recorded 56.10% 43.90%
Diff -0.17%

https://docs.google.com/spreadsheets/d/1sGxtIofohrj3POpwq-85Id2_fYKgvgoWbPZacZw0XlY/edit#gid=1591963017

http://www.cnn.com/election/primaries/polls/pa/Dem

Inline image

 

 
30 Comments

Posted by on April 27, 2016 in 2016 election, Uncategorized

 

Tags: , ,

The NY Democratic Primary Quiz

The NY Democratic Primary Quiz

Richard Charnin
April 23, 2016

There were 1307 NY Exit Poll  respondents at 9 pm and 1391 at the final – an increase of just 84 respondents.  Adjustments made to force the final 1391 exit poll to match the recorded vote in all exit poll categories are mathematically impossible. Therefore, the recorded vote was also mathematically impossible. The impossible adjustments are irrefutable proof of election fraud.

Let’s review the adjustments as a quiz.

1. At 9pm, Clinton had a) 51%, b) 52%, c) 53%
2. Clinton won the recorded vote with a) 57.3%, b) 57.7%, c) 57.9%
3. She had 28% of 18-29 year-olds. In the final she had a) 33%, b) 35%, c) 37%
4. She had 45% of males. In the final she had a) 49%, b) 50%, c) 51%
5. She had 71% of blacks. In the final she had a) 74%, b) 75%, c) 76%
6. She had 57% of Democrats. In the final she had a) 60%, b) 61%, c) 62%
7. She had 55% of Urban voters. In the final she had a) 59%, b) 60%, c) 62%
8. At 9pm, Urban voters comprised 55% of the total vote.
At the final, they comprised a) 62%, b) 64%, c) 66%

9. At 9pm, Clinton had 680 (52%) of 1307 respondents. She had 802 (57.9%) at the final (1391), an increase of 122 among the 84 final respondents.
This was a) a polling error, b) of no consequence, c)an absolute indicator of fraud.
10. At 9pm, Sanders had 622 (48%) and 589(42.1%) at the final, a 33 vote decline.
This was a) a polling error, b) of no consequence, c)an absolute indicator of fraud.

11.The probability of the 11.8% exit poll discrepancy from the recorded vote is
a) 1 in 91,000, b) 1 in 94,000, c) 1 in 102,000
12. The probability that Sanders exit poll share would be greater than his recorded share in 21 of 23 primaries is a) 1 in 13,000, b) 1 in 30,000, c) 1 in 40,000
13. In the NY Cumulative Vote Share analysis, Sanders and Clinton were tied after
a) 400,000, b) 500,000, c) 600,000 of 1.79 million total votes

Answers
1b, 2c,3b,4b,5b,6c,7c,8b,9c,10c,11c,12b,13c

 
14 Comments

Posted by on April 23, 2016 in 2016 election, Uncategorized

 

Tags: , , , , , , ,

NY Democratic Primary: More frustration

NY Democratic Primary: More Frustration

Richard Charnin
April 20, 2016

Matrix of Deceit: Forcing Pre-election and Exit Polls to Match Fraudulent Vote Counts
Proving Election Fraud: Phantom Voters, Uncounted Votes and the National Poll
Election Fraud Slides

Momentum was on Bernie’s side. He had just won a solid victory in Wisconsin. Now he was coming home. A Brooklyn native, he was drawing great crowds. The following entertainers gave speeches on his behalf: Harry Belafonte, Tim Robbins, Susan Sarandon, Mark Ruffalo, Spike Lee, Rosario Dawson.

Pre-election polls showed that Hillary was a likely winner. But Bernie had the Big Mo; he would overcome the systemic fraud which plagued his campaign in the primaries.

NYC was a natural for Bernie, despite the fact that Hillary Clinton was a senator for eight years. But it was Arizona all over again. Thousands of voters reported their registrations were changed or dropped.

Only 22% of approximately 8 million registered voters turned out. Clinton won by 57.9-42.1%. 

The UNADJUSTED exit poll indicated a close race. Hillary led  by just 52-48%,  an 11.8% discrepancy from the recorded vote.  There were 1391 respondents and a 2.6% exit poll Margin of Error. Clinton led by a whopping 62-38% in the vote count with 33% of precincts reporting.

