TruthIsAll
10-23-2008, 08:11 PM
2008 Election Model
A Monte Carlo Electoral Vote Simulation
http://www.geocities.com/electionmodel/2008ElectionModel.htm
Updated: Oct. 23
TruthIsAll
Over the past few days, there has been a sharp rise in Obama’s state and national poll averages. Is it due to Powell’s endorsement? Palin’s 175k wardrobe? The economy? The stock market? Calling Obama a “socialist palling around with terrorists”? Exposing Joe the unlicensed plumber? Job layoffs? Is it the heavy Democratic new voter registration and powerful Obama GOTV campaign? Is it the rampant GOP voter purge and registration fraud? What about returning Gore and Kerry voters royally pissed about the stolen elections? Is it a rebellion against the complicit media and politicians who have avoided discussing election fraud since the SCOTUS 2000 coup? Or is it a combination of all the above?
The Election Model (EM) calculates an average of the two or three most recent state polls and projects five vote share scenarios (5000 trials each) over a range of undecided voter allocations from 40-90%. Even in the worst-case scenario in which Obama captures just 40% of the undecided vote, he won all 5000 election trials with an expected (average) 359.3 electoral votes. He won the base case scenario (60% UVA) with an average 374.7 EV. The median EV was 378.
Obama exceeded 350 EV in 4908 election trials, so there is a 98.2% probability that he will win at least 350 EV. He exceeded 370 EV in 4007 trials (80.1%).
Note how the Monte Carlo mean EV (374.51) matches the theoretical summation formula EV (374.67) based on state win probabilities. This illustrates The Law of Large Numbers (LLN): It took 5000 simulated election trials for the MEAN EV of 5000 to CONVERGE to the THEORETICAL EV. In other words, 5000 trials are more than sufficient; we are in the “long run”. A meta-analysis requiring the calculation of millions of EV combinations is overkill, Princeton!
The EM assumes that current polls reflect the will of the electorate and that a fraud-free election is held today.
National polls reflect current vote preferences; state polls lag the nationals by a week on average. Obama’s projected state aggregate 2-party vote (54.39%) is approaching the national average (54.60%). View the State vs. National vote share projection Trend.
The five most critical states weighted by the electoral vote and poll spread are FL (32.0), CO (7.1), IN (11.1), NC (14.8) and MO (13.0). The values represent the optimal percentage of campaign resources to be allocated to these states as of today (approximately 76% of available funds).
For McCain to win, based on current projections, he needs a minimum of 8.3% (1 in 12) of Obama’s votes switched to his column.
Projected Vote Shares, Electoral Votes and Win Probabilities
There are many electoral vote forecasting models. For example, the DeSart and Holbrook Model produced results similar to the Election Model: The latest run (Oct. 16) predicts that Barack Obama will win the election with 52.85%* of the national 2-party popular vote to to John McCain's 47.15%. The model also predicts that Obama will win the presidency with 354 electoral votes to McCain's 184. Based on these results, they calculate a 99.99% probability that Barack Obama will win the election.
Electoral-vote.com and RealClearPolitics now closely match the Election Model. As indicated in a prior update, these sites assign the full electoral vote to the state poll leader regardless of the spread; they avoid using state win probabilities in calculating the EV and do not allocate undecided voters. When the polls were close and McCain was leading in a number of swing states, the unadjusted EV understated Obama’s total. The daily EV tracking numbers were volatile compared to the Election Model which calculates the expected (mean) and median EV using Monte Carlo simulation. Now that Obama has pulled ahead in all of the battleground states, undecided voter allocation has very little impact on the expected EV.
This was the electoral vote.com map on Nov 1, 2004.
The discrepancy in win probabilities between the Election Model (EM) and fivethirtyeight.com (538) is due to differences in methodology.
. 538 attempts to forecasts the Election Day result; the EM assumes the election is held today.
. 538 weights state poll projections based on pollster rankings and other factors; EM does not rank pollsters.
