IS THERE JUST ONE MATHEMATICIAN OUT THERE WHO WILL CONFIRM OR REFUTE THIS?

[TruthIsAll]

The calculation of the odds that Bush's vote tallies in 16 states would all increase beyond the Exit Poll Margin of Error.

************* ONE IN 200 TRILLION ********************

So far, no one has.

So once again, I ask: If any reseacher or mathematician or statistician disagrees with the use of the Excel Binomial distribution function to calculate the probability, please do so.

In the initial calculation, I used the probability of 5% that a single state would deviate beyond the MOE as input to the Excel Binomial Distribution function. But that was the probability of a move beyond the MOE, regardless of whether the move was favorable to Bush OR Kerry. The odds for this: 1 in 4.5 billion.

In fact, what we really want is the probability that the vote would deviate beyond the MOE to Bush alone, which is exactly what happened. So that's why we use 2.5% and NOT 5.0% as our input probability. It's the Bush tail of the probability. We just split the probability in half, the tail that would go to Bush.

What is the effect of this seemingly small, innocuous change on our final probability estimate? It means that the probability that these 16 deviations could be due to CHANCE is EVEN MORE REMOTE.

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 calculated as P =1-BINOMDIST(16,51,0.025,TRUE)

P = 0.000000000000004996

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

Try it yourself in Excel.

Here are the odds for various scenarios that in N states, Bush's vote tallies would move beyond the MOE:

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

[Angry Girl]

1. link please? (that one's 404)

[TruthIsAll]

3. I just replaced the link with the actual text.

tia

[berniew1]

24. I'm statistician. Either an extremely unlikely event occurred or rigging

I'm a statistician and can confirm the following:
If you have 10 independent events with answer Y or N ,each with probability = 1/2, then the probability that all 10 are Y is 1/2 to the 10th power which is essentially near zero.

So if you have 10 (or etc.) events that are independent, the probability that they will all be Y (or N) or (B or K) is near zero.

If such an event happened, then either an extremely improbable occurance has happened or the events weren't independent (there was a connection or rigging between the events).

[MrUnderhill]

29. or....

Or sampling variation was not the only error in the data.

[TruthIsAll]

44. Thanks, MathGuy. IT'S ONLY 1 OUT OF 13.5 TRILLION. See his post below.

He found a slight flaw in the calculation. See his post below. I sure don't mind being corrected by someone who obviously has an appreciation for the simple math involved.

What irks me is when the debunkers - and you will find them in the thread - throw up bogus strawmen and attempt to fog the analysis.

I asked for a mathematical confirmation - positive or negative. MathGuy gave me what I was looking for. An honest evaluation of the calculation used. He essentially confirmed it, with the exception that I mistakenly input 16 rather than 15 to the Binomial Distribution function.

So the odds come down from 1 out of 200 trillion to 1 out of 13.5 trillion.

I expect to here from the trolls: "See TIA's math was off. Don't trust TIA's numbers".

Just wait.


Here are the revised probabilties:
N 1 out of
2 3
4 26
6 597
8 26,885
10 2,098,096
12 262,019,924
14 49,652,431,051

16 13,544,660,533,445

[tridim]

2. I'm sure there are some freaks out there who will question

why you didn't add Gawd to the function.. As everyone knows, Gawd cancels out impossible odds.

[unblock]

4. you're asking the question to the wrong field

it's obvious that there's SOME systemic bias favoring shrub.

the question is, what could account for that bias? you've made many compelling arguments that dumb luck cannot account for it.

however, there are potentially many effects that could account for it, and fraud is just one. i don't have any great suggestions for what the other ones might be, but i think it's clear that the burden of proof is on the banana republicans to prove that there's a reasonable explanation for this bias.

simple assertions that morning voters are more likely to be the kerry voters and/or more likely to talk to pollsters is not enough. they need to prove these things.

[Old and In the Way]

5. Hi TIA

Check this thread out.

http://www.democraticunderground.com/discuss/duboard.php?az=show_mesg&forum=203&topic_id=107127&mesg_id=108130

I wonder if there are enough '0's to calculate the probability of voting problem incidences that break in Bush's favor?

[TruthIsAll]

10. Here is your answer. We MUST get the DATA on the "gliches".

Here are the probabilties, assuming that N gliches went to Kerry.