At 9:03 pm, there were 1307 exit poll respondents, Clinton led the actual count by 680-627 (52.0-48.0%). With just 84 additional respondents (1391 total), Clinton’s lead increased to 802-589 (57.7-42.3%). She had 122 additional respondents and  Sanders had 38 fewer.

How can Clinton gain 122 of 84 respondents? How can Sanders’ total drop?  They can’t. It is mathematically impossible. Therefore the final vote has to be impossible as well. . The exit poll was forced to match the recorded vote with impossible adjustments.

CNN Exit poll- Gender Clinton Sanders
1307 respondents 9:03pm 680 627
 Vote share 52.0% 48.0%
Final EP: 1391 respondents 802 589
Adjusted Vote share 57.7% 42.3%
Change: +84 respondents  +122  -38

This excellent comprehensive analysis confirms that THE NY PRIMARY EXIT POLL USED IMPOSSIBLE ADJUSTMENTS TO MATCH THE RECORDED VOTE.

In 2014, NY voter registration  was  49D-24R-27I. The split was 85D-15I in the exit poll, which (as always) was forced to match the 57.9-42.1% recorded vote.

Assuming primary voting was proportional to registration, the split would have been 65D-35I and the race would have been a tie.  If Clinton had 58% of Democrats, Sanders won the election by 52.5-47.5%.

Final Exit Poll (adjusted) Mix Clinton Sanders Margin
Democrats 85% 62% 38% 24%
Independents 15% 28% 72% -44%
 Total 100% 56.9% 43.1% 13.8%

 

2014 Registration Mix Clinton Sanders Margin
Democrats 65% 62% 38% 24%
Independents 35% 28% 72% -44%
 Total 100% 50.1% 49.9% 0.20%
True Vote Mix Clinton Sanders Margin
Democrats 65% 58% 42% 16%
Independents 35% 28% 72% -44%
Total 47.50% 52.50% -5.00%

Assuming that Sanders’ 48% exit poll was accurate, he must have won the election due to thousands of suppressed votes. Sanders True Vote = 48% exit poll + suppressed vote.

Let’s assume that 5% of registered voters (400,000) were disenfranchised and Sanders had 75%. Then he had 52.9% assuming his 48% exit poll share.

The following table is a sensitivity analysis of Sanders share as a function of his share of the suppressed vote and the exit poll. As you can see, Sanders wins 24 of the 25 scenarios.

Sanders%   Suppressed Vote
Sanders 65.0% 70.0% 75.0% 80.0% 85.0%
Exit Poll   Sanders  True Vote
50% 52.7% 53.6% 54.6% 55.5% 56.4%
49% 51.9% 52.8% 53.7% 54.6% 55.6%
48% 51.1% 52.0% 52.9% 53.8% 54.7%
47% 50.3% 51.2% 52.1% 53.0% 53.9%
46% 49.5% 50.4% 51.3% 52.2% 53.1%

As always, the final CNN exit poll was forced to match the recorded vote. http://www.cnn.com/election/primaries/polls/NY/Dem

View the Early Exit Poll vs. Final (matched to recorded vote) vs. True Vote

 

Probability of 11.8% exit poll discrepancy
……………..Sanders Clinton Margin
Recorded …..42.1%….. 57.9%……15.8%
Exit poll…….48.0%…..52.0%……4.0%

The probability P of the discrepancy is 1 in 126,000.
 P = 1- normdist(0.579,0.52,.026/1.96, true)

Sanders’ exit poll share declined in the recorded vote in 18 out of 19 primaries.
The probability: P=1-binomdist(17,19,.5,true) =  0.000038 = 1 in 26,000.

Cumulative Vote Shares

NY Primary Congressional District  cumulative vote shares indicate a  trend to Clinton with increasing district size reminiscent of increasing, non-intuitive GOP vote shares in Governor and Senate elections. 

