Thursday, September 02, 2004

More on the Presidential guard: Scary guys The case of the so called elite guards of the President who were part of his first ring of security becomes apparently more clear with a home video of the forty soldiers made the day before the recall referendum. In the video, the soldiers are shown in a room decorated by a picture of Chavez and one of Simon Bolivar. There is a table in the middle covered with ¨No¨ signs a flower arrangement and a symbol of the GRAT unit. Their leader speaks out. The speech expresses love for Chavez and what he stands for. However, they say that those in the Casa Militar (in charge of presidential security) are shielding Chavez from the truth. They make accusations of corruption and ask Chavez to do for the Casa Militar the type of cleansing he did at PDVSA. They tell Chavez that this was the reason why tried to get into his office, they wanted to explain to him what was going on. The tone is eerie, recrimination, love and fundamentalism all in one. They tell Chavez they are not paid by the opposition. All of the soldiers join each other by the forearms and sign what is supposed to be the chant of their unit which includes dying for their country, dying for their leader…   Scary stuff, these guys sound like extreme fundamentalists. Trying to get a copy of the video to post. 8:34:52 PM    comment []

On Mathematical studies of the recall vote and fraud: Part IV A Prof. From Princeton and two from John Hopkins have done a calculation similar to that by Elio Valladares that I posted yesterday. Their conclusions are found here in English and while they are very similar quantitatively to Elio´s they differ in their conclusion. They did a very similar calculation to Elio´s. but 1238 simulations only, compared to the 10,000 Elio did. Their results are very similar, they found the distribution had an average of 360.90 for the Si coincidences and 317 for the No coincidences. This means that the occurrence of the No ones is reasonable, but the one for the Si coincidences is less likely.   However, the authors feel comfortable that 2.3 standard deviations away from the mean are feasible and thus feel this proves little. Elio´s results are much more conclusive, it is unclear if it is because he made almost eight times more simulations, but in the interest of discussion and comparison, here are the results.   I continue to believe that simulations at the Center level would be more conclusive than these, as there seems to be a high concentration of coincidences at the center level.  8:33:11 PM    comment []

Looking for ideas: Some interesting graphs   Jose Huerta sends these two plots with the distributions of machines with Si and No votes. I have no clue if it means anything, but why arethey so different? Ideas welcome Distribution of the number of machines with the same number of SI votes Same for the NO Votes 8:31:40 PM    comment []