ONE IN 4.5 BILLION? IT'S ACTUALLY MUCH, MUCH WORSE THAN THAT.

[TruthIsAll]

In the interest of mathematical accuracy, I must confess: My original analysis was actually too conservative when I estimated that the odds were 1 in 4.5 billion that 16 of 51 states would move beyond the Exit Poll Margin of Error to the actual Bush vote.

In the calculation, I used the probability of 5% that a single state would deviate beyond the MOE as input to the Excel Binomial Distribution function. That was actually correct on its face – but it's the probability of a move beyond the MOE, regardless of whether the move is favorable to Bush OR Kerry.

In fact, what we really want is the probability that the vote would deviate beyond the MOE to Bush, not Kerry. This is exactly what happened. So that is why we must use 2.5% and NOT 5.0% as our input probability. We just split the probability in half, the half that would go to Bush..

So, what is the effect of this seemingly small, innocuous change on our final probability estimate? Well, it means that the probability these 16 deviations could be due to CHANCE alone becomes EVEN MORE REMOTE. And this is actually an understatement of BIBLICAL PROPORTIONS.

Get ready. Grab your chair. Hold your hat.

Here are the odds that 16 out of 51 states would move beyond the MOE in favor of Bush, again using the Binomial Distribution. But this time with .025 (rather than .05) as the probability that a given state would move beyond the MOE to Bush:

The probability P is: P =1-BINOMDIST(16,51,0.025,TRUE)

P = 0.0000000000004996%

The odds are 1/P or 1 out of 200.159 TRILLION that these changes could have occurred due to chance alone.

ONE out of 4.5 BILLION?
ONE out of 200 TRILLION?

There is NO practical difference.
But the second one is MATHEMATICALLY CORRECT.

Here are the odds for N states moving beyond the MOE in favor of Bush:

N The odds are 1 out of:
1- 3
2- 7
4- 113
6- 3,715
8- 223,016
10- 22,192,000
12- 3,432,782,579
14- 788,997,832,405

16- 200,159,983,438,689

That's right. You are seeing correctly. Don't laugh. Don't cry.
It's 1 out of 200 trillion.

[TruthIsAll]

1. HERE IS THE ORIGINAL 1 IN 4.5 BILLION THREAD

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x79760

[The Flaming Red Head]

4. I just tried to kick up that thread but they shut me down

[FULL_METAL_HAT]

2. How about odds being 294,862,972 to 1 ??

http://www.census.gov/cgi-bin/popclock

:^)

[cavanaghjam]

3. So you're still not changing

the bottom line - it's pretty fucking unlikely.

[Okieprogie]

5. weak point?

Not to argue with you, and I appreciate your work... but the basis of your calculations is the assumption that the exit polls achieved a random sampling, correct?

That is where the problem is. Is it really safe to assume that the exit polls were a random sample? The repugs were claiming that the early exit polls were 57% women - clearly not a random sample.

Now, why would the sample not be random? There are a lot of different possibilities, the answer is that nobody knows for sure. However, if it is in fact the case that there is a vast cultural divide between us and the repugs, is it not safe to assume that we might particpate in polls at a greater or lesser rate than the repugs?

[aquart]

8. But...but...BushCo told us those security moms were Bush votes.

And they wouldn't lie.

[TruthIsAll]

17. Then how has Mitofsky made a nice living doing exit polls for 25 years?

Someone must believe in them.

You will have to accept that as my last word with you on the subject.

I suggest you do your own research as to the historical accuracy of exit polling.

Come back and treat us to an informative post.

[Lucky Luciano]

6. How did you arrive at the 5% figure in the first place

that a state would deviate from the MOE from the exit polling?

Do the MOEs use 95% confidence intervals? Or are you basing this on the historical accuracy of exit polls?

Just curious...I know a lot of math, but I did pure math, so my prob/stats is a little weak.

Of course there is the whole random sampling thing.....Exit polls have been very accurate in the past, but this time it would be advantageous for democrats to skew the exit polls by not doing a random sample...both side could be playing games here, though the Repubs are notoriously more criminal of late (Kennedy and Nixon were btoh election crooks back in the day). Just being devil's advocate.

