# Tag Archives: monte carlo simulation

## Historical Overview of Election Fraud Analysis

Richard Charnin
Jan.31, 2013

http://richardcharnin.com/

Historical Overview

I have written two books on election fraud which prove that the official recorded vote has deviated from the True Vote in every election since 1968 – always favoring the Republicans. Voting machine “glitches” are not due to machine failures; they are caused by malicious programming.

In the 1968-2012 Presidential elections, the Republicans won the average recorded vote by 48.7-45.8%. The 1968-2012 National True Vote Model (TVM) indicates the Democrats won the True Vote by 49.6-45.0% – a 7.5% margin discrepancy.

In the 1988-2008 elections, the Democrats won the unadjusted state exit poll aggregate by 52-42% – but won the recorded vote by just 48-46%, an 8% margin discrepancy. The state exit poll margin of error was exceeded in 126 of 274 state presidential elections from 1988-2008. The probability of the occurrence is ZERO. Only 14 (5%) would be expected to exceed the MoE at the 95% confidence level. Of the 126 which exceeded the MoE, 123 red-shifted to the Republican. The probability P of that anomaly is ABSOLUTE ZERO (5E-106). That is scientific notation for

P= .000000000 000000000 000000000 000000000 000000000 000000000 000000000 000000000 000000000 000000000 000005.

The proof is in the 1988-2008 Unadjusted State Exit Polls Statistical Reference. Not one political scientist, pollster, statistician, mathematician or media pundit has ever rebutted the data or the calculation itself. They have chosen not to discuss the topic. And who can blame them? Job security is everything.

Election forecasters, academics, political scientists and main stream media pundits never discuss or analyze the statistical evidence that proves election fraud is systemic – beyond a reasonable doubt. This site contains a compilation of presidential, congressional and senate election analyses based on pre-election polls, unadjusted exit polls and associated True Vote Models. Those who never discuss or analyze Election Fraud should focus on the factual statistical data and run the models. If anyone wants to refute the analytic evidence, they are encouraged to do so in a response. Election forecasters, academics and political scientists are welcome to peer review the content.

The bedrock of the evidence derives from this undisputed fact: National and state actual exit poll results are always adjusted in order to force a match to the recorded vote – even if doing so requires an impossible turnout of prior election voters and implausible vote shares.

All demographic categories are adjusted to conform to the recorded vote. To use these forced final exit polls as the basis for election research is unscientific and irresponsible. The research is based on the bogus premise that the recorded vote is sacrosanct and represents how people actually voted. Nothing can be further from the truth.

It is often stated that exit polls were very accurate in elections prior to 2004 but have deviated sharply from the recorded vote since. That is a misconception. UNADJUSTED exit polls have ALWAYS been accurate; they closely matched the True Vote Model in the 1988-2008 presidential elections. The adjusted, published exit polls have always matched the fraudulent RECORDED vote because they have been forced to. That’s why they APPEAR to have been accurate.

The Census Bureau indicates that since 1968 approximately 80 million more votes were cast than recorded. And these were just the uncounted votes. What about the votes switched on unverifiable voting machines and central tabulators? But vote miscounts are only part of the story. The True Vote analysis does not include the millions of potential voters who were illegally disenfranchised and never got to vote.

In 1988, Bush defeated Dukakis by 7 million recorded votes. But approximately 11 million ballots (75% Democratic) were uncounted. Dukakis won the unadjusted exit polls in 24 battleground states by 51-47% and the unadjusted National Exit Poll by 50-49%. The Collier brothers classic book Votescam provided evidence that the voting machines were rigged for Bush.

In 1992, Clinton defeated Bush by 5.8 million recorded votes (43.0-37.5%). Approximately 9 million were uncounted. The National Exit Poll was forced to match the recorded vote with an impossible 119% turnout of living 1988 Bush voters in 1992. The unadjusted state exit polls had Clinton winning a 16 million vote landslide (47.6-31.7%). The True Vote Model indicates that Clinton won by 51-30% with 19% voting for third party candidate Ross Perot.

In 1996, Clinton defeated Dole by 8.6 million recorded votes (49.3-40.7%); 9 million were uncounted. The unadjusted state exit polls (70,000 respondents) had Clinton winning a 16 million vote landslide (52.6-37.1%). The True Vote Model indicates that Clinton had 53.6%.

In 2000, Al Gore won by 540,000 recorded votes (48.4-47.9%). But the unadjusted state exit polls (58,000 respondents) indicated that he won by 50.8-44.4%, a 6 million vote margin. There were nearly 6 million uncounted votes. The True Vote Model had him winning by 51.5-44.7%. But the Supreme Court awarded the election to Bush (271-267 EV). In Florida, 185,000 ballots were uncounted. The following states flipped from Gore in the exit poll to Bush in the recorded vote: AL AR AZ CO FL GA MO NC TN TX VA. Gore would have won the election if he captured just one of the states. Democracy died in this election.

In July 2004 I began posting weekly Election Model projections based on the state and national polls. The model was the first to use Monte Carlo Simulation and sensitivity analysis to calculate the probability of winning the electoral vote. The final projection had Kerry winning 337 electoral votes and 51.8% of the two-party vote, closely matching the unadjusted exit polls.

The adjusted 2004 National Exit Poll was mathematically impossible since it indicated that there were 52.6 million returning Bush 2000 voters. But Bush had just 50.5 million recorded votes in 2000 – and only 48 million were alive in 2004. Approximately 46 million voted, therefore the adjusted Final NEP overstated the number of returning Bush voters by 6.5 million. In order to match the recorded vote, the NEP required an impossible 110% living Bush 2000 voter turnout in 2004.

The post-election True Vote Model calculated a feasible turnout of living 2000 voters based on Census total votes cast (recorded plus net uncounted), a 1.25% annual mortality rate and 98% Gore/Bush voter turnout. It determined that Kerry won by 67-57 million and had 379 EV. But Kerry’s unadjusted state exit poll aggregate 51.0% share understated his True Vote Model. There was further confirmation of a Kerry landslide.

Consider the Final National Exit Poll adjustments made to Bush’s approval rating and Party–ID crosstabs.

Bush had a 48% national approval rating in the final 11 pre-election polls. But the Final adjusted National Exit Poll indicated that he had a 53% approval rating – even it was 50% in the unadjusted state exit poll weighted aggregate. Given the 3% differential between the Final NEP and state exit poll ratings, let’s deduct 3% from his 48% pre-election approval. This gives Bush a 45% vote share – a virtual match to the True Vote Model. The exit pollsters had to inflate Bush’s final pre-election 48% average rating by 5% in the NEP in order to have it match the recorded vote – and perpetuate the fraud. There was a near-perfect 0.99 correlation ratio between Bush‘s state approval and unadjusted exit poll share.

Similarly, the unadjusted state exit poll Democratic/Republican Party ID split was 38.8-35.1%. In order to force the National Exit Poll to match the recorded vote, they needed to indicate a bogus 37-37% split.

The correlation between state Republican Party ID and the Bush unadjusted shares was a near-perfect 0.93. This chart displays the state unadjusted Bush exit poll share, approval ratings and Party-ID.

The Final 2006 National Exit Poll indicated that the Democrats had a 52-46% vote share. The Generic Poll Trend Forecasting Model projected that the Democrats would capture 56.43% of the vote. It was within 0.06% of the unadjusted exit poll.

In the 2008 Primaries, Obama did significantly better than his recorded vote.

The 2008 Election Model projection exactly matched Obama’s 365 electoral votes and was within 0.2% of his 52.9% share (a 9.5 million margin). But the model understated his True Vote. The forecast was based on final likely voter (LV) polls that had Obama leading by 7%. The registered voter (RV) polls had him up by 13% – before undecided voter allocation. The landslide was denied.

The Final 2008 National Exit Poll was forced to match the recorded vote by indicating an impossible 103% turnout of living Bush 2004 voters and 12 million more returning Bush than Kerry voters. Given Kerry’s 5% unadjusted 2004 exit poll and 8% True Vote margin, one would expect 7 million more returning Kerry than Bush voters – a 19 million discrepancy from the Final 2008 NEP. Another anomaly: The Final 2008 NEP indicated there were 5 million returning third party voters – but only 1.2 million were recorded in 2004. Either the 2008 NEP or the 2004 recorded third-party vote share (or both) was wrong. The True Vote Model determined that Obama won by over 22 million votes with 420 EV. His 58% share was within 0.1% of the unadjusted state exit poll aggregate (83,000 respondents).

In the 2010 Midterms the statistical evidence indicates that many elections for House, Senate, and Governor, were stolen. The Wisconsin True Vote Model contains worksheets for Supreme Court and Recall elections. A serious analyst can run them and see why it is likely that they were stolen.

In 2012, Obama won the recorded vote by 51.0-47.2% (5.0 million vote margin) and once again overcame the built-in 5% fraud factor. The 2012 Presidential True Vote and Election Fraud Simulation Model exactly forecast Obama’s 332 electoral vote based on the state pre-election polls. The built-in True Vote Model projected that Obama would win by 56-42% with 391 electoral votes. But just 31 states were exit polled, therefore a comparison between the True Vote Model and the (still unreleased) state and national unadjusted exit polls (i.e. the red-shift) is not possible. Obama won the 11.7 million Late votes recorded after Election Day by 58-38%. In 2008, he won the 10.2 million late votes by 59-37%. The slight 2% margin difference is a powerful indicator that if a full set of 2012 unajusted state and national exit polls were available, they would most likely show that Obama had 55-56% True Vote share.

## Track Record: 2004-2012 Election Forecast and True Vote Models

Track Record: 2004-2012 Election Forecast and True Vote Models

Richard Charnin
Jan. 19, 2013

This is a summary of 2004-2012 pre-election projections and corresponding recorded votes, exit polls and True Vote Models.

Note that the Election Model forecasts are based on final state pre-election Likely Voter (LV) polls, a subset of the total Registered Voters (RV) polled. The LVs always understate Democratic voter turnout; many new (mostly Democratic) voters are rejected by the Likely Voter Cutoff Model (LVCM). In addition, pre-election polls utilize previous election recorded votes in sampling design, rather than total votes cast. Total votes cast include net uncounted votes which are 70-80% Democratic. The combination of the LVCM and uncounted votes results in pre-election polls understating Democratic turnout – and their projected vote share.

