Now this is not true at all, at least with respect to elections. If anything, Nate’s genius is in marketing his product. Plenty of people have been applying statistics to elections for the past fifty years. I somehow managed to get a degree in grad school doing it almost five years ago. There are quite a number of researchers who do it accurately, some of whom were my professors. They publish tightly-reasoned journal articles for specialists and are hired to consult on major political campaigns.
Nate puts his material up on the internet and writes about it for nonspecialists.
In my view, it is a very recent phenomenon that non-specialists have become interested in applying statistical analysis to elections. In the past, such things were interesting only to one branch of political science. Lots of political scientists reject a game theoretic/econometric approach. I believe that the surfeit of polls and instant feedback information has left people perplexed about what the state of the world actually is. Statistics was designed to answer these kinds of questions.
I will certainly concede this point. I am certainly not a specialist in this field and was not familiar with all the work on elections. I am more familiar with sabermetrics and the history involved there. I would assert Nate is a pioneer in that field, for sure. In elections, I will defer to your experience/expertise.
Silver was on the Colbert Report a couple weeks back, too. He’s made quite a splash this election cycle; I hope he keeps his hand in the game in future cycles.
The truth must be told: I really don’t know diddly about baseball, having given up on the sport when the Cubs choked in '69. I was just looking for an excuse to mention the Lively Ball because baseball pundits who still remembered Tinker to Evers to Chance used to end every description how modern players sucked with, “But that was before the Lively Ball.”
Recently, fivethirtyeight did an article on the various polls and which ones were more or less representative of public opinion. I read the article and came away with the impression that they (Nate?) really know their stuff. Prior to the article I believed Gallop was the last word, now I’m looking more at Rasmussen. Thank you fivethirtyeight.com
How did Silver do predicting the other 5 division races?
Baseball Prospectus is great and I have a lot of respect for Mr. Silver, but it isn’t right all the time. I’m relying on memory here so any inaccuracies are the fault of my failing brain cells, but I seem to recall that in mid September 2007 they said that the Mets had a 90% percent chance of making the playoffs and the Rockies had a less than 10% chance of making it, and yet both of those long shots came through, the Mets choking away their chance and the Rockies going on an unbelievable streak and making it all the way to the Series.
So it sounds like they were right, given your description: the Mets were expected to go to the playoffs, and it was notable that they didn’t, whereas the Rockies were not expected to, and it was amazing that they did. Did they update the chances as each team progressed through the season?
Yeah, they update the playoff odds report every day.
And the odds were a lot more extreme than 90%/10%. The Mets, IIRC, were over 98% likely to make the playoffs at their peak, and quite possibly over 99%. Which is accurate. A team will blow a 7.5 game lead with 17 to play maybe once a century. The Rockies winning streak was similarly historic and unlikely.