Ranking pollsters based on prior election results is a two-edged sword. For example, in 2004 Rasmussen was very accurate in his final projection. But since the election was rigged, does that mean he was more accurate than Zogby who projected that Kerry would win the True Vote? In the 2000 election (before HAVA) Zogby closely matched the recorded vote and Rasmussen was way off.
The Monte Carlo simulation was the basis for the Electoral Vote Simulation Frequency chart. Note that ALL 5000 election trials are to the right of the 270 mark; therefore Obama’s win probability is 100%. His 99%+ popular vote win probability is a function of the aggregate state/national poll margin of error. The popular and electoral vote win probabilities confirm each other.
2008 Election Calculator
http://www.geocities.com/electionmodel/2008ElectionCalculator.htm
In May 2008, the Calculator projected Obama to win the True Vote by 71-59m.
The model has been updated to include new information.
Obama is now projected to win by an 80-58 million vote landslide.
Of course, the assumption is that the election will be fraud-free
1) A 20%+ increase in new registered voters, the great majority of whom are Democratic.
2) A slight increase in the estimated Obama share of returning Bush voters.
3) An increase in third party vote share to 4%.
Oct 2008
2004 Estimated vote share
Turnout Votes Mix Obama McCain Other
DNV - 29.9 20.8% 59% 35% 6%
Kerry 95% 60.6 42.2% 92% 5% 3%
Bush 95% 51.6 35.9% 11% 86% 3%
Other 95% 1.6 1.1% 64% 11% 25%
Total 113.7 143.7 100% 55.7% 40.4% 3.9%
143.7 - 80.1 58.1 5.6
May-08
2004 Estimated vote share
Turnout Votes Mix Obama McCain Other
DNV - 17.2 13.1% 59% 40% 1%
Kerry 95% 60.5 46.2% 89% 10% 1%
Bush 95% 51.6 39.4% 11% 88% 1%
Other 95% 1.6 1.2% 70% 11% 19%
Total 113.7 130.9 100% 54.1% 44.7% 1.2%
130.9 - 70.8 58.5 1.6
If Obama wins just 9% of returning Bush voters and 90% of Kerry voters, he would win by 17.5m votes (54.2-41.6%).
If he wins just 55% of new voters and 90% of Kerry voters, he would win by 17.2m votes (54-42%).
The 1988-2004 Election Calculator was developed as a response to the Final 2004 National Exit Poll.
The Final was forced to match the recorded vote using impossible weights: 43% (52.6m) were Bush 2000 voters; 37% Gore voters.
But Bush only had 50.5m votes in 2000.
Approximately 2.5m died and another 2.5m did not return to vote.
Therefore, only 45.5m Bush 2000 voters could have returned to vote in 2004.
The Final NEP overstated the Bush vote by 7 million in order to match a corrupt miscounted vote.
The 2004 True Vote calculation assumed an estimated 100.1m returning 2000 voters.
100.1m = 110.8m Votes cast in 2000 less 5.4m voter mortality times 95% turnout.
Vote shares were based on the 12:22am National Exit Poll.
The model determined that Kerry won by 66.9-57.1 million.
Kerry did slightly better (53.2%) than the unadjusted state exit poll (52.0%) aggregate.
The results indicate that 5.4m votes (8.0% of Kerry’s total) were switched from Kerry to Bush.