2004 Election Model
Kerry Projected 51.8% (2-party), 337 EV (simulation mean), 322 EV snapshot
Adjusted National Exit Poll (recorded vote): 48.3-50.7%, 252 EV
Unadjusted State exit poll aggregate: 51.1-47.6%, 349 EV snapshot, 336 EV expected Theoretical)
True Vote Model: 53.6-45.1%, 364 EV

2006 Midterms
Democratic Generic 120-Poll Trend Projection Model: 56.4-41.6%
Adjusted Final National Exit Poll (recorded vote): 52.2-45.9%
Wikipedia recorded vote: 57.7-41.8%

2008 Election Model
Obama Projected: 53.1-44.9%, 365.3 expected EV; 365.8 EV simulation mean; 367 EV snapshot
Adjusted National Exit Poll (recorded vote): 52.9-45.6%, 365 EV
Unadjusted State exit poll aggregate: 58.1-40.3%, 419 EV snapshot, 419 expected EV
True Vote Model: 58.0-40.4%, 420 EV

2012 Election Model
Obama Projected: 51.6% (2-party), 332 EV snapshot; 320.7 EV expected; 321.6 EV simulation mean
Adjusted National Exit Poll (recorded): 51.0-47.2%, 332 EV
True Vote Model 56.1%, 391 EV (snapshot); 385 EV (expected)
Unadjusted State Exit Polls: not released
Unadjusted National Exit Poll: not released

Posted by on January 19, 2013 in Uncategorized

## Final Forecast: The 2012 True Vote/ Election Fraud Model

Final Forecast: The 2012 True Vote/ Election Fraud Model

Richard Charnin
Nov.5, 2012

Are there any forecasters in the corporate media who discuss systemic election fraud and include a True Vote analysis in their models? I have not seen any. The pundits ignore election fraud completely by limiting their projections to the recorded vote. But they are missing the big story which can be expressed by the simple formula:

Recorded Vote = True Vote + Fraud factor

The forecast: Obama has 320.7 expected electoral votes (see the definition of the expected value below) and a 332 snapshot EV. His 99.4% win probability is based on 497 electoral vote wins out of 500 trial simulations. His projected popular vote share margin is 51-48%, a 69-65 million vote margin.

But the recorded vote is not the True Vote. The True Vote is never the same as the recorded vote. The True Vote Model indicates Obama would have approximately 371 EV, a 55% vote share and win by 74-59 million votes in a fraud-free election.

Obama’s 332 snapshot EV assumes he will win all the battleground states except for NC. The races are very close in CO (9), FL (29), NC (15), NH (4), NV (6), OH (18) WI (10), VA (13) and that is why the expected EV is just 320.7. But keep in mind that the projections are based on LV polls which a) always understate Democratic turnout and b) are at least partially based on previous election bogus recorded votes.

If FL, OH and NC are stolen, Obama will likely lose. Even though he won the True Vote, it was not enough to overcome the FRAUD FACTOR.

To the pundits, the Fraud Factor is zero. They are not paid to project the True Vote. Their projections are based on Likely Voter polls which are always close to the popular recorded vote. The public has always been led to believe that the recorded vote was in fact the True Vote. It never is. The historical uncounted votes which are 70-80% Democratic prove it. And there have been approximately 40-45 million uncounted ballots in the last six presidential elections – according to the U.S. Census. That’s the bad news.

The good news is that finally, after 12 long years, there is a near critical mass of election fraud awareness. The 2000 and the 2004 elections have been proven to be stolen. Of course, the media pundits know this. But they like their jobs too much to defy their editors.

But the word is finally getting out after decades of media silence and misinformation. Yes, it’s a conspiracy, all right – a conspiracy fact, not a theory. The simple fact is that the conspiracy is the media and politicians who have kept the facts about our broken electoral system hidden from the public. What is the proof? The proof is…they never talk about the millions of uncounted votes or the proprietary voting machines owned and serviced by right-wing organizations – who just so happen to also count the votes..

Unlike the other election forecasters in the media and academia, the 2012 True Vote/ Election Fraud Forecast model projects both the True Vote and the official Recorded vote.

- The Monte Carlo electoral vote simulation is based on the latest state likely voter (LV) polls.
- The True Vote Model is based on plausible turnout estimates of new and returning 2008 voters and corresponding vote shares.

The LV polls are based partially on a Likely Voter Cutoff Model which always reduces projected (Democratic) turnout. Another factor to keep inmind is that the polls are at least somewhat based on prior election recorded votes – which are themselves tainted.

Even so, Obama has a 99% probability of winning the Electoral Vote (EV). Models which indicate an 80% win probability based on the latest polls cannot be correct – probably because they include extraneous factor variables. An experienced modeler knows how to KISS (keep it simple stupid).

Only 500 election simulation trials are necessary to determine the EV win probability. Anything more than that is overkill. Calculating the expected EV does not require a million scenario combinations, either.

Assuming the polls, the state win probabilities p(i) can be calculated. The expected EV is just a simple summation based on the expected state electoral votes: Expected EV = ∑p(i)* EV(i), where i =1,51 states.

Election Model Forecast; Post-election True Vote Model

2004 Election Model (2-party shares)
Kerry 51.8%, 337 EV (snapshot)
State exit poll aggregate: 51.7%, 337 EV
Recorded Vote: 48.3%, 255 EV
True Vote Model: 53.6%, 364 EV

2008 Election Model
Obama 53.1%, 365.3 EV (simulation mean);
Recorded: 52.9%, 365 EV
State exit poll aggregate: 58.0%, 420 EV
True Vote Model: 58.0%, 420 EV

2012 Election Model
Obama Projected: 51.6% (2-party), 332 EV snapshot; 320.7 expected; 321.6 mean
Adjusted National Exit Poll (recorded): 51.0-47.2%, 332 EV
True Vote Model 56.1%, 391 EV (snapshot); 385 EV (expected)
Unadjusted State Exit Polls: not released
Unadjusted National Exit Poll: not released

Posted by on November 5, 2012 in 2012 Election

## Calculating the Projected Electoral Vote

Calculating the Projected Electoral Vote

Oct. 26, 2012

The 2012 Election Forecast Simulation Model calculates the projected electoral vote in three ways.

1. Snapshot EV: The state electoral vote goes to the projected leader based on the pre-election poll. This is a crude estimate in close races in which the projected margin is within 1-3%.

2. Expected EV: The probability of winning each state is calculated using the poll-based projection. The theoretical forecast electoral vote is the weighted sum of the state win probabilities and corresponding electoral votes. EV= ∑ P(i) * EV (i), i =1,51 states. This is the best estimate for the projected Electoral Vote.

3. Simulation Mean EV: The mean electoral vote is a simple average of the simulated trial elections. It calculated mean approaches the theoretical expected EV as the number of trials increase (500 is sufficient), illustrating the Law of Large Numbers. A Monte Carlo simulation is needed to calculate the probability of winning the election. It is simply the number of winning trials divided by 500.

The Final Nov.6 model forecast that Obama would have a 332 Snapshot EV (exactly matching his actual EV), a 320.7 Expected EV and 320.8 Simulation Mean EV. But the Expected EV is a superior forecast tool since it eliminates the need for stating that “the states are too close to call”.

Election Model Forecast; Post-election True Vote Model

2004 Election Model (2-party shares)
Kerry 51.8%, 337 EV (snapshot)
State exit poll aggregate: 51.7%, 337 EV
Recorded Vote: 48.3%, 255 EV
True Vote Model: 53.6%, 364 EV

2008 Election Model
Obama 53.1%, 365.3 EV (simulation mean);
Recorded: 52.9%, 365 EV
State exit poll aggregate: 58.0%, 420 EV
True Vote Model: 58.0%, 420 EV

2012 Election Model
Obama Projected: 51.6% (2-party), 332 EV snapshot; 320.7 expected; 321.6 mean
Adjusted National Exit Poll (recorded): 51.0-47.2%, 332 EV
True Vote Model 56.1%, 391 EV (snapshot); 385 EV (expected)
Unadjusted State Exit Polls: not released
Unadjusted National Exit Poll: not released

Posted by on October 27, 2012 in 2012 Election, Uncategorized

## Final Forecast: 2012 Presidential True Vote/Election Fraud Model

Final Forecast: 2012 Presidential True Vote/Election Fraud Model

Richard Charnin
Nov 5, 2012

The final 2012 Presidential True Vote/Election Fraud Model exactly forecast Obama’s 332 electoral vote. His projected 51.6% two-party recorded share was close to the actual 51.9%. But Obama actually did much better in the True Vote forecast (391 EV, 56% two-party). As usual, the systematic fraud factor was in effect, causing a 4-5% red-shift. But Obama overcame the fraud, just as he did in 2008.

The final 2008 Election Model was also right on the money. It forecast Obama would have 53.1% and 365.3 expected EV compared to his actual recorded 52.9% share and 365 EV. But he had a 58.0% True Vote Model share and 420 EV. His 58.0% share of the unadjusted state exit polls (76,000 respondents) confirmed the True Vote Model.. He won unadjusted National Exit Poll by 61-37% (17,836 respondents).

The Presidential True Vote and Monte Carlo Simulation Forecast Model is updated on a daily basis. The election is assumed to be held on the latest poll date.

Final Forecast: 11/06/2012 9am
Obama: 320.7 expected electoral votes; 99.6% win probability (498 of 500 trials).
He has a 332 snapshot EV (actual total).
He leads the state poll weighted average by 49.3-46.2% (51.6% 2-party share).
He leads 50.4-47.0% in 16 of 18 Battleground states with 184 of 205 EV.

Obama leads Romney in the RCP National average: 48.8-48.1%.
Rasmussen and Gallup are Likely Voter (LV) polls which lean to the GOP.
Rasmussen: Romney 49-48%.
Gallup: Romney 50-49%. It was 51-46% a week ago.

Obama leads in the Rand poll 49.5-46.2% (closely matching the state polls). Unlike the national LV polls, the Rand poll doesn’t eliminate respondents but rather weights them on a scale of 1-10 (based on voter preference and intention to vote).

The 3% Obama margin increase in the Rand poll over the national LV polls illustrates why the LVs understate Obama’s margin by using the Likely Voter Cutoff Model (LVCM). LV polls are a subset of the registered voter (RV) sample. They always understate the Democratic share. The majority of voters eliminated by the Likely Voter Cutoff Model (LVCM) are Democrats.

The True Vote Model indicates that Obama would have 55.2% of the two-party vote with 371 expected EV in a fraud-free election. Will he be able to overcome the systemic fraud factor?

2012 Presidential True Vote and Monte Carlo Simulation Forecast Model (html)
- The Monte Carlo Electoral Vote Simulation is based on the latest state polls and currently assumes an equal split of undecided voters. The expected electoral vote is the sum of the products of the state win probabilities and corresponding electoral votes.

- The True Vote Model is based on plausible turnout estimates of new and returning 2008 voters and corresponding vote shares.

The model calculates an estimated True Vote forecast for the National aggregate or any state. The calculation is displayed below the input data section. State poll-based national vote shares, electoral vote and probabilities are displayed on the right side of the screen.

This worksheet contains the weekly polling trend analysis.

The polling data is from the Real Clear Politics (RCP) and Electoral-vote.com websites. The simulation uses the latest state polls.

View this 500 election trial simulation electoral vote frequency graph.

1988-2008: 274 State exit polls. An 8% Discrepancy

In the six presidential elections from 1988-2008, the Democrats won the average recorded vote by 48-46%. But they led both state and national exit polls by 52-42%. There were approximately 375,000 respondents in the 274 state polls and 90,000 respondents in the six national polls. Overall, an extremely low margin of error.

