What leading digit are they using for the analysis? I can't see it in any of the figures. I teach Benford's in my class, so I'm pretty familiar with it. It's telling that there is a bump in the middle, because most humans make.up data in the 4 to 6 range, but I still don't understand what leading digit they are using for the analysis.
because the most common number of votes was in between 1 and 2 hundred, the ward sizes are all somewhat consistent so this is a pretty bad application of benfords and a waste of energy. I ran the Clinton numbers from 16 and they showed the same thing, but when you look at the vote counts it makes sense, the democrats tend to get 4-600 votes per ward while trump got about 1-200 in each. Benfords should only be used when you have a mix of values with different orders of magnitude: ie some in the 10's, some in the 1000's and some in the 100,000,000's.
What leading digit are they using for the analysis? I can't see it in any of the figures. I teach Benford's in my class, so I'm pretty familiar with it. It's telling that there is a bump in the middle, because most humans make.up data in the 4 to 6 range, but I still don't understand what leading digit they are using for the analysis.
@MikePenceVP then why did Trump's votes follow it Benford's law to a tee?
because the most common number of votes was in between 1 and 2 hundred, the ward sizes are all somewhat consistent so this is a pretty bad application of benfords and a waste of energy. I ran the Clinton numbers from 16 and they showed the same thing, but when you look at the vote counts it makes sense, the democrats tend to get 4-600 votes per ward while trump got about 1-200 in each. Benfords should only be used when you have a mix of values with different orders of magnitude: ie some in the 10's, some in the 1000's and some in the 100,000,000's.