October 4, 2004
Just for Fun
OK, I'm sure most of you have noticed the button to the left chronicling the current electoral vote count as tabulated by Electoral-vote.com I don't know how acurate his projections are, how he chooses which polls to report, and I also note that he just shifted methodology, running a 3 day average of polls rather than relying on daily polls.
Regardless, he's done a lot of work gathering all of this data, and it makes for a fun way to keep at least one measure of the campaign. Also on his site, he projects the final electoral count, evaluating existing data and formulating a line using a linear regression. I question that approach, since I doubt voter movement is going to be anything close to linear, so I use a different way to track voter movement. I'm tracking a 3, 5, and 10 day moving average. No weighting inviolved, just a straight average.
As you can see below, the results are interesting. Bush shows more volatility, but tends to sit well above the 270 line. Kerry works on a more even keel, which isn't good for him since he tends to rest below the 270 threshold. Both candidates show a slight upward trend, Bush a bit stronger than Kerry (3 and 5 day moving up for Bush, Kerry 5 day flatlined), indicating that undecideds are deciding, making states easier to call. Based on the data, I'd have to say that this is Bush's race to lose, but it's not in the bag by any stretch. Kerry could be building a strong base that could break out at any time.
Or this could be totally meaningless, but it's kinda fun.