Inline image

Final NY Exit Poll – forced to match the recorded vote

Gender Mix Clinton Sanders Margin
men 41% 50% 50% 0.00%
women 59% 63% 37% 26.00%
Total 57.67% 42.33% 15.34%
Age Mix Clinton Sanders Margin
18-29 18% 35% 65% -30%
30-44 24% 53% 47% 6%
45-64 39% 63% 37% 26%
65+ 19% 73% 27% 46%
Total 57.46% 42.54% 14.92%
Race Mix Clinton Sanders Margin
white 59% 50% 50% 0%
black 22% 75% 25% 50%
latino 14% 64% 36% 28%
asian 2% n/a n/a
other 3% n/a n/a
Total 54.96% 45.04% 9.92%
Party id Mix Clinton Sanders Margin
democrats 83% 62% 38% 24%
republicans 3% n/a n/a
independents 14% 28% 72% -44%
Total 55.38% 44.62% 10.76%
Ideology Mix Clinton Sanders Margin
very liberal 29% 44% 56% -12%
somewhat liberal 37% 59% 41% 18%
moderate 29% 67% 33% 34%
conservative 5% n/a n/a
Total 54.02% 45.98% 8.04%
When decided Mix Clinton Sanders Margin
last week 23% 57% 43% 14%
earlier 76% 58% 42% 16%
Total 57.19% 42.81% 14.38%
Area Mix Clinton Sanders Margin
urban 64% 62% 38% 24%
suburban 31% 51% 49% 2%
rural 6% n/a n/a
Total 55.49% 44.51% 10.98%
Region Mix clinton sanders Margin
new york city 52% 63% 37% 36%
long island 9% 58% 42% 16%
hudson valley 16% 56% 44% 12%
urban upstate 14% 50% 50% 0%
rural upstate 9% 42% 58% -16%
Total 57.72% 42.28% 15.44%
 
64 Comments

Posted by on April 20, 2016 in 2016 election, Uncategorized

 

Tags: , ,

NY Democratic Primary: Your forecast

NY Democratic Primary: Your forecast

Richard Charnin
April 18, 2016

You can listen to the pundits who told you that Hillary is leading by 17%, implying that it is all over and it’s time to just anoint Hillary, so why even bother to vote? Or you can actually go out and vote.

Now these so-called  election forecasters (who never mention the fraud factor) are going to try and project the vote to within 1% – with Hillary winning of course. I’m not going to tell you what the final vote will be. I’m not that smart.

But I will say this: Win or lose, Bernie will do much better than reported in the media.

So do the forecast. Below is a matrix table which shows approximately how Bernie will do if he gets 45-65% of the white vote and 45-65% of non-whites. That’s 25 scenarios. It’s called a sensitivity analysis.  Bernie’s total vote share is a function of his share of whites and non-whites.

Fully expect that one of the 25 scenarios will closely match Bernie’s final recorded vote.

Know this: if the CNN exit poll indicates that Bernie had 45% of whites and 45% of non-whites (or 45% overall) do not believe it. In fact, expect that Bernie will do 5-8% better than the recorded vote. Remember the exit poll is always adjusted to match the recorded vote.

So here goes.  Pick out the cell which contains Bernie’s share that you believe will be closest to the final recorded vote. In other words predict Bernie’s recorded share of the white vote and non-white vote, knowing that it will surely understate his True Vote.

For instance, cell C2 = 56.2%

SENSITIVITY ANALYSIS

Race………….. NYReg Sanders Clinton
White…………… 76.6% 55.0% 45.0%
Non-white……… 23.4% 60.0% 40.0%
TOTAL…………….100% 56.2% 43.8%

………………………..Sanders % White
……………………..45.0% 50.0% 55.0% 60.0% 65.0%
………………………A……..B……..C……..D…….E
%Non-White………………. Sanders TOTAL
1. 65.0%………….. 49.7% 53.5% 57.3% 61.2% 65.0%
2. 60.0%………….. 48.5% 52.3% 56.2% 60.0% 63.8%
3. 55.0%………….. 47.3% 51.2% 55.0% 58.8% 62.7%
4. 50.0%………….. 46.2% 50.0% 53.8% 57.7% 61.5%
5. 45.0%………….. 45.0% 48.8% 52.7% 56.5% 60.3%

 
13 Comments

Posted by on April 18, 2016 in 2016 election, Uncategorized

 

Tags: , ,

Election Fraud: The 2016 Democratic Primaries

Election Fraud: The 2016 Democratic Primaries

Richard Charnin
4/13/2016

Bernie Sanders is leading 50.4-49.6% based on the unweighted average of all 34 caucuses and primaries. Let’s accept the reasonable premise that the primaries have been fraudulent and Sanders won in MO, MA, AZ, OH,IL, IA, and NV.  Electoral votes are directly proportional to state voting population.  Clinton has won 11 RED states with 160 EV. Sanders won the other 23 states with 188 EV

Based on late exit polls (which had yet to be adjusted to match the recorded vote), Sanders is leading by an unweighted 52.4-47.0%.  The lead must be even greater since votes were stolen from Bernie in the RED states. Proof? Check the average 8.7% exit poll margin discrepancy from the recorded votes in the Democratic Primaries spread sheet.