In other words, out of 1000 gliches, the probability that 600 would be in favor Bush is 1 out of 7.3 BILLION.

N.......Probability
10 2.4833E-278 < FORGET ALL THESE
20 3.2341E-260
50 9.3185E-217
100 6.7017E-162
200 8.22499E-86
300 8.83284E-38

400 1.36423E-10 < 1 out of 7,330,131,113
In Excel: P =BINOMDIST(400,1000,0.5,TRUE)


450 0.000865268 < 1 out of 1,156
475 0.060607133 < 1 out of 16
490 0.273986373 < 1 out of 4

500 0.512612509 < 1 out of 2
THAT'S WHAT YOU WOULD EXPECT TO OCCUR

[hughee99]

6. I think it's more than that...

I'm not exactly sure what the odds of the this are, but it would be based on the assumption that the exit poll was conducted "correctly". Anyone could go out and ask 2000 people how they voted and claim that this will be the poll result. They reason that just anyone can't do it accurately is that in order to be able to truly extrapolate the results, you need to have an accurate representation with accurate proportions of voters. If the exit polling conducted as accurate, then the odds are astronomical (though I believe that 1 in 200 trillion is a little high). If the exit polls underrepresent fundies and other RWers, then the exit polls wouldn't necessarily have any correlation to the actual results.

The bottom line is that if you believe that the exit polling actually represented the demographics of the actual voters, then I think you have to conclude that * most likely stole the election. Either that or this was the "1 in 200 trillion", which I doubt.

ON EDIT: Looking at the math, perhaps 1 in 200 trillion isn't high.

[suston96]

8. Key precincts

The basis of exit polling is taking the polls in "key precincts".

I believe this means that the popluation sample in a "key precinct" is very stable and does not have excesses of incidentals or variations as a poster cited above.

Here is the point: exit polling has worked, and quite well, generally, before this election.

In Florida 2000, exit polling said Gore won there. But he didn't and that trashed exit polling that election.

But you see, as an example, some 1100 people in that county with the infamous butterfly ballot voted for Buchanan when they thought they voted for Gore. But when polled on exit, if they were polled, they would have said that they voted for Gore. Gore did not get those votes.

Who knows what else happened in other counties that year?

The election in 2004 attempted exit polling again and once more, exit polling ran into a perfect storm where none of their results from key precincts were matching with machine tallies.

So, we trash exit polling again, right? It couldn't be the machine tallies, right?

Right!!!!!

[hughee99]

22. Trashing exit polls???

I'm not trashing exit polls, but why should I be skeptical of the machines ONLY? Is it impossible that both the polls and the machines are wrong? Is it not also possible that exit polling doesn't understand the demographics in what they've concidered to be key precincts as well as they claim? I'm not saying that they are wrong, just that they could be wrong.

IMHO the election was probably fixed. I DO NOT believe that you can demonstrate this by using exit polling results unless you can PROVE that the exit polling was as accurate as it is claiming. Unfortunately, I DO NOT believe you can do this without knowing key information about both (A) the participants in the exit polls (which is known) and (B) the same information about all the voters (which is not). The process of picking a "key precinct" is not an exact science.

If you want to prove fraud, you are going to have to do this by finding specific examples of vote stuffing, rigged machines, voter intimidation, not by just comparing potentially suspect polling results with very suspect actual results. Right?

RIGHT!!!

BTW, Welcome to DU.

[TruthIsAll]

33. You believe 1 on 200 trillion is a little high?

Then you must believe that
1) use of the binomial function is not an appropriate model
2) Bush tallies did not deviate beyond the MOE in 16 states or
3) the .025 probability that a state tally would deviate beyond the state exit poll MOE to Bush is not valid.

It has to be one of the three. Which is it?

[hughee99]

42. IF you read the WHOLE thing...

you see the end where, on edit several hours ago, I say that after seeing the math, perhaps its not high. I was not able to get to the link when first posted, and my initial statement was posted before you posted the math. After seeing the math, and I believe the binomial function would be appropriate in this case, I added that it may not be too high...

So is your only issue with me, that I was wrong in initially assuming that 1 in 200 trillion would be too high, without seeing the math?