1988-2008 Unadjusted State and National Exit Poll Database

The Ultimate Smoking Gun that proves Systemic Election Fraud:
1) The Likely Voter Cutoff Model eliminates newly registered Democrats from the LV sub-sample. Kerry had 57-61% of new voters; Obama had 72%.
2) Exit poll precincts are partially selected based on the previous election recorded vote.
3) In the 1988-2008 presidential elections, 226 of 274 exit polls red-shifted to the Republicans. Only about 137 would normally be expected to red-shift. The probability is zero.
4) 126 of the 274 exit polls exceeded the margin of error. Only 14 (5%) would normally be expected. The probability is ZERO.
5) 123 of the 126 exit polls that exceeded the margin of error red-shifted to the Republicans. The probability is ZERO.

No exit polls in 19 states

The National Election Pool (NEP) is a consortium of six corporate media giants which funds the pollster Edison Research to do exit polling in the U.S and abroad. The NEP announced that they would not exit poll in 19 states, 16 of which are universally thought of as being solid RED states. Or are they?

In 2008, Obama won exit polls in AK, AL, AZ, GA, NE, SD. He came close to winning in TX, KY, SC, TN, MS. These former RED states may have turned PURPLE. View this worksheet in the model.

The bad news is that the NEP decision to eliminate the polls makes it easier for vote margins to be padded and electoral votes flipped. Without the polls, it is much more difficult to calculate the statistical probabilities of fraud based on exit poll discrepancies. In the 1988-2008 elections, the Democrats led the unadjusted state exit polls by 52-42%, but by just 48-46% in the official recorded vote. This is a mathematically impossible result which proves systemic election fraud.

The good news is that the post-election True Vote Model should find implausible discrepancies in the recorded state and national votes. After all, that is what it was designed to do.

Sensitivity Analysis

The pre-election TVM built in the 2012 Election Model uses alternative scenarios of 2008 voter turnout and defection rates to derive a plausible estimate of the total final share. The returning voter assumptions are based on Obama’s 58% True Vote (a plausible estimate) and his 53% recorded share. The latter scenario results in vote shares that are close to the LV polls.

The sensitivity analysis of alternative turnout and vote share scenarios is an important feature in the model. The model displays the effects of effects of incremental changes in turnout rates and shares of returning voters. The tables display nine scenario combinations of a) Obama and McCain turnout rates and b) Obama/Romney shares of returning Obama and McCain voters. Obama’s vote share, winning margin and popular vote win probability are displayed for each scenario.

Registered and Likely Voters

Historically, RV polls have closely matched the unadjusted exit polls after undecided voters are allocated and have been confirmed by the True Vote Model.

Likely Voter (LV) polls are a subset of Registered Voter polls and are excellent predictors of the recorded vote – which always understate the Democratic True Vote. One month prior to the election, the RV polls are replaced by LVs. An artificial “horse race” develops as the polls invariably tighten.

The Likely Voter Cutoff Model (LVCM) understates the voter turnout of millions of new Democrats, thereby increasing the projected Republican share. Democrats always do better in RV polls than in the LVs. Based on the historical record, the Democratic True Vote share is 4-5% higher than the LV polls indicate. The LVs anticipate the inevitable election fraud reduction in Obama’s estimated 55% True Vote share.

Media pundits and pollsters are paid to project the recorded vote – not the True Vote. The closer they are, the better they look. They never mention the fraud factor which gets them there, but they prepare for it by switching to LV polls.

The disinformation loop is closed when the unadjusted, pristine state and national exit polls are adjusted to match the LV recorded vote prediction.

2004 and 2008 Election Models

The 2004 model matched the unadjusted exit polls. Kerry had 51.7% and 337 electoral votes. But the election was stolen. Kerry had 48.3% recorded. View the 2004 Electoral and popular vote trend

The 2008 model exactly matched Obama’s 365 EV. The National model exactly matched his official recorded 52.9% share; the State model projected 53.1%. His official margin was 9.5 million votes.

Obama had 58.0% in the unadjusted, weighted state exit poll aggregate (83,000 respondents) which exactly matched the post-election True Vote Model. Obama’s 23 million True Vote margin was too big to steal.

The National Exit Poll displayed on mainstream media websites (Fox, CNN, ABC, CBS, NYT, etc.) indicates that Obama had 52.9% – his recorded vote. Unadjusted state and national exit polls are always forced to match the recorded share.

But the media never discussed the fact that Obama had 61% in the unadjusted National Exit Poll (17,836 respondents). View the 2008 Electoral and popular vote trend

This graph summarizes the discrepancies between the 1988-2008 State Exit Polls and the corresponding Recorded Votes.

The True Vote Model

The 2008 True Vote Model (TVM) determined that Obama won in a landslide by 58-40.3%. Based on the historical red-shift, he needs at least a 55% True Vote share to overcome the systemic 5% fraud factor. The TVM was confirmed by the unadjusted state exit poll aggregate: Obama had an identical 58-40.5% margin (83,000 respondents). He won unadjusted National Exit Poll (17,836 respondents) by an even bigger 61-37% margin.

In projecting the national and state vote, a 1.25% annual voter mortality rate is assumed. The TVM uses estimated 2008 voter turnout in 2012 and corresponding 2012 vote shares. The rates are applied to each state in order to derive the national aggregate result.

There are two basic options for estimating returning voters. The default option assumes the unadjusted 2008 exit poll as a basis. The second assumes the recorded vote. It is important to note that the True Vote is never the same as the recorded vote. The 1988-2008 True Vote Model utilizes estimates of previous election returning and new voters and and adjusted state and national exit poll vote shares.

Monte Carlo Simulation

The simulation consists of 500 election trials. The electoral vote win probability is the number of winning election trials divided by 500.

There are two forecast options in the model. The default option uses projections based on the latest pre-election state polls. The second is based on the state True Vote. The fraud factor is the difference between the two.

The projected vote share is the sum of the poll and the undecided voter allocation (UVA). The model uses state vote share projections as input to the Normal Distribution function to determine the state win probability.

In each election trial, a random number (RND) between 0 and 1 is generated for each state and compared to Obama’s state win probability. If RND is greater than the win probability, the Republican wins the state. If RND is less than the win probability, Obama wins the state. The winner of the election trial is the candidate who has at least 270 electoral votes. The process is repeated in 500 election trials.

The Electoral Vote is calculated in three ways.
1. The Snapshot EV is a simple summation of the electoral votes. It could be misleading if close state elections favor one candidate.
2. The Mean EV is the average of the 500 simulated election trials.
3. The Theoretical EV is the product sum of the state electoral votes and corresponding win probabilities. A simulation or meta-analysis is not required to calculate the expected EV.

The Mean EV approaches the Theoretical EV as the number of election trials increase. This is an illustration of the Law of Large Numbers.

Obama’s electoral vote win probability is his winning percentage of 500 simulated election trials.

The national popular vote win probability is calculated using the national aggregate of the the projected vote shares. The national margin of error is 1-2% lower than the MoE of the individual states. That is, if you believe the Law of Large Numbers and convergence to the mean.

The Fraud Factor

The combination of True Vote Model and state poll-based Monte Carlo Simulation enables an analyst to determine if the forecast electoral and popular vote share estimates are plausible. The aggregate state poll shares can be compared to the default TVM.

The TVM can be forced to match the aggregate poll projection by…
- An incremental change in vote shares. A red flag would be raised if the match required that Obama captured 85% of returning Obama voters and Romney had 95% of returning McCain voters (a 10% net defection).

- Adjusting 2008 voter turnout in 2012. For example, if McCain voter turnout is required to be 10-15% higher than Obama’s, that would raise a red flag.

- Setting the returning voter option to the 2008 recorded vote. The implicit assumption is that the 2008 recorded vote was the True Vote. But the 2008 election was highly fraudulent. Therefore, model vote shares will closely match the likely voter polls.

Check the simulated, theoretical and snapshot electoral vote projections and corresponding win probabilities.

In 2004, Election Model forecasts were posted weekly using the latest state and national polls. The model was the first to use Monte Carlo simulation and sensitivity analysis to calculate the probability of winning the electoral vote. The final Nov.1 forecast had Kerry winning 337 electoral votes with 51.8% of the two-party vote, closely matching the unadjusted exit polls.

2004 Election Model Graphs

In the 2006 midterms, the adjusted National Exit Poll was forced to match the House 52-46% Democratic margin. But the 120 Generic Poll Trend Model forecast that the Democrats would have a 56.4% share – exactly matching the unadjusted exit poll.

The 2008 Election Model projection exactly matched Obama’s 365 electoral votes and was within 0.2% of his 52.9% recorded share. He won by 9.5 million votes. But the model understated his True Vote. The forecast was based on final likely voter (LV) polls that had Obama leading by 7%. Registered voter (RV) polls had him up by 13% – even before undecided voters were allocated. The landslide was denied.

The post-election True Vote Model determined that Obama won by 23 million votes with 420 EV. His 58% share matched the unadjusted state exit poll aggregate (83,000 respondents).

Exit pollsters and media pundits have never explained the massive 11% state exit poll margin discrepancy or the impossible 17% National Exit Poll discrepancy. If they did, they would surely claim that the discrepancies were due to reluctant Republican responders. But they will not even try to explain the impossible returning voter adjustments required to force the polls to match the recorded vote in the 1988, 1992, 2004 and 2008 elections.

Track Record: Election Model Forecast; Post-election True Vote Model

2004 Election Model (2-party shares)
Kerry 51.8%, 337 EV (snapshot)
State exit poll aggregate: 51.7%, 337 EV
Recorded Vote: 48.3%, 255 EV
True Vote Model: 53.6%, 364 EV

2008 Election Model
Obama 53.1%, 365.3 EV (simulation mean);
Recorded: 52.9%, 365 EV
State exit poll aggregate: 58.0%, 420 EV
True Vote Model: 58.0%, 420 EV

2012 Election Model
Obama Projected: 51.6% (2-party), 332 EV snapshot; 320.7 expected; 321.6 mean
Adjusted National Exit Poll (recorded): 51.0-47.2%, 332 EV
True Vote Model 56.1%, 391 EV (snapshot); 385 EV (expected)
Unadjusted State Exit Polls: not released
Unadjusted National Exit Poll: not released

Posted by on October 17, 2012 in 2008 Election, 2012 Election

## 10/06/ 2012 Presidential True Vote/Election Fraud Simulation Model:Obama 295 EV; 89% Win Probability

10/06/ 2012 Presidential True Vote/Election Fraud Simulation Model:Obama 295 EV; 89% Win Probability

Richard Charnin
Oct. 6, 2012

The 2012 Presidential True Vote and Monte Carlo Simulation Forecast Model is updated on a daily basis. The projections assume the election is held on the latest poll date. Two forecasting methods are used.

The Monte Carlo Electoral Vote Simulation is based on the latest state polls and currently assumes an equal split of undecided voters. The expected electoral vote is the sum of the products of the state win probabilities and corresponding electoral votes.

The True Vote Model is based on plausible turnout estimates of new and returning 2008 voters and corresponding vote shares.