Sanders’ exit poll share exceeded his recorded share in n= 17 of N= 18 primaries. The probability P=0.000072 or 1 in 13,797. The spreadsheet function is P= 1-BINOMDIST(n-1,N,0.5,true). There is a 99.9% probability that this anomaly was not due to chance and must have been the result of election fraud.

Wyoming

Bernie was a 56-44% winner in the caucus, yet Hillary won 11 of 18 delegates!  In 12 counties, 54% of Clinton’s votes were surrogates (mail-in), representing 74% of the delegates. Just 27% of Sander’s votes were surrogates. Contrast this to  the Nebraska caucus, where 20% of Clinton’s votes were mail-in.

From CNN: “A Clinton campaign aide said their ‘secret sauce’ in Wyoming was the state’s onerous vote-by-mail rules that required anyone voting by mail to have voted as a Democrat in the 2014 midterms.”  But there is no evidence of such a rule.  The aide was not named.

Wisconsin

Bernie Sanders had 563,127 votes (56.5%) and Hillary Clinton 429.738 (43.1%). But the early exit poll indicates that Bernie most likely  did even better.  At 4pm, the exit poll indicated that Sanders had 68% of white vote.  Whites comprise 88% of  WI voters. Assuming Sanders had just 40% of the non-white vote, he won the election by an estimated 64.6-35.4% (2-party).

The final adjusted exit poll was forced to match the recorded vote. It indicates that whites comprised just 83% of the vote and Sanders had just 59% of them. Blacks  comprised 10% – and Sanders had just 31% . These numbers are not  plausible. A pre-election poll from Public Policy Polling (PPP)  indicated that Sanders was  winning black voters by 51-40%.

The exit poll shows that 7% of voters were Latino (3%), Asian (2%), Other (2%). According to the pollsters,  the vote shares are NA. How is that? Was it because their respective turnout rates were too low? The pollsters could have combined the 7% as Other Non-whites. Without this information, we cannot calculate the total recorded vote shares. The abbreviated totals have Sanders winning by 52.1-40.1%. The 12% margin is close to the official recorded margin.

Arizona

Arizona is the latest poster child of election fraud,  along with Florida 2000 and Ohio 2004. Sanders won Utah (a bordering state) and Idaho primaries with nearly 80% of the vote. But he lost in Arizona by 60-38%. Who believes it?

The  National Exit Pool (NEP) of six major media conglomerates funds exit pollster Edison Research. The NEP decided not to poll AZ.  It’s as if they knew they would have to match the unadjusted poll to a bogus recorded vote; the massive discrepancies would be too obvious. But  the networks called it for Hillary  with less than 1% of the votes in. How did they know this if they did not exit poll? Luckily the Yavapai County Daily Courier did an exit poll – and Bernie led by 63-37%. Hillary won  the county by 54-43%- an impossible 37% difference in margin. But the evidence of fraud goes  much further than this one poll.

Of the 15 Arizona counties, Maricopa (Phoenix) is by far the largest with nearly 60% of the vote. Pima County (Tucson)  is second with 16%. In the 2008 primary, Maricopa voter turnout was  54.3%. In the other 14 counties, there was a 47.2% turnout. In 2016, 13  counties had higher voter turnout rates than in 2008. The 4.1% decline (17,000 votes) in Maricopa 2016 turnout (50.2%) from 2008 is counter-intuitive. Voter  turnout in the other 14 AZ counties increased by 8.8% to 56.0%.

Based on the overall trend, Maricopa should have had an approximate 63.1% turnout. It is  a powerful indicator of  voter suppression. The  probability of the 12.9% difference  (160,000 votes) between Maricopa’s projected 62.1% voter turnout and the actual 50.2% turnout  is approximately  1 in  90 trillion.

