Oct.6
The usual DU naysayers question the Election ModelEV Simulation

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x508722#508961

This is the fivethirtyeight.com electoral vote frequency graph:
http://4.bp.blogspot.com/_ov-pT1x-W8Y/SNU03EJMSzI/AAAAAAAACYE/IQVTZ_z7L9I/S1600-R/0920_evdist.png


This is the electoral vote frequency distribution produced by the Election Model Monte Carlo simulation:
http://www.RichardCharnin.com/2008ElectionModel_30550_image001.gif


Has Febble ever developed a model which uses the Monte Carlo Simulation method?
Does Febble know how to calculate the EXPECTED electoral vote based on the latest state polls?

Can she describe how the Election Model derives the electoral vote distribution histogram step-by-step?
Can she describe how 538.com derives the electoral vote distribution histogram step-by-step?
Can she describe the algorithm she would use to calculate the EV distribution – step-by-step?

Does she realize that Princeton Professor Sam Wang derives virtually identical results as the Election Model using his exhaustive combinatorial meta-analysis model?

Is Febble aware of the fact that 538.com uses factors other than state polls in projecting the EV and win probability?
Is she aware that 538.com projects Election Day results while the Election Model assumes a FRAUD-FREE election is held TODAY?

These are the first 20 of 5000 election trials using polling data as of Oct.6.
http://www.RichardCharnin.com/MonteCarloEVSimulation20trials1006.pdf

Obama won all 5000 election trials. How would Febble calculate his EV win probability? Febble should calculate the average EV of the 20 trials and see how close it is to the expected 358 EV.

OnTheOtherHand talks about pixels and ogives while totally avoiding the mathematics. Look at the 538 curve. About 13% of the election trials are below 270. Graph pixels have nothing do with the basic simulation results.

THIS IS WHAT IS RELEVANT: Look at the Oct. 6 graph. ALL 5000 election trials result in Obama winning more than 270 EV. The histogram breaks up the distribution into increments of 5 EV which is a reasonable bin size. Would OTOH suggest a bin size of ONE?

In fact, Excel has a limit of approximately 40 bins in its Histogram program. Does OTOH know what a frequency distribution curve is? Does he know what the cumulative distribution curve represents?

The basic statistics which define the frequency histogram are what counts:
The MEAN, MEDIAN, MAXIMUM, MINIMUM, 95% CONFIDENCE INTERVAL.

Note that the mean and median are virtually identical, a clear indication of the LAW OF LARGE NUMBERS. It proves that 5000 simulations are SUFFICIENT TO ILLUSTRATE THE CONVERGENCE TO THE THEORETICAL MEAN (EXPECTED ELECTORAL VALUE).
The expected EV (mean) is calculated as a simple summation:

Expected EV = ∑ P(i) * EV (i), for i=1, 51 states.

P(i) is the probability of winning state (i) and its EV(i) electoral votes.
P(i) is calculated using the most recent state poll average vote share adjusted for an undecided voter allocation normal distribution. A 4% state poll margin of error is assumed.

OTOH may be interested in this graph of 5000 election trials, which is not a histogram but shows that Obama won all 5000 trials:
http://www.geocities.com/electionmodel/2008ElectionModel_3523_image001.gif

http://www.intrade.com/

Obama 364

McCain 174

The issue is not the EV spread now that Obama is leading in the Battleground states.
I had him leading by 100-200 EV all along because I projected that he would win 60% of the undecided vote. Now Intrade, RCP, 538 and electoral-vote.com have Obama in the lead with 340-370 EV. Better late than never, I guess.

The point of contention is the EV distribution curve. In the Election Model, the curve represents 5000 Monte Carlo election trials. Obama won them all, so his EV win probability is 100%. In each trial, he had more than 270EV. Febble and OTOH take issue with the bin size but avoid discussing the mathematics.

Now Intrade is often quoted as the definitive market probability estimate of the election outcome. But it just reflects the herd instinct of investors who go by their gut response to the latest polls. BUT THE BETTING ODDS ON INTRADE DO NOT REFLECT THE TRUE MATHEMATICAL ODDS, WHICH ARE DERIVED USING MONTE CARLO SIMULATION.

THE ELECTION MODEL ASSUMES THE ELECTION IS HELD TODAY AND IS FRAUD-FREE. BASED ON THE CURRENT POLLS, OBAMA HAS A 100% PROBABILITY OF WINNING THE ELECTORAL VOTE SINCE HE WON ALL 5000 MONTE CARLO ELECTION TRIALS.

Not only that, Obama has a 98-100% probability of winning the popular vote. The probability is a function of the standard deviation (equal to MoE/1.96) and the projected vote share. It confirms the Monte Carlo simulation ELECTORAL VOTE WIN PROBABILITY. THE EV WIN PROBABILITY IS A FUNCTION OF THE STATE WIN PROBABILITIES.