On Oct.6 Obama led the weighted average of the state pre-election polls by 48.3-45.6% and the national poll average by 48.3-47.2%. Romney’s “bounce” from the debate reduced Obama’s expected electoral vote to 295, down from 325 last week and 342 two weeks ago.

The Gallup poll taken after the debate(10/4-10/6) was tied at 47-47%. Obama led by 50-45% before the debate (9/30-10/2). The Rasmussen poll is tied at 48%.

If the election were held today, the Monte Carlo simulation indicates that Obama would have an 89% probability of winning the electoral vote.

Likely voter (LV) polls discount the pervasive systematic fraud factor. They are traditionally excellent predictors of the recorded vote – which always understate the Democratic True Vote. The LV polls anticipate the inevitable election fraud reduction in Obama’s estimated 55.6% True Vote share and 381 electoral votes. Based on the historical record, Obama’s True Vote share is about 4-5% higher than the latest polls indicate. It is a certainty that he will lose millions of votes on Election Day to fraud.

The only question is: Will Obama be able to overcome the systemic fraud factor?

Forecast Summary

The source of the polling data is the Real Clear Politics (RCP) website. The simulation uses the latest state polls. Recorded 2008 vote shares are used for states which have not yet been polled.

2012 Presidential True Vote and Monte Carlo Simulation Forecast Model

10/06/2012
UVA = undecided voter allocation = 50/50%
True Vote Model Obama Romney
True Vote...... 55.6% 44.4% (see model)
Expected EV.... 381 157 EV = sum(state win prob (i) * EV(i)), i=1,51
Snapshot EV.... 391 147 Sum of state EV
EV Win Prob.... 100% 0%

State Polls
Average........ 48.3% 45.6% (state vote-weighted average)
Projection..... 51.4% 48.6% (RCP Polls + UVA)
Pop. Win Prob.. 81% 19% (3.0% MoE assumed for calculation)
Expected EV.... 295.3 242.7 EV = sum(state win prob(i) * EV(i)), i=1,51
Snapshot EV.... 263 275 Sum of winning state electoral votes

National Polls
Average........ 48.3% 47.2% (RCP poll average)
Projection..... 50.6% 49.4% (RCP polls + UVA)
Pop. Win Prob.. 71% 29% (2.0% MoE assumed for calculation)
Gallup......... 47% 47% (1387 RV, 3.0% MoE)
Rasmussen...... 48% 48% (1500 LV, 3.0% MoE)

Monte Carlo Simulation (500 Election trials)
Projection..... 51.4% 48.6% (RCP state polls + UVA)
Mean EV........ 296.3 241.7 (average of 500 election trials)
Maximum EV..... 348 190
Minimum EV..... 240 298
EV Win Prob.... 89% 11% (445 wins/500 election trials)

2004 and 2008 Election Models

The 2004 model matched the unadjusted exit polls. Kerry had 51.7% and 337 electoral votes. But the election was stolen. Kerry had 48.3% recorded. View the 2004 Electoral and popular vote trend

The 2008 model exactly matched Obama’s 365 EV. The National model exactly matched his official recorded 52.9% share; the State model projected 53.1%. His official margin was 9.5 million votes.

Obama had 58.0% in the unadjusted, weighted state exit poll aggregate (83,000 respondents) which exactly matched the post-election True Vote Model. Obama’s 23 million true vote vote margin was too big to steal.

The National Exit Poll displayed on mainstream media websites (Fox, CNN, ABC, CBS, NYT, etc.) indicates that Obama had 52.9% – his recorded vote. Unadjusted state and national exit polls are always forced to match the recorded share.

But the media never discussed the fact that Obama had 61% in the unadjusted National Exit Poll (17,836 respondents). View the 2008 Electoral and popular vote trend

In the six presidential elections from 1988-2008, the Democrats won the average recorded vote by 48-46%. But they led both state and national exit polls by 52-42%. There were approximately 375,000 respondents in the 274 state polls and 90,000 respondents in the six national polls. Overall, an extremely low margin of error.

This graph summarizes the discrepancies between the1988-2008 State Exit Polls vs. the corresponding Recorded Votes.

The True Vote Model

The 2008 True Vote Model (TVM) determined that Obama won in a landslide by 58-40.3%. Based on the historical red-shift, he needs at least a 55% True Vote share to overcome the systemic 5% fraud factor. The TVM was confirmed by the unadjusted state exit poll aggregate: Obama had an identical 58-40.5% margin (83,000 respondents). He won unadjusted National Exit Poll (17,836 respondents) by an even bigger 61-37% margin.

In projecting the national vote, the TVM uses estimated returning 2008 voter turnout in 2012 and corresponding 2012 vote shares. The rates are applied to each state in order to derive the national aggregate result.

A 1.25% annual voter mortality rate is assumed. There are two options for estimating returning voters. The default option assumes that 2008 voters return in proportion to the unadjusted 2008 exit poll aggregate (Obama won by 58-40.5%).

It is important to note that the True Vote is never the same as the recorded vote. The 1988-2008 True Vote Model utilizes estimates of previous election returning and new voters and and adjusted state and national exit poll vote shares.

Sensitivity analysis

The TVM displays the effects of effects of incremental changes in turnout rates and shares of returning voters. Three tables each display nine scenario combinations of a) Obama and McCain turnout rates and b) Obama/Romney shares of returning Obama and McCain voters. Obama’s vote share, winning margin and popular vote win probability are displayed for each scenario.

Monte Carlo Simulation

The simulation consists of 500 election trials. The electoral vote win probability is the number of winning election trials divided by 500.

There are two forecast options in the model. The default option uses projections based on the latest pre-election state polls. The second is based on the state True Vote. The fraud factor is the difference between the two.

The projected vote share is the sum of the poll and the undecided voter allocation (UVA). The model uses state vote share projections as input to the Normal Distribution function to determine the state win probability.

In each election trial, a random number (RND) between 0 and 1 is generated for each state and compared to Obama’s state win probability. If RND is greater than the win probability, the Republican wins the state. If RND is less than the win probability, Obama wins the state. The winner of the election trial is the candidate who has at least 270 electoral votes. The process is repeated in 500 election trials.

The Electoral Vote is calculated in three ways.
1. The Snapshot EV is a simple summation of the electoral votes. It could be misleading if close state elections favor one candidate.
2. The Mean EV is the average of the 500 simulated election trials.
3. The Theoretical EV is the product sum of the state electoral votes and corresponding win probabilities. A simulation or meta-analysis is not required to calculate the expected EV.

The Mean EV approaches the Theoretical EV as the number of election trials increase. This is an illustration of the Law of Large Numbers.

Obama’s electoral vote win probability is his winning percentage of 500 simulated election trials.

The national popular vote win probability is calculated using the normal distribution using the national aggregate of the the projected vote shares. The national aggregate margin of error is 1-2% lower than the average MoE of the individual states. That is, if you believe the Law of Large Numbers and convergence to the mean.

The Fraud Factor

Election fraud reduced the 1988-2008 Democratic presidential unadjusted exit poll margin from 52-42% to 48-46%. View the 1988-2008 Unadjusted State and National Exit Poll Database

The combination of True Vote Model and state poll-based Monte Carlo Simulation enables an analyst to determine if the forecast electoral and popular vote share estimates are plausible. The aggregate state poll shares can be compared to the default TVM.

The TVM can be forced to match the aggregate poll projection by…
- Adjusting vote shares by an incremental change. A red flag would be raised if the match required, if for example Obama captured 85% of returning Obama voters and Romney had 95% of returning McCain voters (a 10% net defection).

- Adjusting 2008 voter turnout in 2012. For example, if McCain voter turnout is required to be 10-15% higher than Obama’s, that would raise a red flag.

- Setting the returning voter option to the 2008 recorded vote. The implicit assumption is that the 2008 recorded vote was the True Vote. But the 2008 election was highly fraudulent. Therefore, model vote shares will closely match the likely voter polls.

Check the simulated, theoretical and snapshot electoral vote projections and corresponding win probabilities.

Election Model Projections

In 2004, Election Model forecasts were posted weekly using the latest state and national polls. The model was the first to use Monte Carlo simulation and sensitivity analysis to calculate the probability of winning the electoral vote. The final Nov.1 forecast had Kerry winning 337 electoral votes with 51.8% of the two-party vote, closely matching the unadjusted exit polls.

2004 Election Model Graphs

In the 2006 midterms, the adjusted National Exit Poll was forced to match the House 52-46% Democratic margin. But the 120 Generic Poll Trend Model forecast that the Democrats would have a 56.4% share – exactly matching the unadjusted exit poll.

The 2008 Election Model projection exactly matched Obama’s 365 electoral votes and was within 0.2% of his 52.9% recorded share. He won by 9.5 million votes. But the model understated his True Vote. The forecast was based on final likely voter (LV) polls that had Obama leading by 7%. Registered voter (RV) polls had him up by 13% – even before undecided voters were allocated. The landslide was denied.

The post-election True Vote Model determined that Obama won by 23 million votes with 420 EV. His 58% share matched the unadjusted state exit poll aggregate (83,000 respondents).

Exit pollsters and media pundits have never explained the massive 11% state exit poll margin discrepancy or the impossible 17% National Exit Poll discrepancy. If they did, they would surely claim that the discrepancies were due to reluctant Republican responders. But they will not even try to explain the impossible returning voter adjustments required to force the polls to match the recorded vote in the 1988, 1992, 2004 and 2008 elections.

2008 Election Model Graphs

Pre-election RV and LV Polls

Virtually all early pre-election polls are of Registered Voters (RV). The Rasmussen tracking poll is an exception, using a Likely Voter (LV) subset of the full RV sample. Rasmussen is an admitted GOP pollster.

One month prior to the election, pollsters replace the full RV sample polls with LV sub-samples, helping to promote an artificial “horse race” as the poll shares invariably tighten. The Likely Voter Cutoff Model (LVCM) effectively understates the voter turnout of millions of new Democrats, increasing the projected Republican share. Democrats always do better in the full RV polls than the LVs.

Media pundits and pollsters are paid to project the recorded vote – not the True Vote. The closer they are, the better they look. But they never mention the fraud factor which gets them there. They prepare for it by switching to LV polls which are usually excellent predictors of the recorded vote.

Historically, RV polls have closely matched the unadjusted exit polls after undecided voters are allocated and have been confirmed by the True Vote Model. The disinformation loop is closed when the unadjusted, pristine state and national exit polls are adjusted to match the LV recorded vote prediction.

In pre-election and exit polls:
1) The Likely Voter Cutoff Model eliminates newly registered Democrats from the LV sub-sample. Kerry had 57-61% of new voters; Obama had 72%.
2) Exit poll precincts are partially selected based on the previous election recorded vote.
3) In the 1988-2008 presidential elections, 226 of 274 exit polls red-shifted" to the Republicans. Only about 137 would normally be expected to red-shift. The probability is zero.
4) 126 of the 274 exit polls exceeded the margin of error. Only 14 (5%) would normally be expected. The probability is ZERO.
5) 123 of the 126 exit polls that exceeded the margin of error red-shifted to the Republicans. The probability is ZERO.