2000 12:22am NEP Vote shares
Turnout Votes Mix Kerry Bush Other
DNV - 25.6 20.4% 57% 41% 2%
Gore 95% 49.7 39.5% 91% 8% 1%
Bush 95% 46.6 37.1% 10% 90% 0%
Other 95% 3.8 3.0% 64% 17% 19%
Total 100.1 125.7 100% 53.2% 45.4% 1.4%
Votes 125.7 66.9 57.1 1.7
Recorded 48.3% 50.7% 1.0%
Vote 122.3 59.0 62.0 1.2
Unadj Exit Poll - 52.0% 47.0% 1.0%
Deviation from True - -1.2% 1.6% -0.4%
A Monte Carlo Electoral Vote Simulation
http://www.geocities.com/electionmodel/2008ElectionModel.htm
Updated: Oct. 23
TruthIsAll
Over the past few days, there has been a sharp rise in Obama’s state and national poll averages. Is it due to Powell’s endorsement? Palin’s 175k wardrobe? The economy? The stock market? Calling Obama a “socialist palling around with terrorists”? Exposing Joe the unlicensed plumber? Job layoffs? Is it the heavy Democratic new voter registration and powerful Obama GOTV campaign? Is it the rampant GOP voter purge and registration fraud? What about returning Gore and Kerry voters royally pissed about the stolen elections? Is it a rebellion against the complicit media and politicians who have avoided discussing election fraud since the SCOTUS 2000 coup? Or is it a combination of all the above?
The Election Model (EM) calculates an average of the two or three most recent state polls and projects five vote share scenarios (5000 trials each) over a range of undecided voter allocations from 40-90%. Even in the worst-case scenario in which Obama captures just 40% of the undecided vote, he won all 5000 election trials with an expected (average) 359.3 electoral votes. He won the base case scenario (60% UVA) with an average 374.7 EV. The median EV was 378.
Obama exceeded 350 EV in 4908 election trials, so there is a 98.2% probability that he will win at least 350 EV. He exceeded 370 EV in 4007 trials (80.1%).
Note how the Monte Carlo mean EV (374.51) matches the theoretical summation formula EV (374.67) based on state win probabilities. This illustrates The Law of Large Numbers (LLN): It took 5000 simulated election trials for the MEAN EV of 5000 to CONVERGE to the THEORETICAL EV. In other words, 5000 trials are more than sufficient; we are in the “long run”. A meta-analysis requiring the calculation of millions of EV combinations is overkill, Princeton!
The EM assumes that current polls reflect the will of the electorate and that a fraud-free election is held today.
National polls reflect current vote preferences; state polls lag the nationals by a week on average. Obama’s projected state aggregate 2-party vote (54.39%) is approaching the national average (54.60%). View the State vs. National vote share projection Trend.
The five most critical states weighted by the electoral vote and poll spread are FL (32.0), CO (7.1), IN (11.1), NC (14.8) and MO (13.0). The values represent the optimal percentage of campaign resources to be allocated to these states as of today (approximately 76% of available funds).
For McCain to win, based on current projections, he needs a minimum of 8.3% (1 in 12) of Obama’s votes switched to his column.
Projected Vote Shares, Electoral Votes and Win Probabilities
There are many electoral vote forecasting models. For example, the DeSart and Holbrook Model produced results similar to the Election Model: The latest run (Oct. 16) predicts that Barack Obama will win the election with 52.85%* of the national 2-party popular vote to to John McCain's 47.15%. The model also predicts that Obama will win the presidency with 354 electoral votes to McCain's 184. Based on these results, they calculate a 99.99% probability that Barack Obama will win the election.
Electoral-vote.com and RealClearPolitics now closely match the Election Model. As indicated in a prior update, these sites assign the full electoral vote to the state poll leader regardless of the spread; they avoid using state win probabilities in calculating the EV and do not allocate undecided voters. When the polls were close and McCain was leading in a number of swing states, the unadjusted EV understated Obama’s total. The daily EV tracking numbers were volatile compared to the Election Model which calculates the expected (mean) and median EV using Monte Carlo simulation. Now that Obama has pulled ahead in all of the battleground states, undecided voter allocation has very little impact on the expected EV.
This was the electoral vote.com map on Nov 1, 2004.
The discrepancy in win probabilities between the Election Model (EM) and fivethirtyeight.com (538) is due to differences in methodology.
. 538 attempts to forecasts the Election Day result; the EM assumes the election is held today.
. 538 weights state poll projections based on pollster rankings and other factors; EM does not rank pollsters.