ON EDIT: Perhaps I should have stuck to my initial instincts, or actually done the math myself. 13.5 trillion sounds right

[mdhunter]

7. My qualifications aren't that hefty...

I'm a graduate student in a Masters of Public Health program with some coursework in biostats, but your analysis looks spot on to me.

In my world your question reads, what are the odds based on chance that out of 51 trials, 16 would be sucessful if you expected .05 (for any variation) or .025 (for a Bush "success") to be successful at the start. I can confirm that number is exceedingly small.

One small point, though it doesn't really change the impact of your analysis. It may not be the case that a Bush success and a Kerry success were equally likely. As someone mentioned above, there might have been some systemic bias for Bush. The result of that would be, while .05 would be the chance for any deviation outside of the MOE, it may not end up being .025 Kerry and .025 Bush. It may well be .01 Bush and .04 Kerry. Who knows? When stating your findings I would range it between the most conservative estimate, .025, and .05.

On edit. I was just reading the post above on the veracity and validity of the exit polling results. I haven't seen the raw exit polling numbers, nor do I know their methodology. However, even if we assume the exits polls were twice as likely to be wrong (90% confidence) then the worst you can do is your original number - assuming a conservative .05/.05 split. That Bush won by chance alone is still about as likely as Dennis Miller's becoming funny again any time this century.

[libertypirate]

9. Hey anyone want to look at the full New Mexico Precinct numbers

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x107478

[On the Road]

11. Let's Take New Hampshire

and the precincts with large deviations between exit polls and the official count. Eleven wards were recounted, and the results confirmed the official vote over the exit polls.

What is your interpretation of those results?

[TruthIsAll]

20. You asked me exactly this question on another thread. I answered you.

Why don't you just go back to that thread?

[Emillereid]

30. Could you provide the link to the discussion of the NH results

you're referring to? The New Hampshire results keep messing my understanding of this too. Thanks in advance.

The aspect of the election results that keep haunting me is not any one analysis of some oddball deviation from expected results but how the vast majority of the anomalies all seem to break in the Bushit's favor -- that's just too anomalous to believe. When the theory (Bush won the election) doesn't fit the data, it's time to throw out the theory.

[TruthIsAll]

48. Here is a LINK

http://www.democraticunderground.com/discuss/duboard.php?az=show_topic&forum=201&topic_id=6005#6557

[Emillereid]

53. Thanks!

[Zan_of_Texas]

12. regardless

Regardless if you can get a mathematician to say it's true, as a smart person with a head for statistics, I have to say that my first reaction when I see that number

ONE IN 200 TRILLION

is to toss it into the trash!

One trillion is so large as to be nearly incomprehensible -- the specificity of 200 trillion just sounds like hyperbole.

Just my opinion, but I think using that number, true or no, does not bolster the argument for the average reader.

[TruthIsAll]

19. Zan, I felt the same way: people would not believe it.

But I would rather be mathematically correct than worry about what people might think. Besides, 1 out of a million or a billion or a trillion. However you slice it the answer is: NO WAY.

Let's let the professional statistical researchers and mathematicians have their say on this.

Let's get hundreds of thousands of college math students to do their homework on this one calculation. That's a real grass roots operation.

[keepthemhonest]

13. keep at em TIA

you are and will be heard.
we are so lucky to have people with your mathematical ability on our side. I know the people on their side with mathematical abilities are just ignoring the numbers all together( they have to.)

[badc0der]

14. Think systemic problems

This makes a good story, but the exit polls were all conducted by the same company. If they have a flaw in their sampling methods than it could easily be reflected in many states.

[milkyway]

15. The obvious repug refutation of this is that the methodology of the exit

polls was flawed. The work of yourself, Freeman, and others has proved beyond any doubt that it wasn't just random mathematical coincidence that resulted in the discrepancy between the exit polls and actual tallies.

The repugs are left to argue that it's not the methodology of the vote counting that was fucked up, it was the methodology of the exit polls. This, of course, is highly debatable.

[skids]

17. The repugs can also point to...

... the adjusted exit poll numbers. Which we know are trash.

But if they did point to the adjusted numbers, it would be nice if you folks doing this calculation would also run those -- even after the adjustments, though the red shift may be within the margin of error, they are still oddly bigger in the battlegrounds.