Posted by on October 7, 2012 in 2012 Election

## 9/26/ 2012 Presidential True Vote/Election Fraud Simulation Model:Obama 342 EV; 100% Win Probability

9/26/ 2012 Presidential True Vote/Election Fraud Simulation Model:Obama 342 EV; 100% Win Probability

Richard Charnin
Sept. 26, 2012

The 2012 Presidential True Vote and Monte Carlo Simulation Forecast Model uses two forecast methods. The Monte Carlo Electoral Vote Simulation is based on the latest state polls. The True Vote Model calculates vote shares based on a feasible estimate of new and returning 2008 voters and corresponding vote shares. The model is updated periodically for the latest state and national polls. The projections assume the election is held on the latest poll date.

Obama increased his expected electoral vote from 320 to 342 by gaining the lead in the North Carolina and Iowa polls. The expected electoral vote is based on the state win probabilities. If the election were held today, the Monte Carlo electoral vote simulation indicates that Obama would have a 100% probability of winning as he won all 500 election simulation trials. But there are six weeks to so.

Obama’s 49.2-44.3% margin in the weighted state pre-election polls is very close to his 48.9-44.9% lead in the RCP national average – a joint confirmation. His lead increased to 50-44% in the Gallup RV tracking poll (2% MoE, 3050 sample).

2004 and 2008 Election Models

The 2004 model matched the unadjusted exit polls. Kerry had 51.7% and 337 electoral votes. But the election was stolen. Kerry had 48.3% recorded. View the 2004 Electoral and popular vote trend

The 2008 model exactly matched Obama’s 365 EV. The national model exactly matched his official recorded 52.9% share; the state model projected 53.1%. His official margin was 9.5 million votes. But Obama had 58.0% in the unadjusted, weighted state exit poll aggregate (83,000 respondents) which exactly matched the True Vote Model. His 23 million vote margin was too big to steal. This is the whopper no one in the media talks about: Obama had 61% in the unadjusted National Exit Poll (17,836 respondents). View the 2008 Electoral and popular vote trend

Forecast Summary

Approximately 7% of voters are undecided and may hold the key to the election. I suspect they are mostly Democrats disillusioned with Obama but scared by Romney and Ryan. The model currently assumes an equal split of the undecided vote. If undecided voters break for Obama, he will be in a commanding position to win re-election.

The Likely Voter (LV) polls anticipate the inevitable election fraud reduction in Obama’s estimated 56.3% True Vote share and 402 electoral votes.

The source of the polling data is the Real Clear Politics (RCP) website. The simulation uses the latest state polls. Recorded 2008 vote shares are used for states which have not yet been polled.

9/26/2012
UVA = undecided voter allocation = 50/50%
True Vote Model Obama Romney
True Vote...... 56.3% 43.7% (see model)
Expected EV.... 402 136 EV = sum(state win prob (i) * EV(i)), i=1,51
Snapshot EV.... 410 128 Sum of state EV
EV Win Prob.... 100% 0%

State Polls
Average........ 49.2% 44.3% (state vote-weighted average)
Projection..... 52.4% 47.6% (RCP Polls + UVA)
Pop. Win Prob.. 94.5% 5.5% (3.0% MoE)
Expected EV.... 342.4 195.6 EV = sum(state win prob(i) * EV(i)), i=1,51
Snapshot EV.... 343 195 Sum of winning state electoral votes

National Polls
Average........ 48.9% 44.9% (RCP poll average)
Projection..... 52.0% 48.0% (RCP polls + UVA)
Pop. Win Prob.. 97.5% 2.5% (2.0% MoE)
Gallup......... 50% 44% (3050 RV sample, 2.0% MoE)
Rasmussen...... 46% 46% (1500 LV sample, 3.0% MoE)

Monte Carlo Simulation (500 Election trials)
Projection..... 52.4% 47.6% (RCP state polls + UVA)
Mean EV........ 341.8 196.2 (average of 500 election trials)
Maximum EV..... 375 163
Minimum EV..... 309 229
EV Win Prob.... 100% 0% (500 wins/500 election trials)

Polling samples are based on prior election recorded votes – not the previous True Vote or unadjusted exit poll. Likely voter (LV) polls discount the pervasive systematic fraud factor. They are traditionally excellent predictors of the recorded vote – which always understate the Democratic True Vote.

In the six presidential elections from 1988-2008, the Democrats won the average recorded vote by 48-46%. But they led both state and national exit polls by 52-42%. There were approximately 375,000 respondents in the 274 state polls and 90,000 respondents in the six national polls. Overall, an extremely low margin of error.

This graph summarizes the discrepancies between the1988-2008 State Exit Polls vs. the corresponding Recorded Votes

Based on the historical record, Obama’s True Vote share is about 4-5% higher than the latest polls indicate. It is a certainty that he will lose millions of votes on Election Day to fraud. The only question is: Will he overcome the systemic fraud factor? As of today, it appears he will.

The 2008 True Vote Model (TVM) determined that Obama won in a landslide by 58-40.3%. Based on the historical red-shift, he needs at least a 55% True Vote share to overcome the systemic 5% fraud factor. The TVM was confirmed by the unadjusted state exit poll aggregate: Obama had an identical 58-40.5% margin (83,000 respondents). He won unadjusted National Exit Poll (17,836 respondents) by an even bigger 61-37% margin.

The National Exit Poll displayed on mainstream media websites (Fox, CNN, ABC, CBS, NYT, etc.) indicate that Obama had 52.9% – his recorded vote. Unadjusted state and national exit polls are always forced to match the recorded share.

The True Vote Model

In projecting the national vote, the required input to the TVM are returning 2008 voter turnout rates in 2012 and estimated 2012 vote shares. The rates are applied to each state in order to derive the national aggregate turnout . A 1.25% annual voter mortality rate is assumed. There are two options for estimating returning voters. The default option assumes that 2008 voters return in proportion to the unadjusted 2008 exit poll aggregate (Obama won by 58-40.5%). In this scenario, Obama wins by 55-45% with 380 EV and has a 100% EV win probability.

It is important to note that the True Vote is never the same as the recorded vote. The 1988-2008 True Vote Model utilizes estimates of previous election returning and new voters and and adjusted state and national exit poll vote shares.

Sensitivity analysis

The TVM displays the effects of effects of incremental changes in turnout rates and shares of returning voters. Three tables are generated consisting of nine scenario combinations of a) Obama and McCain turnout rates and b) the Obama/Romney shares of returning Obama and McCain voters. The output tables display resulting vote shares, vote margins and popular vote win probabilities.

Monte Carlo Simulation: 500 election trials

There are two forecast options in the simulation model. The default option uses projections based on the latest pre-election state polls. The second uses projections based on the state True Vote. The difference between the two approximates the fraud factor.

The projected vote share is the sum of the poll share and the undecided voter allocation (UVA). The model uses state vote share projections as input to the Normal Distribution function to determine the state win probability.

The simulation consists of 500 election trials. The electoral vote win probability is the number of winning election trials divided by 500.

In each election trial, a random number (RND) between 0 and 1 is generated for each state and compared to Obama’s state win probability. If RND is greater than the win probability, the Republican wins the state. If RND is less than the win probability, Obama wins the state. The winner of the election trial is the candidate who has at least 270 electoral votes. The process is repeated in 500 election trials.

2008 State Exit Poll and recorded vote data is displayed in the ‘2008‘ worksheet. The latest state polls are listed in the ‘Trend/Chart” worksheet, The data is displayed graphically in the ‘PollChart’ worksheet. A histogram of the Monte Carlo Simulation (500 trials) is displayed in the ‘ObamaEVChart’ worksheet.

The Electoral Vote is calculated in three ways.
1. The Snapshot EV is a simple summation of the state electoral votes. It could be misleading since there may be several very close elections which favor one candidate.
2. The Mean EV is the average electoral vote of the 500 simulated elections.
3. The Theoretical (expected) EV is the product sum of all state electoral votes and corresponding win probabilities. A simulation or meta-analysis is not required to calculate the expected EV.

The Mean EV approaches the Theoretical EV as the number of election trials increase. This is an illustration of the Law of Large Numbers.

Obama’s electoral vote win probability is his winning percentage of 500 simulated election trials.

The national popular vote win probability is calculated using the normal distribution using the national aggregate of the the projected vote shares. The national aggregate margin of error is 1-2% lower than the average MoE of the individual states. That is, if you believe the Law of Large Numbers and convergence to the mean.

The Fraud Factor

Election fraud reduced the 1988-2008 Democratic presidential unadjusted exit poll margin from 52-42% to 48-46%. View the 1988-2008 Unadjusted State and National Exit Poll Database

The combination of True Vote Model and state poll-based Monte Carlo Simulation enables the analyst to determine if the electoral and popular vote share estimates are plausible. The aggregate state poll shares can be compared to the default TVM.

The TVM can be forced to match the aggregate poll projection by…
- Adjusting vote shares by an incremental change. A red flag would be raised if the match required, if for example Obama captured 85% of returning Obama voters and Romney had 95% of returning McCain voters (a 10% net defection).

- Adjusting 2008 voter turnout in 2012. For example, if McCain voter turnout is required to be 10-15% higher than Obama’s, that would raise a red flag.

- Setting the returning voter option to the 2008 recorded vote. The implicit assumption is that the 2008 recorded vote was the True Vote. But the 2008 election was highly fraudulent. Therefore, model vote shares will closely match the likely voter polls.

Check the simulated, theoretical and snapshot electoral vote projections and corresponding win probabilities.

Election Model Projections: 2004-2010

In 2004, I created the Election Model , and posted weekly forecasts using the latest state and national polls. The model was the first one to use Monte Carlo simulation and sensitivity analysis to calculate the probability of winning the electoral vote. The final Nov.1 forecast had Kerry winning 337 electoral votes with 51.8% of the two-party vote. The forecast closely matched the unadjusted exit polls.

2004 Election Model Graphs

In 2006, the adjusted National Exit Poll indicated that the Democrats won the House by a 52-46% vote share. But the 120 Generic Poll Forecasting Regression Model indicated that they would have 56.4% – exactly matching the unadjusted exit poll.

The 2008 Election Model projection exactly matched Obama’s 365 electoral votes and was within 0.2% of his 52.9% recorded share. He won by 9.5 million votes. But the model understated his True Vote. The forecast was based on final likely voter (LV) polls that had Obama leading by 7%. Registered voter (RV) polls had him up by 13% – before undecided voter allocation. The landslide was denied. The post-election True Vote Model determined that Obama won by 23 million votes with 420 EV. His 58% share matched the unadjusted state exit poll aggregate (83,000 respondents).

2008 Election Model Graphs

Exit pollsters and media pundits have never explained the massive 11% state exit poll margin discrepancy or the impossible 17% National Exit Poll discrepancy. If they did, they would surely claim that the discrepancies were due to reluctant Republican responders. But they will not even try to explain the impossible returning voter adjustments required to force the polls to match the recorded vote in the 1988, 1992, 2004 and 2008 elections.

