Maybe the correction for a misunderstanding: when I said that I am done with what I have boldly called the absolute mathematical proof, I meant that I do expect to search for other "cheat" detectors. The oscillation detector is like mass spectroscopy. I do not need more or better. But I will publish more results obtained with them.
affa, you might want to add them to the beginning of your summary thread as EDITs, so that the substantive analysis remains frontloaded.
Ok, back to investigation in NH (honestly guys, if not debunked, this could become a book!).
A smart alec mentioned in a post that he wanted to see the table where I have shown the layman's probability of Romney's score to be "natural" in various SC counties for other candidates. Smart, fair and instructive, so wish granted.
Let's start with Coots, NH. On all my analysis, it has never exhibited any anomaly. Here is the oscillation doodle.
Remember, if the score of the candidate converges naturally towards the final score as you count more and more ballots, it will zig-zag above and under the 50% line all the time. Boy, Coots illustrates that very well: mad criss-crossing. Now nifty maths give me the following table:
The number in the table are a tabular form of the doodle chart. See how everyone's probability to reach his final score looks normal? Gingrich is drifting a bit low at 22%, but comes back immediately.
Now Merrimack, NH. Doodle chart: strong mathematical anomaly. Where are Romney and Paul's oscillations gone?
Are they gone a little, or are they entirely gone? Well the tabular form tells you the probality of that:
See the difference with Coots
Paul is vampirized by Romney here.
Possibly something funny going on in the last precinct in favour of Huntsman, but it's only 1 data point.