That way they can say: the polls were screwed up, and they adjusted for that, and you can toss back: well even if we use those numbers, the odds are still X.

(P.S. You can add a bit of legitimacy to the numbers by assuming a bias in the battleground polls. To do this, find the average shift, and then recalculate the shifts using the average as the base. You won't get such spectacular numbers, but you will be able to say "we accounted for systematic bias in the raw polls and still the odds are X.")

[skids]

18. duh...

Oh wait -- are we actually supposed to be talking about exit polls here. 600 glitches as determined by what?

[Gregorian]

16. There was a statistician who posted not long ago.

A phd husband and wife. I don't remembe the name. But they were offering their services here.

[Yupster]

21. Your analysis assumes randomness

When you see a number in the trillions, obvious you are not dealing with randomness.

For instance the odds of rolling a 6 on a die is one out of six of course.

If you roll a die 5000 times and it comes up six all 5000 times, you are no longer dealing with a statistics question.

There's either something wrong with the die (the exit poll was wrong) or something wrong with the person writing down the results (the votes weren't recorded properly).

Just to keep the analogy going, you could go back and watch the person recording the results for 25 rolls on videotape, and you'd see sure enough the die did roll six all 25 times and he wrote them down properly. (This would be the analogy to the test of the New Hampshire precincts which matched the recorded votes.)

Well, in that case, you could say "gawd, I can't believe it rolled 5,000 sixes in a row", but someone else would just throw the die away and buy a new one because obviously that die (exit poll) was no good.

[IndyPriest]

25. Exit polls BETTER than random

If I read your post correctly, you're using the same logic to defend the results as has been often used: there must have been something wrong with the exit polls.

If so, you must deal with the fact that exit polls are designed to IMPROVE on randomness. Survey polls typically have a MoE of between 2-4% depending on the number of persons surveyed. Exit polls, by contrast, have a MoE about HALF that BECAUSE they are designed to screen for the errors that are typically part of survey polls.

The argument that TIA has been making for weeks (months?) is that it is IMPOSSIBLE for the exit polls were as wrong as the actual vote pretends they were.

[MrUnderhill]

28. Hardly "impossible".

No offense. But it would obviously be a lot easier to mess with the exit polling than it would be to mess with the actual election results. Either is "possible".

[IndyPriest]

39. On the contrary...

And no offense taken. Given what we know about black box voting, it would be MUCH easier to mess with the "actual" election results than to mess with the exit polls.

[TruthIsAll]

35. Thank you, thank you, thank you.

tia

[IndyPriest]

40. My pleasure. n/t

[Yupster]

49. Indy - do me a favor?

read this long article from the New York Times magazine.

http://www.nytimes.com/2004/11/21/magazine/21OHIO.html/?ex=1102222800&en=ae18d4a994ac94d5&ei=5070&oref=login

It's the most interesting one of I've read about the election.

A reporter followed around the Ohio chairman of the 527 "America Coming Together" all through election day.

I think there's this view that people have that the exit poll showed Kerry up big abd then whammo the vote came in and he lost. I don't think that was the case. It wasn't for this insider anyway.

Yes the first exit poll batch of data came out and Kerry was up 4 % and this guy was elated. But, the next batch to come out was not favorable, and each successive batch that came out was tighter and tighter.

In another scene, the guy decides to drive to some heavy Republican precincts and check the paper the election workers post and update throughout the day on how many people had voted so far. That took away the last bit of joy the guy had as the precinct was having a record day.

Another interesting scene was buying $ 3,000 worth of McDonalds for people waiting in line to vote at a heavy Democratic precinct. When he gets there with the food, he talks to the local precinct chairman who shocks him by telling him that many of the people on line are Republicans. How can this be in this precinct? The chairman pointed him to a new ex-urb neighborhood and said see that area? It didn't exist four years ago. Now these Republicans from there have come to vote after work.

By the time the polls closed, this guy knew Kerry was in trouble in Ohio, though he was still hopeful.

Anyway, my point is that some people here seem to think that it's a straight story of Kerry up big, everyone knew he would win big, and then he lost and everyone was shocked.

At least to this one leader on the ground in Ohio, a Democratic insider, it was not at all like that at all.