Pre-election RV and LV Polls

Virtually all early pre-election polls are of Registered Voters (RV). An exception is the Rasmussen poll. It uses the Likely Voter (LV) subset of the full RV sample. Rasmussen is an admitted GOP pollster.

One month prior to the election, pollsters replace the full RV sample polls with LV subsamples. The RV polls are transformed to LVs to promote an artificial “horse race” – and the poll shares invariably tighten. The Likely Voter Cutoff Model (LVCM) effectively understates the turnout of millions of new Democratic voters – and therefore increases the projected Republican share. Democrats always do better in RV polls than in the LVs.

Media pundits and pollsters are paid to project the recorded vote – not the True Vote. And they are usually right. The closer they are, the better they look. They expect there will be fraud, so they prepare the public for it by switching to LV polls which are usually excellent predictors of the recorded vote. But they never mention the fraud factor which gets them there.

Historically, RV polls have closely matched the unadjusted exit polls after undecided voters were allocated< They have been confirmed by the True Vote Model. The loop is closed when unadjusted, pristine state and national exit polls are adjusted to match the LV recorded vote prediction.

In pre-election and exit polls:
1) The Likely Voter Cutoff Model eliminates newly registered Democrats from the LV sub-sample. Kerry had 57-61% of new voters; Obama had 72%.
2) Exit poll precincts are partially selected based on the previous election recorded vote.
3) In the 1988-2008 presidential elections, 226 of 274 exit polls red-shifted" to the Republicans. Only about 137 would normally be expected to red-shift. The probability is zero.
4) 126 of the 274 exit polls exceeded the margin of error. Only 14 (5%) would normally be expected. The probability is ZERO.
5) 123 of the 126 exit polls that exceeded the margin of error red-shifted to the Republicans. The probability is ZERO.

Election Model Forecast; Post-election True Vote Model

2004 (2-party vote shares)
Model: Kerry 51.8%, 337 EV (snapshot)
State exit poll aggregate: 51.7%, 337 EV
Recorded Vote: 48.3%, 255 EV
True Vote Model: 53.6%, 364 EV

2008
Model: Obama 53.1%, 365.3 EV (simulation mean);
Recorded: 52.9%, 365 EV
State exit poll aggregate: 58.0%, 420 EV
True Vote Model: 58.0%, 420 EV

2012 (2-party state exit poll aggregate shares)
Model: Obama 51.6%, 332 EV (Snapshot)
Recorded : 51.6%, 332 EV
True Vote 55.2%, 380 EV

Posted by on September 26, 2012 in 2012 Election

## 2012 Presidential True Vote and Monte Carlo Simulation Forecast Model

2012 Presidential True Vote and Monte Carlo Simulation Forecast Model

Richard Charnin
June 22,2012

The model will be run on a periodic basis up to Election Day.
a) The True Vote Model is based on estimated turnout and vote share assumptions
b) The Simulation model is based on the latest state and national polls.

NOV.5 UPDATE: VIEW FINAL FORECAST HERE

It is important to note that the True Vote is never the same as the recorded vote. In 2008, Obama had 58% in the unadjusted state exit poll aggregate and 58% in the 1988-2008 True Vote Model. But his recorded vote share was just 52.9%. Therefore, assuming the same 5% red-shift differential, Obama needs at least a 55% True Vote share to win the popular vote.

Rasmussen is a GOP pollster who provides a Likely Voter (LV) subset of the total number of Registered Voters (RV). The majority of registered voters excluded by the Likely Voter Cutoff Model are Democrats.

Election Model Projections: 2004-2010

The 2004 Election Model weekly projections started in July and were based on the latest state and national polls. The model was the first to use Monte Carlo Simulation and sensitivity analysis to calculate the probability of winning the electoral vote. It projected Kerry winning 337 electoral votes with 51.8% of the two-party vote, closely matching the unadjusted National Exit Poll (51.7%). The election was stolen.

The 2006 House Trend Forecast Model was based on 120 Generic polls. It projected that the Democrats would capture 56.43% of the vote and was virtually identical to the unadjusted National Exit Poll (56.37%). The NEP was forced to match the recorded 52-46% vote share. The landslide was denied. Election fraud cost the Democrats 15-20 House seats.

The 2008 Election Model projection was published weekly. The final projection exactly matched Obama’s 365 electoral votes and was within 0.2% of his 52.9% share (a 9.5 million margin). But the model understated his True Vote. The forecast was based on likely voter (LV) polls that had Obama leading by 7%. Registered voter (RV) polls had him up by 13% – before undecided voter allocation. The True Vote Model determined that Obama won by over 22 million votes with 420 EV. His 58% share was within 0.1% of the unadjusted state exit poll aggregate (83,000 respondents). The landslide was denied.

The 2010 Election Forecast Model predicted a 234-201 GOP House based on the final 30 likely voter (LV) Generic polls (the GOP led by 48.7-41.9%). It predicted a 221-214 GOP House based on the final 19 registered voter (RV) polls (the GOP led by 45.1-44.4%). The Final National Exit Poll was a near match to the LV pre-election poll average. The Democratic margin was 6.1% higher in the RV polls than the LVs.

The model predicted a 50-48 Democratic Senate based on 37 LV polls in which the GOP led by 48.1-43.5%. It predicted a 53-45 Democratic margin based on a combination of 18 RV and 19 LV polls in which they led by 45.2-44.6%, a 5.2% increase in margin.

There were no RV polls in the realclearpolitics.com final polling averages. CNN/Time provided RV and LV polling data for 18 Senate races. The Democrats led the RV polls in 11 states (49.2-40.6%) and the LV subset in 8 (46.6-45.8%), an 8% difference in margin.

Take the Election Fraud Quiz.

Election Model Forecast; Post-election True Vote Model

2004 (2-party vote shares)
Model: Kerry 51.8%, 337 EV (snapshot)
State exit poll aggregate: 51.7%, 337 EV
Recorded Vote: 48.3%, 255 EV
True Vote Model: 53.6%, 364 EV

2008
Model: Obama 53.1%, 365.3 EV (simulation mean);
Recorded: 52.9%, 365 EV
State exit poll aggregate: 58.0%, 420 EV
True Vote Model: 58.0%, 420 EV

2012 (2-party state exit poll aggregate shares)
Model: Obama 51.6%, 332 EV (Snapshot)
Recorded : 51.6%, 332 EV
True Vote 55.2%, 380 EV

## 9/19/ 2012 Presidential True Vote and Election Fraud Simulation Model: Obama 320 EV; 100% Win Probability

9/19/ 2012 Presidential True Vote/Election Fraud Simulation Model:Obama 320 EV; 100% Win Probability

Richard Charnin
Sept.19, 2012

The analysis assumes the election is held on the latest poll date:
2012 Presidential True Vote and Monte Carlo Simulation Forecast Model

Forecast Summary
Obama has jumped out to a commanding 49-44% lead in the battleground state polls. He has 320 expected electoral votes based on his state win probabilities. The 500 trial Monte Carlo simulation indicates that if the election were held today, he would have a 100% win probability (he won all 500 election simulation trials). But it’s still too early to project him a winner.

The 7% of voters who are still undecided hold the key to the election. The model currently assumes an equal split of the undecided vote. I suspect they are mostly Democrats disillusioned with Obama but scared by Romney and Ryan. If the undecided voters break for Obama, he will be in a commanding position to win re-election. But look for an October surprise.

Obama needs at least a 55% True Vote to overcome the Fraud factor. He has held a steady 4% lead in the state polls since April. The polls are anticipating the inevitable 5% reduction in Obama’s True Vote. Immediately after the Democratic Convention, Obama moved into a 5% lead in the Gallup (RV) and Rasmussen (LV) national tracking polls, but the polls are tied once again.

The forecast model is a combination of a) a pre-election Monte Carlo Simulation Model, which is based on the latest state polls, and b) the True Vote Model, based on a feasible estimate of new and returning 2008 voters and corresponding estimated vote shares. The model will be updated periodically for the latest state and national polls.

The source of the polling data is the Real Clear Politics (RCP) website. The simulation uses the latest state polls. Recorded 2008 vote shares are used for states which have not yet been polled.

9/19/2012
Model.......... Obama Romney
True Vote...... 55.25% 44.75%
Expected EV.... 379.64 158.36
Snapshot EV.... 380 158
EV Win Prob.... 99.97% 0.03%

State Polls
Average........ 49.3% 44.4%
Projection..... 52.5% 47.5%
Pop. Win Prob.. 94.8% 5.2%
Expected EV.... 320.2 217.8
Snapshot EV.... 322 216

National Polls
Average....... 48.20% 45.30%
Projection.... 51.45% 48.55%
Pop. Win Prob.. 92.2% 7.8%
Gallup......... 47.0% 46.0%
Rasmussen...... 46.0% 47.0%

Simulation
Projection..... 52.5% 47.5%
Mean EV........ 320.4 217.6
Max EV......... 351 187
Min EV......... 278 260
EV Win Prob.... 100.0% 0.0%

The 2008 True Vote Model (TVM) determined that Obama won in a landslide by 58-40.3%. Based on the historical red-shift, he needs at least a 55% True Vote share to overcome the systemic 5% fraud factor. The TVM was confirmed by the unadjusted state exit poll aggregate: Obama had an identical 58-40.5% margin (76,000 respondents). He won unadjusted National Exit Poll (17,836 respondents) by an even bigger 61-37% margin.

The National Exit Poll displayed on mainstream media websites (Fox, CNN, ABC, CBS, NYT, etc.) indicate that Obama had 52.9%. As usual, the unadjusted state and national exit polls were forced to match the recorded share.

The True Vote Model

In projecting the national vote, the required input to the TVM are returning 2008 voter turnout rates in 2012 and estimated 2012 vote shares. The rates are applied to each state in order to derive the national aggregate turnout . A 1.25% annual voter mortality rate is assumed. There are two options for estimating returning voters. The default option assumes that 2008 voters return in proportion to the unadjusted 2008 exit poll aggregate (Obama won by 58-40.5%). In this scenario, Obama wins by 55-45% with 380 EV and has a 100% EV win probability.

It is important to note that the True Vote is never the same as the recorded vote. The 1988-2008 True Vote Model utilizes estimates of previous election returning and new voters and and adjusted state and national exit poll vote shares.

Sensitivity analysis

The TVM displays the effects of effects of incremental changes in turnout rates and shares of returning voters. Three tables are generated consisting of nine scenario combinations of a) Obama and McCain turnout rates and b) the Obama/Romney shares of returning Obama and McCain voters. The output tables display resulting vote shares, vote margins and popular vote win probabilities.

Monte Carlo Simulation: 500 election trials
There are two options for the simulation model. Both should be used and the results compared. The default option uses the TVM projected state vote shares. The second option uses projections based on the latest pre-election state polls.

The projected vote share is the sum of the poll share and the undecided voter allocation (UVA). The model uses state vote share projections as input to the Normal Distribution function to determine the state win probability.