[TruthIsAll]

10. If you use MOE, you are using 95% confidence intervals..

Otherwise, what's the point of a margin of error if you don't assign a probability to it?

In 95 times out of 100, the election results will fall WITHIN the polling MOE, whatever it may be.

And in the other 5 times, the results will fall OUTSIDE of the MOE. Half of these results can be expected to fall on either side of the Population Mean.

By the way, the MOE for any sample size N is equal to 1/sqrt(N), at least in standard polls. The MOE is even smaller in EXIT polls.

So for sample size N, the MOE is calulated as:
N MOE
900: 1/30 = 3.33%
1600: 1/40 = 2.50%
2500: 1/50 = 2.00%

Note: MOE = 1.96 * Standard deviation
so

Std Dev= MOE / 1.96

The probability is 95% that the population Mean will fall within 1.96 standard deviations on either side of the Sample mean

[Lucky Luciano]

12. Sure,

but how does this hold up to the historical exit polls? It would be interesting to see if 95% of the time in previous elections the MOE held up....if not, then it could be a good indicator of either fraud or poor random sampling.

[TruthIsAll]

15. Exit polls have been and continue to be very accurate.

In the last three German elections, they have been correct to within .2%.

And they have always been extremely accurate in U.S. elections. At least until 2000 in Florida. I'm sure you have heard about Florida.

175,000 spoiled punched cards in Democratic minority precincts: 60,000 hanging chads and 110,000 double or triple-punched or butterflied. The vast majority of those were Gore votes. He would have won the state by at least 50,000 votes, if the votes weren't spoiled.

[Bread and Circus]

7. So....you are saying....it could still happen due to chance....

(channeling Lloyd Christmas from Dumb and Dumber)

[TruthIsAll]

16. Yes, it could still happen due to chance.

And the chances are equal to the odds that you would personally count every single grain of sand on earth.

[Bread and Circus]

22. Well I'm not going to Iraq if that's what you are suggesting.

[americanwhothinks]

9. Local Affidavits to Support Statistical Findings...

I am new here, and my posts are always on the bottom of a thread...ib other words, no one has ever read any of them! but in this case, i won't let this happen, so i'm reappearing in a new thread..

Anyway, the problem with statistical analyses is that they don't get the people-- ie, most people-- activated, excited, aware and protesting..

So, in response to Trudyco's suggestion (on another thread) that a LOCAL VOTE COUNTING effort be made to PROVE the fraud the statistics show....So the basic idea is: get people to sign affidavits saying, I voted for Kerry..and compare those numbers with the "offical" tallies--

I said--

local, grass-roots approach to democracy seems to be exactly what we need, as it asks Americans to revisit election day: I voted, and this is who I voted for-- ie, MAKE IT REAL! GET THE PEOPLE TO SAY WHO THEY VOTED FOR, THEN DO A LITTLE COUNTING... Get the folks from one precinct to come out and sign affidavits (using the poll books as proof of their voting) and see ONE precinct radically underrepresented in official tallies-- and you have the potential for a MAJOR LOCAL STORY that INVOLVES THE PEOPLE! The people then realize, my god, my vote vanished, disappeared! And then the PEOPLE are realizing what the statisticians can't make anyone see/feel-- the most intimate, direct realization of disenfranchisement.


And all it takes is for ONE precinct to have this experience, I think...People will then start waking up..the post November 2nd slumber will be interrupted... and once the PEOPLE wake up and realize, my god! the guy I voted for may have really won!

well, that's the beginning of the groundswell that WILL make MSM take note. That's all we need to get regular people OUT ON THE STREETS PROTESTING!!

[Old and In the Way]

13. Welcome aboard!

I've also been making the case for comprehensive canvassing of selected precincts with affidavitts.

As a legally binding document, I think it carries a lot more weight. Actually, it does a couple of things-

(1) Proves/disproves voter fraud. Either way, it's important for us to understand the truth.

(2) It shows that the Democratics are doing something concrete to support the rank-and-file. So far there's been precious little communication from the top, down.

I'd open it up to everyone who voted in the precincts. I assume solidly pro-Bush voters won't bother to participate, but if they were so stupid as to say they voted for Kerry, they'd only be hurting the Republicans. And, of course, I'll bet there are many Republicans who would be willing to say they voted Kerry....because they did and because they really think Bush is going to destroy this country.