[Yupster]

51. Just to reply to my own post here

a scary part of this NY Times article is how well it says the 527's did.

Well, Bush was caught flat footed in the 527 wars this time. He thought they'd be ruled illegal and therefore the Republicans were very slow getting theirs organized and their efforts were dwarfed by Move On and America Coming Together etc.

Don't expect that to happen again.

If the rules of the game is unlimited donations from wealthy donors, the Republicans will be up and running in plenty of time next time, and they will have overwhelming resources compared to the Democratic groups.

You think Rove couldn't have raised a billion dollars if he thought he'd be allowed to this time? Well he, or someone else will next time.

[RevolutionStartsNow]

55. Great (sad) article, thanks for posting it --

Particularly interesting to me, as I worked in Ohio for ACT (in Dayton).

The one point the reporter missed though was how much easier it was to vote in "exurbia" than in the cities. I have no doubt that this was part of the GOP strategy.

I can also confirm that there was almost no sign of Repub GOTV effort where I was. I never saw a single door hanger, precinct walker or any other sign of feet on the street from their side. I also mistakenly took this as a positive sign for us.

[Nye Bevan]

23. Hi TIA

I have been following your posts with interest and your plea for a mathematician has finally converted me from a lurker to a poster!

I think your analysis is very close to being correct, except for one minor change. I would compute the probability P as

P = 1-BINOMDIST(15,51,0.025,TRUE)
P = 0.0000000000000738298311376

which would give odds of 1 in 13.5 trillion, compared to your 1 in 200 trillion.

The only change I made was to replace your 16 with 15 in the BINOMDIST function. The reason for this is as follows. Here is the specification for BINOMDIST from the Excel help:-

BINOMDIST(number_s,trials,probability_s,cumulative)
Number_s is the number of successes in trials.
Trials is the number of independent trials.
Probability_s is the probability of success on each trial.
Cumulative is a logical value that determines the form of the function. If cumulative is TRUE, then BINOMDIST returns the cumulative distribution function, which is the probability that there are at most number_s successes; if FALSE, it returns the probability mass function, which is the probability that there are number_s successes.

Let's define an "anomaly" as a state where the Bush vote exceeds the exit poll margin of error. We are assuming that the probability of each state being an anomaly is 2.5%. Our goal is to determine the probability of the following event: that the number of anomalies is AT LEAST 16; that is, 16 or more, out of 51 trials (each trial corresponding to a state).

Since the BINOMDIST function returns the probability that there are AT MOST N anomalies, we cannot use it directly to answer our question. Instead we need to check the probability of our event NOT happening and subtract this probability from 1. Our event NOT happening means that the number of anomalies is 15 or fewer, i.e. AT MOST 15.

In other words: we are looking for the probability of 16 or more anomalies. If there are NOT 16 or more anomalies it follows that there are 15 or fewer anomalies, i.e. AT MOST 15 anomalies. So we need to determine the probability of at most 15 anomalies and subtract it from 1, which explains why 15 is the correct parameter to use in the BINOMDIST function.

It is clearly still safe to say that even with the reduced odds of 1 in 13.5 trillion this event cannot be explained by chance alone!

[keepthemhonest]

31. alright a really smart newbie.. wlcome

this really makes me wish i had taken more than stat 101 in college

[Nye Bevan]

32. Thanks (nt)

[TruthIsAll]

37. You are right. I just confirmed it. It's 1 out of 13.5 trillion.

Flip a coin 4 times:

Prob (zero heads) = prob (4 tails) = 1/2*1/2*1/2*1/2= 1/16

Prob (at least 1 head)= 1- prob(4 tails) = 15/16

= 1 -BINOMDIST(0,4,0.5,TRUE)= 1 -.0625

[faithnotgreed]

54. hi mathguy - welcome to du. were glad youre here

and i couldnt help but notice the avatar. i grew up in ny and in my family, the mets were the team. this was early 70s with felix millan, torre, seaver, matlack

we still hold on even when theyre ragged dont we?? ah, the mets....

[Yupster]

56. hey faith

another Mets fan here too. First game at Shea in 1965. Still Have my scorebook.

Pitcher for one game I went to in 68 was Rayan. Guess he never made it. Also Willie Mays leading off and playing first base in 1972.