The  probability of the  5.8% difference in voter turnout  between 14 AZ counties (56.0%) and Maricopa (50.2%) is approximately 1 in 13,000).

Super Tuesday

In the five unadjusted exit polls there were 7,220 respondents. Clinton led by 53.2-44.7%. In the final adjusted polls , there were 7979 respondents (759 additional). She led the final adjusted polls (which were matched to the recorded vote) by 55.6-42.4%. Clinton had 586 (77.2%) of the FINAL 759 respondents, or 21.9% above her unadjusted share. Sanders had 20% (24.7% below his unadjusted share).

Inline image

Michigan

Sanders did much better than his recorded vote in the Michigan primary.  Sanders had 590,386  votes (49.8%) and Clinton 570,948 (48.3%).   Sanders won in 73 of 83  MI counties with 56% of the vote. He won the preliminary exit poll by 52.1-45.9%, a 97% win probability. Clinton won urban counties Wayne and Oakland  with approximately 55% of the vote.

Once again, we have multiple confirmation indicating fraud: Cumulative vote shares, preliminary exit poll, absentee vote anomalies and other anecdotal information.

Cumulative Vote Shares  are a likely indicator of fraud. The lines should be nearly parallel, but invariably, vote shares rise for establishment candidates in urban Democratic counties. It should be conventional wisdom by now: in state elections, fraud abounds in heavily populated urban and suburban locations. Of course, the media never talks about it. They report the recorded numbers as if there was not a fraud factor. 

In the CVS analysis, Sanders had approximately 56% at the 600,000 mark. Notice the abrupt change to straight lines at the 600,000 vote mark. They represent the largest counties (Wayne and Oakland) which used ES&S optical scanners exclusively. 

Sanders had   54% of approximately 500,000 votes cast on AccuVote and Sequoia  voting machines. Clinton had  75% of approximately 240,000 absentee votes and  51.2% of approximately 700,000 votes cast on ES&S Mod 100 machines. The percentages are highly suspect.

Sanders’ county vote shares  were negatively correlated to machine types. The ES&S Model 100 correlation was  -0.68. The bigger the county the lower Sanders’ vote share. Wayne and Oakland counties used ES&S Model 100 optical scanners. Macomb used both ES&S and Premier/Diebold/Dominion AccuVote optical scanners.

 Massachusetts

Late changes to the exit poll indicate that the election was likely stolen.  Sanders  led the Unadjusted Exit Poll Gender crosstab  (1297 respondents) by 52.3-45.7% a 97% win probability.. The poll was captured from CNN at 8:01pm.

But as always, the exit poll was adjusted to match the recorded vote. Clinton led the adjusted exit poll (1406 respondents) by 50.3-48.7%,  a near-exact match to the  RECORDED vote margin.  But her 50.3% share was IMPOSSIBLE.  The proof is self-explanatory: How could Clinton gain  114 respondents and Sanders just 7 among the final 109 exit poll respondents?

Clinton won  by 51-49% on electronic voting machines from ES&S, Diebold and Dominion. Sanders won 68  hand-counted precincts by 58-41% (32,360 votes, 2.7% of votes cast).  He won 250 of  351 jurisdictions and had at least 58% in 110. 

There is a 97%  probability  that Sanders won the election given the 3.55% Margin of Error. The MoE includes the exit poll cluster effect  (30% of the 2.72% calculated MoE). Sanders 53.4% two-party share and the MoE are input to the Normal distribution function to calculate his win probability.

Matrix of Deceit: Forcing Pre-election and Exit Polls to Match Fraudulent Vote Counts
Proving Election Fraud: Phantom Voters, Uncounted Votes and the National Exit Poll (E-book)
LINKS TO WEB/BLOG POSTS FROM 2004

Election Fraud Overview

 

 
14 Comments

Posted by on April 13, 2016 in 2016 election, Uncategorized

 

Tags: , , , ,

The 2016 Democratic Primary Quiz

The 2016 Democratic Primary Quiz

Richard Charnin
April 11, 2016

Data and calculations used in this quiz are from the Democratic Primaries Spreadsheet  which is linked from my blog posts in the 2016 election category.

