Flipping the vote against Ron Paul in South Carolina?

[cachemaster]

Can we hypothesize what will happen today in Michigan and Arizona and look for this vote flipping? Can we somehow use this additional proof?

 

[bcreps85]

Wow. Well done Liberty...would like to see someone take a stab at debunking this. I got nothing. The hint of doubt that I was still holding onto just shrunk significantly.

 

[bbwarfield]

wow..... Im amazed


as a side note.... the last of liberties mass post was #999 so congrats! you pushed us above 1,000 ; )

 

[cjm]

Wow. Well done Liberty...would like to see someone take a stab at debunking this. I got nothing. The hint of doubt that I was still holding onto just shrunk significantly.

Post #981

 

[cachemaster]

I have a question: How well will only statistical data hold up when presenting this? Is there anything that can be done to find physical evidence?

 

[tremendoustie]

This is different from the previous work. The following argument is mathematically stronger than the previous ones by several orders of magnitude.

Counting votes in a ballot is like taking marbles fron an urn until you've got them all out. Take an urn with 10 marbles, 5 reds 5 blues. You draw the 1st: probability if it being red? 50% Let' say it is red. Pick the second. Probability of being red? 4 reds left in there, so you know it is now 4 chances out of 9.

Is this assumption correct, though? Aren't these cities made up of many wards, which may differ politically?

For example, if more rural wards report first, with more urban wards reporting later, it would be expected that candidates who are popular in rural areas would loose percentage points, while those popular in urban areas would gain.

That is, perhaps it's not like drawing marbles from an urn.

 

[Aden]

Hey guys, I have been following this thread off and on since it started. I am too busy IRL to keep up on it though. Hey Liberty, can you make a blog post here on RPF, or somehow put all the latest information together in one spot for us to read? Or maybe you guys can just tell me what post #s are the most important to read, to know what is going on with this thread. Also, what are the links to the data sets you guys are using? And, when somebody writes this up can you please state where to get the data sets or any other information? I want to email the writeup to all the math professors on campus.

 

[Liberty1789]

In layman's terms

 

[bbwarfield]

Is this assumption correct, though? Aren't these cities made up of many wards, which may differ politically?

For example, if more rural wards report first, with more urban wards reporting later, it would be expected that candidates who are popular in rural areas would loose percentage points, while those popular in urban areas would gain.

That is, perhaps it's not like drawing marbles from an urn.

these are not based on time reporting but by precinct size. Alot of the demographic switches are not significant enough (less that 2% of the vote was minority even though some precincts were 92% black)

this is showing a clean statystical anamoly that is not explainable based on demographics

The urn thing is an example and is not really 1:1 .... the analogy is the weakest part of the argument... but the data is very very strong

 

[bcreps85]

Post #981

Although these are sorted by voter turnout, voter turnout does not really have much to do with any geographical data, since voter turnout is relative and not tied to area population. Also, since polling data indicates that Paul gets pretty much the same amount of support in different areas sort of affirms this. The chosen method should be random enough for this test.

There are two real keys here. First, there is a huge number of statistical anomalies. If we saw them once, we'd shrug it off...but we see it constantly and only in favor of a single candidate. Second, if the astronomical chances of the first reason were not unlikely enough, we now have to compound that with the chances of a near 1:1 vote projection flip based on an acceptable sample, and every number in Liberty's last post above becomes "beyond Excel"...

 

[Mark37snj]

BOOM goes the Dynamite!!!

EDIT: Mitt Romney, the man whose Magic is beyond Excel!

 

[cachemaster]

can we compare this to probability of the other candidates by county as well?

 

[amonasro]

 

[bdmarti]

From the start I have found these graphs to be fascinating. I want very much to understand what it is that is happening here. The "flipping" algorithm fits well but I don't want to don a tinfoil hat and look like an idiot.

If there is a reason that larger vote precincts would vote in a higher percentage for Romney, we need to know what that is. Has early voting been eliminated?

I like the last batch of graphs, but I don't find it convincing because there could, possibly, be a reason that the large precincts vote differently and thus the samples aren't random in the same way picking marbles from a jar would be. Clearly if you lined up precincts first by ones Paul won, then Santorum, Then Gingrich, then Romney you'd end up with bizzaro looking graphs and you'd have astronomical odds of ending up at your end percentage, just like Romney does in that last series of graphs. yet, when we look at the precincts by size, that is kind of what we are doing...we know that Romney wins the big precincts and Paul wins the small ones. This will have the obvious effect of having strange graphs like what was just produced.

I do not claim to know why Romeny would so dominate large precincts. I do think the graphs look fishy and I want to understand them better. I don't however currently think the last batch of graphs are a slam dunk, because the ordering of Paul wins vs Romney wins precincts is too obviously flawed compared to a random sample. Since we know the data shows Romney wins the big precincts, this would need to be eliminated as a factor before I think that last set of graphs means much.

I wish I was more skilled with statistics or had better debunking ideas. I'm glad for the people in this thread and the work they are doing though.

 

[rb3b3]

Like I said in a previous post, I don't know statistics to save my life, but I do know that if u put 10 marbles in a bag 5 of which are red and 5 of which are green, I know in the end if u pick 10x 5 will be red and 5 will be green hahahahaha........ The most important question I have is, will this all get swept under the rug by the establishment??? Judges, media, attorneys???????????????????

 

[chri5opher]

//

 

[Aden]

We need demographic and socioeconomic data for each precinct so that we can run cross tabs and look for relationships.

 

[Mark37snj]

can we compare this to probability of the other candidates by county as well?

Well only results where the algorithym did not take votes from someone would/should show realistic odds. All voting stats for Ron Paul where he had votes taken from him would also show crazy mathmatical odds. The only odds that should show normal would be those where the algorithym did not TAKE or ADD votes. So impossible odds for Ron Paul and others would also strengthen the arguement for fraud.

 

[Mark37snj]

We need demographic and socioeconomic data for each precinct so that we can run cross tabs and look for relationships.

No we don't. Plz read thread.

 

[Liberty1789]

Just a quick word. You can resort along any demographic parameter you like. Rich/poor? The rich deciles will oscillate, the poor will oscillate. Rural/urban? will oscillate. Black/white? Will oscillate. Gay/straight, fat/lean, whatever. Oscillation will be there. Step 4 is about the total, binary disappearance of oscillation. And that is impossible naturally, which ever way you slice the electorate.