If you're still interested in the old teams check out www.whatifsports.com. My favorite website.

Sorry - I mean second favorite of course.

[faithnotgreed]

57. hey - thanks for that. i love the older teams

and you got to see ryan and mays. excellent/

ill go check out that link now. hope i dont get addicted to it like this place

thanks again.

[Yupster]

58. Check out the forums

"baseball classifieds" and "baseball owners" forums. But no politics allowed.

[MrUnderhill]

26. I'm happy to help you out...

since you've posted this same thread about a dozen times you obviously like the topic. In all caps as well.


The problem isn't your math. It's your assumptions. Some were too generous (as you've said before), but too many factors are just plain left out.

[TruthIsAll]

43. I NEVER said the assumptions were too generous. You did.

I like posting in all caps. If I feel I have something important to say, I will do so.

You did not provide one bit of useful information in this post.
Just a smarmy left-handed rebuke. Of no substance.

[MrUnderhill]

47. Perhaps you misunderstand

By "generous" I mean that the statistical MOE is likely to be smaller than you were originally assuming... and thus it's likely MORE states fall in the category of "outside the MOE". You HAD said as much on previous threads... though I don't know if you've changed your presumed MOE.

"Didn't provide one piece of useful information?" Only if you define your terms to mean "something that supports your theory" - I'm not "rebuking" you, merely pointing out that there are numerous possibilities not accounted for in your calculations (as there were in your pre-election stats - little wrong with the math... but too far off in the assumptions)

[TruthIsAll]

60. OK, I'll give you that. The MOE's are really much smaller..which means

a much higher (if that's possible) probability of fraud.

Exit polls are much more accurate (at least double) than standard polls. So, yes, I was being conservative when I used traditional standard polling MOE calculations based on sample size. But I did NOT want to make any presumptions.

If that was your point, than I misunderstood and apologize for the comment.

ONE OUT OF 13.5 TRILLION IS HIGH ENOUGH FOR CIRCUMSTANTIAL EVIDENCE OF FRAUD.


TIA

[Ranec]

27. Bias!

My reaction is BIAS! The exit polls are biased! They have tried to correct this, but the magnitude of the bias is evident at the state level. I mean this in the sense that they tend to predict higher vote counts for Kerry. There are two sorts of explanations:
1) The poll over-sampled Kerry supporters in a systematic and unforseen way.
2) The election counters undercounted the votes of Kerry supporters.

Either explanation is equally likely at this point.

We need to demand more information! Ask for recounts, and clarification from the pollsters.

The outrage I feel is the amount of secrecy surrounding the process. A democratic country must have an open government. The election procedures should be transparent. This is not a partisan issue. Votes should be counted.
:grr:

WARNING: Don't be wedded to the exact probabilities. Just note that the exit poll results are unlikely. The exit polls didn't pick a random group of 2000 people out of the population. They have a very structured sample that first chooses precincts, and then samples within precincts. This effects the probability calculations, and increases the probability of error somewhat.

Without knowing more about how the exit polling data is generated (as well as to what sort of weighting is being done), I would be slightly worried about these calculations.

[TruthIsAll]

34. No need to worry about the calculations.

We should worry about the fact that exit polls have traditionally been very accurate.

Why should they not have been this time?
Did attitudes towards pollsters change?
Were Republicans less likely top speak to pollsters. If so, why?
Were women polled earlier than men? Aren't more women working than ever?
Were the pollsters not trained well enough- in 16 states?
What about the states which were well within the MOE? Why no problems there?

Remember FL 2000? Remember the 175,000 spoiled punched cards, 75% of whom voted for Gore and who were exit p
Just asking.

[understandinglife]

36. For consideration 'truthisall' and everyone

We now have the WA State Dems financing a state-wide, every ballot, manual count because of the Governor's race.

WA State had a 2.9 'red shift'

Comparison of the manual count ballots to each of the tables in the each county 'central tabulating system' could be illuminating to say the least (as opposed to selecting a few precincts, as was done in NH).

As some of you know, I've hypothesized that WA State may be the gotcha opportunity -- gotcha in the sense that I think it reasonable to postulate that someone was just a bit too greedy. Said another way, if I were going to do what might have happened, I would have Bush emerging with ~ 1M margin and with, at most 3 'red shift' states.