The simulation consists of 500 election trials. The electoral vote win probability is the number of winning election trials divided by 500.

In each election trial, a random number (RND) between 0 and 1 is generated for each state and compared to Obama’s state win probability. If RND is greater than the win probability, the Republican wins the state. If RND is less than the win probability, Obama wins the state. The winner of the election trial is the candidate who has at least 270 electoral votes. The process is repeated in 500 election trials.

2008 State Exit Poll and recorded vote data is displayed in the ‘2008‘ worksheet. The latest state polls are listed in the ‘Polls” worksheet which will be used for trend analysis. The data is displayed graphically in the ‘PollChart’ worksheet. A histogram of the Monte Carlo Simulation (500 trials) is displayed in the ‘ObamaEVChart’ worksheet.

The Electoral Vote is calculated in three ways.

1. The snapshot EV is a simple summation of the state electoral votes. It could be misleading since there may be several very close elections which go one way.
2. The Theoretical EV is the product sum of the state electoral votes and win probabilities. A simulation or meta-analysis is not required to calculate the expected EV.
3. The Mean EV is the average electoral vote in the 500 simulated elections.

The Mean EV will be close to the Theoretical EV, illustrating the Law of Large Numbers. The snapshot EV will likely differ slightly from the Theoretical EV, depending on the number of state election projections that fall within the margin of error.

Obama’s electoral vote win probability is the percentage of 500 simulated election trials that he won.

The national popular vote win probability is calculated using the normal distribution using the national aggregate of the the projected vote shares. The national aggregate margin of error is 1-2% lower than the average MoE of the individual states. That is, if you believe the Law of Large Numbers and convergence to the mean.

The Fraud Factor

Election fraud reduced the 1988-2008 Democratic presidential unadjusted exit poll margin from 52-42% to 48-46% recorded. View the 1988-2008 Unadjusted State and National Exit Poll Database

The combination of True Vote Model and state poll-based Monte Carlo Simulation enables the analyst to determine if the electoral and popular vote share estimates are plausible. The aggregate state poll shares can be compared to the default TVM.

The TVM can be forced to match the aggregate poll projection by…
- Adjusting the vote shares by entering an incremental adjustment in the designated input cell. A red flag would be raised if the match required that Obama capture 85% of returning Obama voters while Romney gets 95% of returning McCain voters (a 10% net defection).

- Adjusting 2008 voter turnout in 2012 in order to force a match to the aggregate projected poll shares. For example, if McCain voter turnout is required to be 10-15% higher than Obama’s, that would also raise a red flag.

- Setting the returning voter option to assume the 2008 recorded vote. This is an implicit assumption that the 2008 recorded vote was the True Vote. Of course, the 2008 election was highly fraudulent, but this is what the election forecasters effectively do: they ignore the fraud factor. The resulting model vote shares would then closely match the LV polls and would suggest that Romney has a good chance of winning a rigged election.

In any case, check the simulated, theoretical and snapshot electoral vote projections and the corresponding win probabilities.

Election Model Projections: 2004-2010

In 2004, I created the Election Model , and posted weekly forecasts using the latest state and national polls. The model was the first to use Monte Carlo simulation and sensitivity analysis to calculate the probability of winning the electoral vote. The final Nov.1 forecast had Kerry winning 337 electoral votes with 51.8% of the two-party vote. The forecast closely matched the unadjusted exit polls.

In 2006, the adjusted National Exit Poll indicated that the Democrats won the House by a 52-46% vote share. My 120 Generic Poll Forecasting Regression Model indicated that the Democrats would capture 56.43% of the vote. It was within 0.06% of the unadjusted exit poll.

The 2008 Election Model projection exactly matched Obama’s 365 electoral votes and was within 0.2% of his 52.9% recorded share. He won by 9.5 million votes. But the model understated his True Vote. The forecast was based on final likely voter (LV) polls that had Obama leading by 7%. Registered voter (RV) polls had him up by 13% – before undecided voter allocation. The landslide was denied. The post-election True Vote Model determined that Obama won by 23 million votes with 420 EV. His 58% share matched the unadjusted state exit poll aggregate (83,000 respondents).

The exit pollsters have never explained the massive 11% state exit poll margin discrepancy, much less the impossible National Exit Poll 17% discrepancy. If they do, they will surely claim that the discrepancies were due to flawed polling samples. Nor can they explain the rationale of using impossible returning voter weight adjustments to force the exit polls to match the recorded votes in the 1988, 1992, 2004 and 2008 presidential elections.

Pre-election RV and LV Polls

Virtually all early pre-election polls are of Registered Voters (RV). An exception is the Rasmussen poll. It uses the Likely Voter (LV) subset of the full RV sample. Rasmussen is an admitted GOP pollster.

One month prior to the election, the pollsters replace the full RV sample polls with LV subsamples. The RV polls are transformed to LVs to promote an artificial “horse race” – and the poll shares invariably tighten. The Likely Voter Cutoff Model (LVCM) effectively understates the turnout of millions of new Democratic voters – and therefore increases the projected Republican share. Democrats always do better in RV polls than in the LVs.

Media pundits and pollsters are paid to project the recorded vote – not the True Vote. And they are usually right. The closer they are, the better they look. They expect there will be fraud, so they prepare the public for it by switching to LV polls which are usually excellent predictors of the recorded vote. But they never mention the fraud factor which gets them there.

Historically, RV polls have closely matched the unadjusted exit polls – after undecided voters are allocated. They have also been confirmed by the True Vote Model. The loop is closed when implausible/impossible exit polls are forced to match bogus recorded votes that were predicted by biased LV pre-election polls.

Election Model Forecast; Post-election True Vote Model

2004 (2-party vote shares)
Model: Kerry 51.8%, 337 EV (snapshot)
State exit poll aggregate: 51.7%, 337 EV
Recorded Vote: 48.3%, 255 EV
True Vote Model: 53.6%, 364 EV

2008
Model: Obama 53.1%, 365.3 EV (simulation mean);
Recorded: 52.9%, 365 EV
State exit poll aggregate: 58.0%, 420 EV
True Vote Model: 58.0%, 420 EV

2012 (2-party state exit poll aggregate shares)
Model: Obama 51.6%, 332 EV (Snapshot)
Recorded : 51.6%, 332 EV
True Vote 55.2%, 380 EV

Posted by on April 26, 2012 in 2012 Election, True Vote Models

## Monte Carlo Simulation: Election Forecasting and Exit Poll Modeling

Richard Charnin

Updated: July 8, 2012

2004 Monte Carlo Electoral Vote Simulation (pre-election and  exit polls)

The simulation model consists of 200 election trials based on pre-election state polls and post-election exit polls. It is strong circumstantial evidence that the election was stolen.

In the pre-election model, the state and national polls are adjusted for the allocation of undecided voters. The post-election model is based on unadjusted and adjusted state exit polls. Monte Carlo simulation is used to project state and aggregate vote shares and calculate the popular and electoral vote win probabilities.

The state win probability is a function of 1) the projected vote shares (after allocating undecided voters) and 2) the state poll margin of error.

The expected (theoretical) electoral vote can be calculated using a simple summation formula. It is just the product sum of the state win probabilities and corresponding electoral votes.

The purpose of the simulation is to calculate the overall probability of winning the electoral vote. As the number of election trials increase, the average (mean) electoral vote will approach the theoretical expected value.

The electoral vote win probability is the ratio of the number of winning election trials to the total number of trials.

In every presidential election, millions of voters are disenfranchised and millions of votes are uncounted. Forecasting models should have the following disclaimer:

Note: The following forecast will surely deviate from the official recorded vote. If they are nearly equal, then there must have been errors in the a) input data, b) assumptions, c) model logic and/or methodology.

Kerry led the weighted pre-election state and national polls by 1%. After allocating 75% of undecided voters to him, he was projected to win by 51.4-47.7%. Kerry had 51.1% in the unadjusted state exit poll aggregate (76,000 respondents) and 51.7% in the unadjusted National Exit Poll (13,660 respondents).

The National Election Pool, a consortium of six media giants, funds the exit polls. The published National Exit Poll is always forced to match the recorded vote. The Final 2004 NEP adjusted the actual exit poll responses to force a match to the recorded vote (Bush by 50.7-48.3%).

The large discrepancy between the exit polls and the vote count indicates that either a) the pre-election and unadjusted exit polls were faulty or b) the votes were miscounted, or c) a combination of both. Other evidence confirms that the votes were miscounted in favor of Bush.

The True Vote always differs from the official recorded vote due to uncounted, switched and stuffed ballots. Were the pollsters who forecast a Bush win correct? Or were Zogby and Harris correct in projecting that Kerry would win?

None of the pollsters mentioned the election fraud factor – the most important variable of all.

MODEL OVERVIEW

The workbook contains a full analysis of the 2004 election, based on four sets of polls:

(1) Pre-election state polls
(2) Pre-election national polls
(3) Post-election state exit polls
(4) National Exit Poll

Click the tabs at the bottom of the screen to select:
MAIN: Data input and summary analysis.
SIMULATION: Monte Carlo Simulation of state pre-election and exit polls.
NATPRE: Projections and analysis of 18 national pre-election polls.
In addition, three summary graphs are provided in separate sheets.

Calculation methods and assumptions are entered in the MAIN sheet:
1) Calculation code: 1 for pre-election polls; 2 for EXIT polls.
2) Undecided voter allocation (UVA): Kerry’s share (default 75%).
3) Exit Poll Cluster Effect: increase in margin of error (default 30%).
4) State Exit Poll Calculation Method:
1= WPD: average precinct discrepancy.
2= Best GEO: adjusted based on recorded vote geographic weightings.
3= Composite: further adjustment to include pre-election polls.

Note: The Composite state exit poll data set (12:40am) was downloaded from the CNN election site by Jonathan Simon. The polls were in the process of being adjusted to the incoming vote counts and weighted to include pre-election polls.The final adjustment at 1am forced a match to the final recorded votes.

2004 ELECTION MODEL

The Election Model was executed weekly from August to the election. It tracked state and national polls which were input to a 5000 trial Monte Carlo simulation. The final Nov. 1 forecast had Kerry winning 51.8% of the two-party vote and 337 electoral votes> He had a 99.8% electoral vote win probability: the percentage of trials in which he had at least 270 electoral votes.

Simulation forecast trends are displayed in the following graphs:

POLL SAMPLE-SIZE AND MARGIN OF ERROR

Approximately 600 were surveyed in each of the state pre-election polls (a 4% margin of error). The national aggregate has a lower MoE; approximately 30,000 were polled. In 18 national pre-election polls, the samples ranged from 800 (3.5% MoE) to 3500 (1.7% MoE).

In the exit polls, 76,000 voters were sampled. Kerry won the unadjusted state exit poll aggregate by 51.1-47.5%. He also won the unadjusted National Exit Poll (NEP) by 51.7-47.0%. The NEP is a 13,660 sample subset of the state exit polls.The NEP was adjusted to match the recorded vote -using the same 13660 respondents.

