Richard Charnin

Aug. 29, 2017

**77 Billion to One: 2016 Election Fraud**

**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**

**LINKS TO POSTS**

**Last 3 Elections: Exact Forecast of Electoral Vote**

2016 Pre-election Model – Calculating the Expected Electoral Vote

This is for those interested in Electoral Vote math based on pre-election polls. It discusses basic probability and spreadsheet functions. You won’t see a discussion of this anywhere else.The MSM doesn’t care for critical thinking. Perhaps because they are incapable of it.

One of the methods I have used in pre-election forecast modeling is to calculate the Expected Recorded Electoral Vote as well as the True Vote. Important Note: the RECORDED EV is based on MSM pre-election polls which are usually biased for the establishment candidate. In 2016, Clinton was the establishment candidate.

As I did not have 51 state pre-election polls, I used the following method to estimate them based on the average of nine pre-election national polls and Party-ID:

1) Each state’s estimated Party-ID was calculated using the proportional change from the 2012 National Party-ID to the 2016 Gallup National Voter affiliation survey: 40% Independents, 32% Democrats and 28% Republicans.

2) The average vote shares of nine national pre-election polls were applied to the Party-ID of each state to derive the projected state vote shares.

The Expected EV is based on state win probabilities. Calculating the pre and post-election TRUE EV is much more complicated.

In the 2016 Forecast Model, Trump’s Expected EV (before undecided voters) was 305.5, exactly matching his recorded 306 EV. His Snapshot 307 EV is the sum of the EVs for states that he was projected to win. Trump led the weighted average pre-election polls (before undecided voter allocation) by 44.1-43.1%.

View the Recorded votes and two True Vote Models for all the states:

https://docs.google.com/spreadsheets/d/10dlTnin814phKJWjYdkG-ujNKak3zo6ywIP0u0-TGFg/edit#gid=667189511

The following steps calculate the Expected RECORDED EV:

1. Using state forecasts derived from the National Gallup Voter Affiliation survey, calculate the probability P(i) of winning each state using Trump’s projected 2-party vote share. Assume a 3.0% margin of error.

**P(i) = normdist(Trump%/(Trump%+Clinton%),0.5,.03/1.96,true)**

2. Multiply the state win probability by the state electoral vote.

S(i) = P(i)* EV(i), i =1,51

3. **Expected EV = sum [P(i)* EV(i)], i = 1,51**

View the spreadsheet https://docs.google.com/spreadsheets/d/10dlTnin814phKJWjYdkG-ujNKak3zo6ywIP0u0-TGFg/edit#gid=1036175945

State Electoral votes are in the range B129:B179

Trump’s state forecasts are in the range D129:D179

Corresponding state win probabilities are in the range J129:J179

The Expected EV calculation is in cell I128.

**Expected EV = 305.5 = sumproduct(J129:J179, B129:B179)**

John Kesich

September 7, 2017 at 6:58 pm

Have you applied Benford’s law to election data to detect potential fraud? Any thoughts on this approach?

Richard Charnin

September 12, 2017 at 1:02 am

No, I have not. I am not aware of evidence to suggest that Benson’s Law detects fraud.

There is evidence that Cumulative Vote shares is an indicator.

https://richardcharnin.wordpress.com/2016/12/22/2016-cumulative-vote-shares-illinois-michigan-california/

https://richardcharnin.wordpress.com/2016/01/06/2014-governor-election-sensitivity-analysis-voter-registration-and-turnout/