While cavassing, I'd also re-ask the issues part of the survey. Seems that there'd be plenty of Democrats who voted on morals - lying to start an immoral war, bankrupting social programs, etc.

Seems like this should be the highest priority for the Democrats right now. Get some accepted unbiased polling organization to conduct a comprehensive survey of key precincts in Florida, OH, and where-ever....lets see if these results jive with the exit polls or the "unofficial" results.

[americanwhothinks]

14. Thanks! And you made me think....Local Affidavits, continued...

You made me think that one weird possibility is that there WERE a lot of republicans who voted for Kerry but who wouldn't sign an affidavit-- they won't stand up and say they voted for kerry because they're NOT democrats and they don't want anyone to know who the REALLY voted for! (in other words, the privacy of regular election is undone...and this could affect people!)Anyway, because of this potential problem, canvassing/affidavit approach must, I think, occur where people have traditionally voted democratic...they're not afraid of what any neighbors will think, in other words...they're in good company and not afraid to let others know!

and yes!! This effort-- asking the PEOPLE! to make and prove the case! is exactly what is needed-- to make concrete, to galvanize, to energize and encourage EVERYONE to revisit NOVEMBER 2nd.

Even I-- I who am certain this election was stolen!-- even I am beginning to waver, to lose hope and sight...and this MUST NOT HAPPEN! This is why MSM coverage is so critical...and this is why/how DU is so essential...

And this is why a GRASSROOTS, BOTTOM-TOP, EFFORT MUST BE MADE, AS YOU SAY, AS I SAY, AS TRUDY CO SAYS...

CAN YOU POST YOUR OWN LINE/"NEW THREAD" ON THIS???

(I am so grateful to/for all of you!!)

[Old and In the Way]

19. Certainly, we can expect no help from Bush Republicans.

Those Republicans that voted for Kerry, OTOH, may be motivated out of pure fear that Bush will take this country down the tubes. I wouldn't discount anyone who'd be willing to sign an affidavit either way. zIrregardless, I think we can make our case on the affidavits of registered Democrats.

Secret ballots are important...and they have been a historical cornerstone of our democratic electoral process. But we've only recently had the introduction of E-voting. We know 3 things about the process today:

(1) It is possible to manipulate the aggregate data with no paper trail.
(2) The equipment is owned by Republicans.
(3) Republicans are hell-bent on doing whatever it takes to maintain control of Congress and the Presidency.

When secret ballots are combined with no paper back-ups, it opens up the opportunity to commit massive voter fraud with no way of knowing.

At the very least, selected massive canvassing of districts that are suspect would at least assure ourselves whether the election was stolen or not....either way a known outcome is better than wondering if the election was stolen from us. And the course of action - we either have got to do a better job of selecting candidates and articulating our message....or demanding that the rightful person be sworn in as President in January.

[TruthIsAll]

26. Means (Diebold) , motive (power), opportunity (spineless dems)

.

[newyawker99]

25. Hi americanwhothinks!!

Welcome to DU!! :toast:

[americanwhothinks]

30. thanks, new yawker!!

[lowmanknows]

11. I know I'm new, but you need this info..................

http://www.firstcoastnews.com/politics/news-article.aspx?storyid=26823

This is odd. Say you wanted to rig an election. How would one do it? What would be the easiest way to do it, without involving too many people? A good old fashioned "crisis." Scare people away from the crime scene so the dirty business can be done. Everyone knows that if Kerry was going to win he would need OHIO or FLORIDA or both. Those are the only two that would need to be messed with to assure victory for either side. In Ohio there was the famous "Warren County Lockdown" because of "terror" threats. What about the less infamous "Bomb Threat" at the State Elections Office in Tallahassee the day before the election. I guess my point is this....... The only two places that needed to "fixed" (Ohio& Florida) both had a scare tactic applied to a government building the day of, or the day before the election. If they both had the same "crisis" on the same day it would look odd even to people who are not avid news watchers.
The State Elections Office in Tallahassee holds what? The main tabulating computers. It would actually be easier to hack in the main tabulating computers before the actual election because you can place "kernels" in place and never have to mess with it again, and you would still get what looks like "actual" results. The Warren County Courthouse holds what? The optical scan equipment that can be hacked into and votes can be messed with from as low as the local level and as far as the state and national level. By doing it this way you would actually have control before and after the election and you run a smaller chance in getting caught because your using to different modes at two different times.
To get this done all you would need is 2 or 3 people in Ohio and Florida to get this done. Not an army of computer hackers. Does anyone else besides me notice that there were means, opprotunity, and motive here? This even explains the exit polls......