Ranking pollsters based on prior election results is a two-edged sword. For example, in 2004 Rasmussen was very accurate in his final projection. But since the election was rigged, does that mean he was more accurate than Zogby who projected that Kerry would win the True Vote? In the 2000 election (before HAVA) Zogby closely matched the recorded vote and Rasmussen was way off.
The Monte Carlo simulation was the basis for the Electoral Vote Simulation Frequency chart. Note that ALL 5000 election trials are to the right of the 270 mark; therefore Obama’s win probability is 100%. His 99%+ popular vote win probability is a function of the aggregate state/national poll margin of error. The popular and electoral vote win probabilities confirm each other.
2008 Election Calculator
http://www.geocities.com/electionmodel/2008ElectionCalculator.htm
In May 2008, the Calculator projected Obama to win the True Vote by 71-59m.
The model has been updated to include new information.
Obama is now projected to win by an 80-58 million vote landslide.
Of course, the assumption is that the election will be fraud-free
1) A 20%+ increase in new registered voters, the great majority of whom are Democratic.
2) A slight increase in the estimated Obama share of returning Bush voters.
3) An increase in third party vote share to 4%.
Oct 2008
2004 Estimated vote share
Turnout Votes Mix Obama McCain Other
DNV - 29.9 20.8% 59% 35% 6%
Kerry 95% 60.6 42.2% 92% 5% 3%
Bush 95% 51.6 35.9% 11% 86% 3%
Other 95% 1.6 1.1% 64% 11% 25%
Total 113.7 143.7 100% 55.7% 40.4% 3.9%
143.7 - 80.1 58.1 5.6
May-08
2004 Estimated vote share
Turnout Votes Mix Obama McCain Other
DNV - 17.2 13.1% 59% 40% 1%
Kerry 95% 60.5 46.2% 89% 10% 1%
Bush 95% 51.6 39.4% 11% 88% 1%
Other 95% 1.6 1.2% 70% 11% 19%
Total 113.7 130.9 100% 54.1% 44.7% 1.2%
130.9 - 70.8 58.5 1.6
If Obama wins just 9% of returning Bush voters and 90% of Kerry voters, he would win by 17.5m votes (54.2-41.6%).
If he wins just 55% of new voters and 90% of Kerry voters, he would win by 17.2m votes (54-42%).
The 1988-2004 Election Calculator was developed as a response to the Final 2004 National Exit Poll.
The Final was forced to match the recorded vote using impossible weights: 43% (52.6m) were Bush 2000 voters; 37% Gore voters.
But Bush only had 50.5m votes in 2000.
Approximately 2.5m died and another 2.5m did not return to vote.
Therefore, only 45.5m Bush 2000 voters could have returned to vote in 2004.
The Final NEP overstated the Bush vote by 7 million in order to match a corrupt miscounted vote.
The 2004 True Vote calculation assumed an estimated 100.1m returning 2000 voters.
100.1m = 110.8m Votes cast in 2000 less 5.4m voter mortality times 95% turnout.
Vote shares were based on the 12:22am National Exit Poll.
The model determined that Kerry won by 66.9-57.1 million.
Kerry did slightly better (53.2%) than the unadjusted state exit poll (52.0%) aggregate.
The results indicate that 5.4m votes (8.0% of Kerry’s total) were switched from Kerry to Bush.
2000 12:22am NEP Vote shares
Turnout Votes Mix Kerry Bush Other
DNV - 25.6 20.4% 57% 41% 2%
Gore 95% 49.7 39.5% 91% 8% 1%
Bush 95% 46.6 37.1% 10% 90% 0%
Other 95% 3.8 3.0% 64% 17% 19%
Total 100.1 125.7 100% 53.2% 45.4% 1.4%
Votes 125.7 66.9 57.1 1.7
Recorded 48.3% 50.7% 1.0%
Vote 122.3 59.0 62.0 1.2
Unadj Exit Poll - 52.0% 47.0% 1.0%
Deviation from True - -1.2% 1.6% -0.4%