What the Fox Knows
fivethirtyeight.com
fivethirtyeight.com
Just a quick glance seems to indicate this will become a staple in my RSS list and a great complement to my current favorite non-tech blog (Grantland).
http://grantland.com/contributors/sean-mcindoe/feed/
It might not work for everything, but it should work for a lot of things.
We are going to screw some things up. We hope our mistakes will be honest ones. We hope you’ll gain insight and pleasure from our approach to the news and that you’ll visit us from time to time. We hope to demonstrate the value of data journalism as a practical and sustainable proposition.
It’s time for us to start making the news a little nerdier.
> And news accounts routinely estimate the number of attendees at political rallies.
In my experience, news organizations routinely over-estimate the attendance of events to which the reporter sympathizes, and dramatically under-estimates the attendance of events to which the reporter does not.
As someone who attends a lot of rallies encompassing both 'left' and 'right'-leaning movements, I can say that I've witnessed this first hand. I attended an event in Annapolis to oppose some gun control measures last October. The attendance, by my count, was nearly 4,000. I can prove by the testimonial registry that we had over 1400 people not only in attendance, but that provided testimony before the legislature. At the same event, those in favor of the rally were protesting as well. They had, by my count, a few dozen supporters in attendance. Their testimonial registry was exactly 22.
The Baltimore Sun and Washington Times reported attendance for both camps at 'about equal', which a 'few hundred in attendance for either side'.
When protesting in support of the fourth amendment, or in support of free speech rights for the media, those numbers were conversely inflated.
The governor's testimony before the Senate Judicial Proceedings Committee came after at least 1,500 gun-rights advocates rallied outside the State House in opposition to the legislation...
Meanwhile, advocates on either side of the issue packed the hearing room and an overflow room downstairs. Hundreds more stood in a line that extended down a stairway to the floor below.
Legislative aides said they had never seen such an outpouring of people seeking to testify on a bill. By an overwhelming number, they were signing up in opposition to the governor's proposal.
http://articles.baltimoresun.com/2013-02-06/news/bs-md-omall...
"Hundreds testify for, against proposals on assault weapons, licensing" is a sort of true statement, except that exactly 22 is not hundreds.
I remember more biased reports from the time, but I can't find them now. So while there's some validity to your point, I've seen it happen enough to say confidently that those estimates are best-guess estimates, and those estimates are subject to bias.
[1] http://www.washingtontimes.com/news/2013/feb/6/gun-rights-ad...
I'll try to revisit this after my dentist appointment with some evidence, or a retraction.
Hence why you use Bayesian analysis in the first place.
I expect he will also continue to fail to quantify the performance of his models against other benchmarks. In the case of the NCAA how much value does his more complex model add compared to each of the rankings he uses as inputs alone?
As commenters pointed out in an earlier New York Mag article (http://nymag.com/daily/intelligencer/2014/03/nate-silver-int...) and that Silver alludes to via hyperlink, the "hedgehog vs. the fox" quote could be interpreted as coming out in favor of the hedgehog:
http://en.wikiquote.org/wiki/Archilochus
> The fox knows many tricks; and the hedgehog only one; but that is the best one of all.
I think if 538 had picked a hedgehog as its mascot, it would still make sense: it's the pundits/prognosticators who have "many tricks", whereas Silver is proposing just one good trick -- empirical data analysis -- which happens to be effective and valuable in all contexts.
The idea that Tetlock puts forward is that pundits divide into two categories, Foxes and Hedgehogs. Hedgehogs explain the world and make predictions according to one large all encompassing principle. Foxes don't, and apply different principles to different circumstances. Tetlock concludes, and Silver endorses this, that Foxes are better at understanding the world than Hedgehogs.
An example might be the (proposed) principle that free markets lead to increased prosperity. A hedgehogy capitalist might state the above as a broadly applicable principle. A foxy capitalist might opt for something like: "yes, many times increased freedom makes us prosper, but sometimes it doesn't, it depends, let's talk about the specifics".
I take this to be an argument about complexity. The fox side of the argument is roughly that the world is more complex than most people allow for, and one principle or world view will usually not cover an area of knowledge well.
I don't think there is a clear division between those two. Both, the fox and the hedgehog use only one view– their own to seperate signal from noise. The fox side is just more open to integrating conflicting point of views into their worldview.
What I mean to say is that in the end both use only one trick, one skill- filtering information. This of course doesn't mean that the metaphor is wrong, just that it's wrong to think that the fox doesn't filter information the same way the hedgehog does.
BTW, this reminds me of something from Laozi "Through what do I know the nature of all things? Precisely through them." While the things are many, the principle of understanding them through themself remains singular.
But seriously, I think talking about what _exactly_ something is meant to represent isn't "lawyering", it's essential for a serious discussion. I got another quote, too: "The boundaries of our language are the boundaries of our thoughts".
Telepathy would propably be better though, I agree.
Silver is adopting an analogy to describe modern journalism, and the analogy largely fits. Traditional political pundits spent a lot of time in 2012 crafting narratives that fit with what they felt was blowing in the wind. In this case of using the analogy the hedgehog sucks.
However, Archilochus was definitely praising the idea of the hedgehog in his poem. A good analogous English saying would be, "Better a live chicken than a dead duck".
This may not work for everybody, but I love the speed of access and the way my bookshelf looks as a result.
Edit: I have done this with my icons in the past, too: http://flickr.com/photos/porkfriedrice/3753844018/
Something I noticed last night while watching ESPN is that they must have mentioned FTE a dozen or so times within an hour. Are they backing FTE? I even saw an interview on ESPN with one of the cofounders.
http://www.nytimes.com/2013/07/20/business/media/nate-silver...