Thank you,
Stacy Lowman

[trudyco]

18. The only thing I would add

Is that I would think you'd want a way to trigger the kernel to do its dirty deed and then self destruct or just to self destruct. If the vote was going Bush's way then they would trigger option 2 and there would be no fraud. If it wasn't he'd get on early on the MSM and announce something (like he'd won?) and they'd trigger option 1. Heck, he might have even had a coded message to pick one of many options and for various states.

To do the triggering they'd need networking capability. In Volusia it seems like they had to have network drives installed, the code enabled remotely??, and then somebody removed the drives.

Computer hackers would still be needed to write the kernal, trigger the kernal and possibly install it. They also need to record the real vote versus the fraud vote so the other parties would know how many extra ballots to switch in counties where fake homeland security guys or police or whatever did the groundwork. Maybe not.

[TruthIsAll]

20. Hard to fathom? Beyond comprehension? Impossible to accept?

Did you ever see odds of this magnitude? For anything?

I never did. I'm still reeling. It can't be.

But it is.

[BeFree]

21. What? ya don't believe in "Miracles"

The Miracle Balloting musta made it happen. Don't ya just feel great knowing such miracles can happen? Gives one hope... hope that we will win the lottery, hope we may one day find 10 M cash in a bag, hope I will be elected POTUS!

I imagine my inaugural speech:

"My fellow A-merry-klans, I know it is hard to believe the odds were something like a billion to one that I was selected, but it happened. Miracles happen, and this election proves it."

******************* Bull.

What are the odds the first 30 percent of the vote (absentees) would show as much as 15% difference from the last 70%, as was the case in so many NC precincts?

[TruthIsAll]

23. I believe in miracles. This is BEYOND a miracle.

tia

[TruthIsAll]

24. 200 TRILLION = THE NUMBER OF STARS WITHIN 100 MILLION LIGHT YEARS

200 TRILLION = THE NUMBER OF STARS WITHIN 100 MILLION LIGHT YEARS
http://www.anzwers.org/free/universe/virgo.html

[TruthIsAll]

27. These odds are equivalent to 1 Billion Bush TriFectas

He's a VERY Lucky Little Dictator.

[TruthIsAll]

28. In scientific notation, the probability is 4.9960E-15

Just in case you want to know.

[Peace Patriot]

29. 200 trillion...

...the size of the federal deficit when Bush Inc. is finished with us.

We'll need a new Department of Homeland Data just to house the zeros.

On another topic: This is it for the Dems, isn't it? The line has been drawn.

[Peace Patriot]

31. Another thought, or rather a question...

Freeman put the odds for the 10 out of 11 states he studied that skewered to Bush at 1 in about 700,000 (revised down from 1 in 250 million, I recall from his update). Can you say why yours differs from his (in English--I have poor math skills)?

Or maybe I missed something in other discussions.

[TruthIsAll]

32. Here are a few.

1. Freeman did not indicate the formulas that he used.

2. His original estimate was just for 3 states - OH,FL,PA, if I recall. My odds for Bush to win FL and OH based on exit polls were 1 out of 667.

3. I consider all 51 states in the probability calculation.

4. I calculated the MOE for all 51 states to get the corresponding standard deviation to determine the odds of Bush's tally exceeding the MOE - regardless of whether or not he won the state. This is input to the NORMAL DISTRIBUTION.

5. I then use the total number of states (16) which exceed the MOE (all for Bush) as input to the BINOMIAL DISTRIBUTION to determine the odds of that occurrence.

It's straightforward. My math is correct. No one has questioned it.

[mackdaddy]

33. Wow, Odds same as Bush National Debt! :-) N/T