      1. Iowa caucuses: Hillary won how many consecutive coin flips?  a) 6, b)8, c)10
      2. Nevada caucuses: which group changed their shirts from blue  to red to make it appear they were changing their allegiance  from Bernie to Hillary a) retail workers, b) union workers, c) nurses.
      3.  Massachusetts: Clinton won by 50.3-48.7%.  Sanders won the exit poll by 52.1-45.7%. He won hand counted paper ballots with a) 54%, b) 55%, c) 58%.
      4. Online Debate polls:  Bernie won a) 60%, b) 70%, c) 90% of viewers.
      5. Michigan exit poll: Bernie had a) 50%, b) 51%,c) 52% of the vote.
      6. Missouri: Bernie had 49.9% of the two-party vote.  He had a) 50.9%, b) 51.9%, c) 52.9%% in the unadjusted exit poll.
      7. South Carolina: Hillary had  73% of the vote. Voting was by a) DRE, b) Optiscan, c) Paper ballots.
      8. Georgia: Hillary had  72% of the vote. Voting was by a) DRE, b) Optiscan, c) Paper ballots.
      9. Arizona: 4.1% decline (17,000 votes) in Maricopa County  voter turnout (50.2%) from 2008. Turnout in the other 14 AZ counties increased by  8.3%. This is a) a statistical fluke, b) of no consequence, c) indicates massive fraud.
      10. Arizona Yavapai County Daily Courier exit poll: Bernie led by 63-37%. Hillary won  the county  with a) 52%, b) 54%, c) 56%
      11. Sanders performance:  He  has won 4 of 21 primaries with a 41% share and 12 of 13 caucuses with a 65% share. The 24% vote share discrepancy is a) of no significance, b) just a statistical fluke, c) indicates massive election fraud. 
      12. Wisconsin final exit poll:  vote shares are not available for 124 (7%) minority respondents.  This is because a) they did not respond to the exit poll, b) their responses were insignificant, c) there is no logical explanation
      13. Wisconsin: Sanders had a 51% share of blacks in the final pre-election poll. But he had only 31% in the final exit poll (which is always forced to match the recorded vote). The margin of error is 3%.  The 20% difference is a) plausible, b) extremely unlikely, c) indicates the exit poll was changed  to match Sanders 56.5% recorded share.
      14. Wisconsin: Sanders had 68% of white voters at the 4pm exit poll timeline, but just 59% in the final exit poll (which is always forced to match the recorded vote). The margin of error is 3%. The 9% difference is a) plausible, b) extremely unlikely, c) indicates the exit poll was changed  to match Sanders 56.5% share.
      15. Adjusted exit polls: forced to match the recorded vote. The pollsters will not release the unadjusted polls. This is a) standard operating procedure, b) unscientific, c) makes it appear they have something to hide, d) all of the above.
      16. Wyoming caucus: Bernie won by 56-44% yet Hillary picked up 11 delegates to Bernie’s 7. This is a) no big deal, b) the way the system works, c) is more proof of DNC rigging the election for Hillary.
      17. Average vote shares: Based on the unweighted average of 34 caucuses and primaries, Sanders has a) 46.5%, b)48.9%, c) 50.4%
      18. Average exit poll shares: Based on the unweighted average of 34 caucuses and primaries, Sanders has a) 48.9%, b)51.9%, c)52.4%
      19. New York: Hillary won the recorded vote 57.9-42.1%. She won the exit poll by 52-48%. The probability that the 11.8% discrepancy was due to chance is: a) 1 in 97,000, b) 1 in 126,000, c) 1 in 242,000
      20. Probability of Sanders exit poll discrepancy from the recorded vote exceeding the margin of error in 9 of 23 primaries is                                          a ) 1 in 70 million, b) 1 in 80 million, c) 1 in 90 million

Answers: 1a 2c 3c 4c 5c 6b 7a 8a 9c 10c 11c 12c 13c 14c 15d 16c 17c 18c 19b 20c

 
7 Comments

Posted by on April 11, 2016 in 2016 election, Uncategorized

 

Tags: , ,

 
Richard Charnin's Blog

JFK Conspiracy and Systemic Election Fraud Analysis

Follow

Get every new post delivered to your Inbox.

Join 1,600 other followers