Franken to Win!

7 thoughts on “Franken to Win!”

1. And there is a large degree of uncertainty. All Nate really said was that the odds are maybe 52% Franken, 48% Coleman. It’s basically still a coin flip.

   And if when the counting is over they are tied, it literally is a coin flip there.

   Log in to Reply

   1. Silver’s model for the election worked partly because all the polls more or less agreed, so averaging them was likely to give a good prediction (no matter how you weighted them).

      But this is making some rather questionable assumptions, and to predict a very specific vote margin based on some very rough techniques is a little dishonest.

      Log in to Reply

      1. First of all, it’s only dishonest if you don’t acknowledge the likely error.  If you read the article, it’s very clear about the error.

         Second, Nate’s model certainly performed far better than you imply.  There are, after all, a number of poll aggregators out there: among them RCP and pollster.com.  Nate beat them all handily.  He did it because instead of just eyeballing a formula for weighting polls, he spent a lot of time figuring the correct weights, looking at all the factors.  It’s a bit misleading to chalk that up to any average being just as good as another.  Credit where credit is due, Nate had the right weights for his averages, as well as the right means of projecting trends, and accounting for other factors that would eventually impact the results.

         Log in to Reply

         1. NS missed Indiana. I don’t know how RCP did. But that’s a big electoral gap. NS also missed Nebraska’s electoral vote. If you consider only the swing states, not the ones that every intelligent person should have known were sure things, NS doesn’t seem as successful as others were. Anybody could look at polling averages and make predictions, but Silver wasn’t the best among them. I think some random dude on Daily Kos did better.

            Give me a list of your lottery ticket bets and how well you did, and I can give you a regression. That won’t give you any meaningful information about the next lottery ticket, since it’s still all random. But there’s a ton of analysis you can do with statistics regardless.

            Log in to Reply

            1. You’re apparently just looking at win-loss records for each state.  That’s going to inherently be somewhat random.  Also, due to the small number of decisions that have to be made, it’s almost assured that some random guy from Daily Kos did a better job of it than any one given source.

               I was looking at how far off the predictions were from the actual vote margin in each state, and fivethirtyeight.com was way better than any other polling data aggregator.  That’s also the only way it’s really sensible to look at the quality of data from a statistical standpoint.  Whose errors happened to flip states is a matter of pure chance; who had the most error is a matter of the quality of the model.  Would Nate’s model have been better if it had overestimated Obama in Indiana, rather than the other way around?  No, not really.  He just would have gotten lucky when it came to predicting electoral votes.

               Log in to Reply

            2. Aside from IN and the Omaha electoral vote, Nate predicted every state correctly.

               And cdsmith is right about Nate’s statisitcal analysis, it was much better than any of the other sites.

               Log in to Reply

2. According to 538.

   The nightly, 8 PM update from the Minnesota Secretary of State now shows Norm Coleman with a nominal lead of 172 votes over Al Franken. According to the state’s accounting, Franken has gained 43 votes on Coleman since the recount process began. However, the high number of challenges on both sides probably obscures any true movement in the vote counts.

   Getting it close only counts in horse shoes and hand-grenades. We’ll just have to wait an see if Nate is right on this one as well.

   Log in to Reply