Granted, the probabilities may seem too high...BUT THEY ARE MATHEMATICALLY CORRECT GIVEN THE PROJECTED VOTE SHARES.

Febble and OTOH are expert at misinformation and always avoid the devil in the detail - the mathematics. Febble posted today that voter disenfranchisement was THE ONLY ANOMALY in the 2004 election. HERE IT IS 2008 AND FEBBLE STILL MAINTAINS WITHOUT QUALIFICATION THAT ELECTRONIC VOTE-SWITCHING WAS NON-EXISTENT IN 2004.

FEBBLE, TWO SIMPLE QUESTIONS:
1) HOW DO YOU KNOW THERE WAS NO VOTE-SWITCHING IN 2004/2006? SHOW US THE EVIDENCE.
2) WHISTLEBLOWER STEPHEN SPOONAMORE IS A REPUBLICAN EXPERT IN COMPUTER SECURITY. DO DO YOU CLAIM TO KNOW MORE THAN HE DOES ABOUT ELECTRONIC FRAUD? IF YOU HAVE NOT HEARD OF HIM, DO YOUR HOMEWORK. IT'S ALL RIGHT HERE.
http://www.afterdowningstreet.org/node/35475

Obviously Intrade only reflects the gut feelings of "investors" who are willing to put their money on who they think are the winners. Obviously 100% of the investors don't think Obama will win.

I guess what I was getting at is how close these "investors" match your simulations on the Electoral College Votes.

Obviously Intrade only reflects the gut feelings of "investors" who are willing to put their money on who they think are the winners. Obviously 100% of the investors don't think Obama will win.

I guess what I was getting at is how close these "investors" match your simulations on the Electoral College Votes.

From Intrade on Monday:

As of Monday afternoon, Intraders give Obama the edge in states carrying 338 electoral votes. He could lose 69 of those and still win. For example Intrade has FL, VA, OH, NV as the most vulnerable blue states right now.

They could all turn red and Obama still wins.

Intrade markets are historically 100% predictive in US elections when the prices are higher than 70 or lower than 30. If we take the states currently in that range, and add in Colorado, which is at 69 for Obama, we get 44 states that, given the level of volatility and the time left, are unlikely to turn. From those states, Obama controls 273 electoral votes to 163 for McCain (Karl Rove himself pointed this out today).


http://www.intrade.com/jsp/intrade/misc/blog/?initialBlogId=stumpy_1

..................................

Since Monday, Colorado, Virginia, Ohio, and Nevada are trading over 70 and Florida is now 67.

And today Obama is trading at 78 and McCain at 22.

So I would say Intrade is predicting 100% that Obama will win at this point.

From Intrade on Monday:

As of Monday afternoon, Intraders give Obama the edge in states carrying 338 electoral votes. He could lose 69 of those and still win. For example Intrade has FL, VA, OH, NV as the most vulnerable blue states right now.

They could all turn red and Obama still wins.

Intrade markets are historically 100% predictive in US elections when the prices are higher than 70 or lower than 30. If we take the states currently in that range, and add in Colorado, which is at 69 for Obama, we get 44 states that, given the level of volatility and the time left, are unlikely to turn. From those states, Obama controls 273 electoral votes to 163 for McCain (Karl Rove himself pointed this out today).


http://www.intrade.com/jsp/intrade/misc/blog/?initialBlogId=stumpy_1

..................................

Since Monday, Colorado, Virginia, Ohio, and Nevada are trading over 70 and Florida is now 67.

And today Obama is trading at 78 and McCain at 22.

So I would say Intrade is predicting 100% that Obama will win at this point.

Intrade markets are HISTORICALLY 100% predictive in US elections WHEN the prices are higher than 70 or lower than 30. They have always been correct since they started predicting elections under those conditions. But how many elections are they referring to? Four? Eight? Not enough data points to draw any firm conclusion.

They need quite a few more elections to claim that the predictive relationship reflects a 100% certainty, something like the Election Model Monte Carlo simulation in which Obama won all 5000 election trials.

The Intrade win probability is 78%, not 100%.

Intrade markets are HISTORICALLY 100% predictive in US elections WHEN the prices are higher than 70 or lower than 30. They have always been correct since they started predicting elections under those conditions. But how many elections are they referring to? Four? Eight? Not enough data points to draw any firm conclusion.

They need quite a few more elections to claim that the predictive relationship reflects a 100% certainty, something like the Election Model Monte Carlo simulation in which Obama won all 5000 election trials.

The Intrade win probability is 78%, not 100%.

Since in any given election year there can be hundreds of elections to bet on.

Since in any given election year there can be hundreds of elections to bet on.