Assuming a 30% exit poll “cluster effect” (1.1% MoE), Kerry had a 98% probability of winning the popular vote. The Monte Carlo simulation indicates he had better than a 99% probability of winning the Electoral Vote.

The “cluster effect” is the percentage increase in the theoretical Exit Poll margin of error. When it is not practical to carry out a pure random sample, a common shortcut is to use an area cluster sample: primary Sampling Units (PSUs) are selected at random within the larger geographic area.

The Margin of Error (MoE) is a function of the sample size (n) and the polling percentage split: MoE = 1.96* Sqrt(P*(1-P)/n)

ELECTION FORECASTING METHODOLOGY

The Law of Large Numbers is the basis for statistical sampling. All things being equal, polling accuracy is directly related to sample size – the larger the sample, the smaller the margin of error (MoE). In an unbiased random sample, there is a 95% probability that the vote will fall within the MoE of the mean.

There are two basic methods used to forecast presidential elections:
1) Projections based on state and national polls
2) Time-series regression models

Academics and political scientists create multiple regression models to forecast election vote shares and run the models months in advance of the election. The models utilize time-series data, such as: economic growth, inflation, job growth, interest rates, foreign policy, historical election results, incumbency, approval rating, etc. Regression modeling is an interesting theoretical exercise, but it does not account for daily events which affect voter psychology.

Polling and regression models are analogous to the market value of a stock and its intrinsic (theoretical) value. The latest poll share is the equivalent of the current stock price. The intrinsic value of a stock is based on forecast cash flows. The intrinsic value is rarely equal to the market value.

The historical evidence is clear: state and national polls, adjusted for undecided voters and estimated turnout, are superior to time-series models executed months in advance.

Inherent problems exist in election models, the most important of which is never discussed: Election forecasters and media pundits never account for the probability of fraud. The implicit assumption is that the official recorded vote will accurately reflect the True Vote and that the election will be fraud-free.

MONTE CARLO SIMULATION

Monte Carlo is a random process of repeated experimental “trials” applied to a mathematical system model. The Election Simulation Model runs 200 trial “elections” to determine the expected electoral vote and win probability.

Statistical polling (state and national) ideally is an indicator of current voter preference. Pre-election poll shares are adjusted for undecided voters and state win probabilities are calculated. The probabilities are input to a Monte Carlo simulation based on random numbers. The final probability of winning the electoral vote is simply the number of winning election trials divided by the total number of trials (200 in the ESM; 5000 in the Election Model).

The only forecast assumption is the allocation of undecided/other voters. Historically, 70-80% of undecided voters break for the challenger. If the race is tied at 45-45, a 60-40% split of undecided voters results in a 51-49% projected vote share.

ELECTORAL AND POPULAR VOTE WIN PROBABILITIES

The theoretical expected electoral vote for a candidate is a simple calculation. It is just the sum of the 51 products: state electoral vote times the win probability. In the simulation, the average (mean) value will converge to the theoretical value as the number of election trial increase.

The probability of winning the popular vote is a function of the projected 2-party vote share and polling margin of error. These are input to the Excel normal distribution function. The simulation generates a electoral vote win probability that is not sensitive to minor changes in the state polls.

Prob (win) = NORMDIST (P, 0.50, MoE/1.96, True)

For each state in an election trial, a random number (RND) between 0 and 1 is generated and compared to the probability of winning the state. For example, if Kerry has a 90% probability of winning Oregon and RND is less than 0.90, Kerry wins 7 electoral votes. Otherwise, if RND is greater than 0.90, Bush wins. The procedure is repeated for all 50 states and DC. The election trial winner has at least 270 EV.

The electoral vote win probability is directly correlated to the probability of winning the national popular vote. But electoral vote win probabilities in models developed by academics and bloggers are often incompatible with the projected national vote shares.

For example, assume a 53% projected national vote share. If the corresponding EV win probability is given 88%, the model design/logic is incorrect; the 53% share and 88% win probability are incompatible. For a 53% share, the win probability is virtually 100%. This is proved using Monte Carlo simulation based on state win probabilities in which there is a 53% aggregate projected national share.

The state win probability is based on the final pre-election polls which typically sample 600 likely voters (a 4% MoE). In the 2004 Election Simulation Model, the electoral vote is calculated using 200 election trials. The average (mean) electoral vote is usually within a few votes of the median (middle value). As the number of simulation trials increase, the mean approaches the theoretical expected value. That is due to the Law of Large Numbers.

SENSITIVITY ANALYSIS

A major advantage of the Monte Carlo Simulation method is that the win probability is not sensitive to minor deviations in the state polls. It is not an all-or-nothing proposition as far as allocating the electoral vote is concerned. A projected 51% vote share has less electoral “weight” than a 52% share, etc. Electoral vote projections from media pundits and Internet bloggers use a single snapshot of the latest polls to determine a projected electoral vote split. This can be misleading when the states are competitive and often results in wild electoral vote swings.

In the Election Model, five projection scenarios over a range of undecided voter allocation assumptions display the effects on aggregate vote share, electoral vote and win probability.

Snapshot projections do not provide a robust expected electoral vote split and win probability. That’s because unlike the Monte Carlo method, they fail to consider the two bedrocks of statistical analysis: The Law of Large Numbers and the Central Limit Theorem.

For example, assume that Florida’s polls shift 1% from 46-45 to 45-46. This would have a major impact in the electoral vote split. On the other hand, in a Monte Carlo simulation, the change would have just a minimal effect on the expected (average) electoral vote and win probability. The 46-45 poll split means that the race is too close to clearly project a winner; both candidates have a nearly equal win probability.

NORMAL DISTRIBUTION

The Excel function has a very wide range of applications in statistics, including hypothesis testing.

NORMDIST (x, mean, stdev, cumulative)
X is the value for which you want the distribution.
Mean is the arithmetic mean of the distribution.
Stdev is the standard deviation of the distribution.

EXAMPLE: Calculate the probability Kerry would win Ohio based on the exit poll assuming a 95% level of confidence.
Sample Size = 1963; MoE =2.21%; Cluster effect = 20%; Adj. MoE = 2.65%
Std Dev = 1.35% = 2.65% / 1.96

Kerry win probability:
Kerry = 54.0%; Bush = 45.5%; StdDev = 1.35%
Kerry 2-party share: 54.25%
Probability = NORMDIST(.5425, .5, .0135, TRUE)= 99.92%

BINOMIAL DISTRIBUTION

BINOMDIST is used in problems with a fixed number of tests or trials, where the outcome of any trial is success or failure. The trials are independent. The probability of success is constant in each trial (heads or tails, win or lose).

EXAMPLE: Determine the probability that the state exit poll MoE is exceeded in at least n states assuming a 95% level of confidence. The one-tail probability of Bush exceeding his exit poll share by the MoE is 2.5%. Therefore the probability that at most N-1 states fall within the MoE is:
Prob = BINOMDIST (N-1, 50, P, TRUE)
N = 16 states exceeded the MoE at 12:22am in favor of Bush.
The probability that the MoE is exceeded in at least 16 exit polls for Bush:
= 1- BINOMDIST (15, 50, 0.025, TRUE)
= 5.24E-14 or 1 in 19,083,049,268,519

A SAMPLING PRIMER

The following is an edited summary from http://www.csupomona.edu/~jlkorey/POWERMUTT/Topics/data_collection.html

A random sample is one which each outcome has an equal probability of being included. It is an unbiased estimate of the characteristics of the population in which the respondents are representative of the population as a whole.

The reliability of the sample increases with the size of the sample. Ninety-five times out of a hundred, a random sample of 1,000 will be accurate to within about 3 percentage points. The sample has a margin of error of approximately plus or minus (±) 3 percent at a 95 percent confidence level. If a random sample of 1,000 voters shows that 60 percent favor candidate X, there is a 95 percent chance that the real figure in the population is in the 57 to 63 percent range.

Beyond a certain point, the size of the population makes little difference. The confidence interval is not reduced dramatically. Therefore pre-election national polls usually don’t survey more than about 1,500 respondents. Increasing this number increases the cost proportionately, but the margin of error will be reduced only a little.

Often it is not practical to carry out a pure random sample. One common shortcut is the area cluster sample. In this approach, a number of Primary Sampling Units (PSUs) are selected at random within a larger geographic area. For example, a study of the United States might begin by choosing a subset of congressional districts. Within each PSU, smaller areas may be selected in several stages down to the individual household. Within each household, an individual respondent is then chosen. Ideally, each stage of the process is carried out at random. Even when this is done, the resulting sampling error will tend to be a little higher than in a pure random sample,[7] but the cost savings may make the trade-off well worthwhile.

Somewhat similar to a cluster sample is a stratified sample. An area cluster sample is appropriate when it would be impractical to conduct a random sample over the entire population being studied. A stratified sample is appropriate when it is important to ensure inclusion in the sample of sufficient numbers of respondents within subcategories of the population.

Even in the best designed surveys, strict random sampling is a goal that can almost never be fully achieved under real world conditions, resulting in non-random (or “systematic”) error. For example, assume a survey is being conducted by phone. Not everyone has one. Not all are home when called. People may refuse to participate. The resulting sample of people who are willing and able to participate may differ in systematic ways from other potential respondents.

Apart from non-randomness of samples, there are other sources of systematic error in surveys. Slight differences in question wording may produce large differences in how questions are answered. The order in which questions are asked may influence responses. Respondents may lie.

Journalists who use polls to measure the “horse race” aspect of a political campaign face additional problems. One is trying to guess which respondents will actually turn out to vote. Pollsters have devised various methods for isolating the responses of “likely voters,” but these are basically educated guesses. Exit polls, in which voters are questioned as they leave the voting area, avoid this problem, but the widespread use of absentee voting in many states creates new problems. These issues are usually not a problem for academic survey research. Such surveys are not designed to predict future events, but to analyze existing patterns. Some are conducted after the election. The American National Election Study, for example, includes both pre and post election interviews. Post election surveys are not without their own pitfalls, however. Respondents will sometimes have a tendency to report voting for the winner, even when they did not.

The American National Election Study split its sample between face to face and telephone interviews for its 2000 pre-election survey. The response rate was 64.8 percent for the former, compared to 57.2 percent for the latter. An analysis of a number of telephone and face-to-face surveys showed that face-to-face surveys were generally more representative of the demographic characteristics of the general population. Note that many telephone surveys produce response rates far lower than that obtained by the ANES.

Another approach is the online poll, in which the “interaction” is conducted over the Internet. Like robo polls, online polls are also less expensive than traditional telephone surveys, and so larger samples are feasible. Because they require respondents to “opt in,” however, the results are not really random samples.

When samples, however obtained, differ from known characteristics of the population (for example, by comparing the sample to recent census figures), samples can be weighted to compensate for under or over representation of certain groups. There is still no way of knowing, however, whether respondents and non-respondents within these groups differ in their political attitudes and behavior.

1 Comment

Posted by on September 1, 2011 in 2004 Election