Greed does frequently kill and in the case of WA State I do not think anyone could have ever expected a Rossi, Gregoire difference of 42 votes being the trigger to a 2.9 million ballot manual count -- the irony that a race that involves the current State AG might be the one that truly 'opens the box.'

Key issue -- major push to have the Pres/VP checked on each ballot as the Gov is checked. How to make that happen??

Peace.

"Did Bush Know?"

[suston96]

38. Exit polls are biased?

Of course they are.

In fact, they take exit polls in "key precincts" because they know that the results will show their "bias" toward the candidate who received the most votes in polls taken in those key precincts.

Here's a thought: if exit polls are so, uh - undependable - why did we sell or pay for an exit polling process for the recent election in the Ukraine - And....

....Why did we swear - Colin Powell and Senator Lugar - both Republicans - why did they swear that the elections there were crooked because the final tallies did not agree with the exit polls.

Why is everyone saying that the exit polling in the US cannot be trusted, but the exit polling in the Ukraine is just fine.

Now help me out, here. Which exit polling system are we supposed to believe?

The one we used in the US on election day? Or that Ukraine exit polling?

The former says that the election in the US was just fine. The latter says that the election in the Ukraine is NFG.

[understandinglife]

41. Fully aware of the 'irony' of the Ukraine, USA ....

....what I call 'exit poll perfect storm.'

The way to put considerable pressure on the need for a full investigation of ballots versus the tables in the central tabulating systems' databases is that that comparison is likely to reveal that the USA 'exit polls' didn't suddenly go crappy -- something else was responsible for the 'miracle of 211.'

Peace.

"Did Bush Know?"

[suston96]

45. Exactly

Blaming the exit polls is like blaming your yardstick when you screw up a measurement.

Exit polling is a science. It worked for the longest time.

Then somebody figured out how to hack the results of the voting and tabulating machines.

That throws off the exit polling.

So, do we throw away the yardstick?

[understandinglife]

46. Nope -- we put the hacker in jail....and, just to be sure....

....he's not lonely, we send along his boss ;-)

We will prevail.

"Did Bush Know?"

[Carolab]

59. Citizens can challenge

I believe. I thought I read that 5 citizens brought this forward. (Someone should check this out but I thought that's what I read on BBV's website tonight). If so, why not have them ask for the presidential recount?

[jfern]

50. I get 1 in 14 trillion

51 choose 16 * 0.025^16 * 0.975^35

[TruthIsAll]

61. I did your calc and got 1 in 14.5 trillion (no round). Which is correct?

Is it 1 in 14.5 trillion or 1 in 13.5 trillion?
Maybe we should just split the difference and say its 1 in 14 trillion.

For the sake of mathematical purity (and the trolls), we had better get to the bottom of this discrepancy. Hehehehehehe.

"51 choose 16" is the number of combinations of 51 states taken 16 at a time.

Use the Excel combinatorial function:
COMBIN(number,number_chosen)

Returns the number of combinations for a given number of items.
Use COMBIN to determine the total possible number of groups for a given number of items, where:
-number is the number of items.
-number_chosen is the number of items in each combination.

Here is the calculation sequence:

1. Number of combinations of 51 states taken 16 at a time:
C = COMBIN(51,16) = 7.17452E+12

2. Calculate the 51 joint probabilties (16 beyond MOE, 35 within) for each combination:
JP = 0.025^16 * 0.975^35 = 9.59846E-27

3. Calculate the total joint probability (all combinations):
Probability P = C*JP = 0.0000000000000688643
or P = 6.88644E-14

4. Calculate the odds: 1 / P

***** 1 in 14,521,300,254,785 ******

[jfern]

62. I did it ballpark

It's ballpark 1 in 14 trillion.

For the exact calcualtion, you need a sum.
That's why some people were getting 1 in 13.5 billion, and you got 1 in 14.5 billion.

1 in 13.5 billion is the true answer we're interested in.

[Patsy Stone]

52. House Dems Request Mitofsky Exit Poll Data -- DU Thread

http://www.democraticunderground.com/discuss/duboard.php?az=show_mesg&forum=203&topic_id=109716&mesg_id=109716