Thanks, but it answered only one of my three questions, and I shouldn’t have to go to an FAQ to find it. The whole point of a chart or graph is to show data in a concise and meaningful way, or at least so I think.
I’m sure the data on this site is well thought out and meaningful. It’s just that it’s presented badly for a non-statistician. Most people don’t know what “regression” means in this context, for example.
Maybe the target audience are the statistically sophisticated, in which case never mind.
A bit more explanation -
My understanding is that the Electoral vote distribution graphic shows the results of runs of the simulation model and how many times the model would result in various results. His model results in a very low probabilty for some results in either direction and has a peak probability right now of Obama with 335 electoral votes. 75% of the runs have Obama getting 270 or more.
On the left the percent gives the percentage of runs of te model that have each winning the state and the color shows how it went in 2004, solid Red, solid Blue, or close and therefore Purple.
Yeah, it’s got kind of a “for stat geeks by stat geeks” mindset. I also don’t like the charts as presented, there’s not enough information, looks like someone followed his Tufte too far and started trimming muscle along with the fat.
Once you realize what you’re looking at, they are really quite informative, but that should not be my responsibility as a viewer to find out.