Statistics, probability, and Nate Silver
kfury.com
kfury.com
In a way, he's a spokesman for statistical reasoning to the world. It's odd to see how much pushback he gets for saying "A candidate with a slight lead in October wins 75% of the time". This is a fact, easy to look up. But because the other side and the media don't like it, Nate is put into the position of defending common sense to America. And he does an admirable job.
Hopefully he can find a way to use this for good outside of politics. Maybe he can educate people about statistics and probability in other ways. Starting a blog that correctly predicts the behavior of reality tv stars might be the best thing he could do for the cause of reason in this country.
On the Democrat side of the table, many treated 538 like a source of religious prophecy and source of comfort - when truth be told Nate is just a smart statistician who developed a model that's accurately predicted the outcomes of three elections (the first being the Hillary / Obama primary.) On the Republican side of the table, major trends in national and state-level polling data were ignored because pundits believed that the pollsters were oversampling Democrats when the truth is that the Democratic base has expanded since 2008.
Other legitimate statisticians, like the University of Colorado, predicted a Romney win and obviously their models were wrong.
Scientists, statisticians, economists, and engineers shouldn't be worshiped as sources of incontrovertible truth - they're just humans trying to do the best objective, accurate job that they can!
True, but they're a big step in the direction of progressively less wrong [1] from the traditional alternative punditry and 'gut' analysis [2].
If I were a movie producer I'd buy the rights to Moneyball 2, script it about this election, and recast the gossiping scouting staff as all the blowhards [3] and bs artists [4].
Nate Silver, Andrew Tannenbaum [5], Drew Linzer [6], Sam Wang [7] and the other modelers deserve a lot of credit for helping the public cut through the bullshit.
[1]: http://serendip.brynmawr.edu/sci_cult/lesswrong/lesswrong/
[2]: http://blogs.wsj.com/peggynoonan/2012/11/05/monday-morning/
[3]: http://www.theatlanticwire.com/politics/2012/10/people-who-c...
[5]: http://electoral-vote.com
[6]: http://votamatic.org
My point is that it is thoroughly, truly annoying to see masses "smart" but still obviously ignorant people apply religious qualities to scientists and statisticians because the outcomes are either a source of divine comfort or blasphemy.
Watch either of the "Go God Go" episodes of South Park if you want a more concrete example of the behavior I'm describing: http://www.southparkstudios.com/full-episodes/s10e12-go-god-...
I'm glad someone pointed this out, since I think it gets obscured a bit in most discussion of 538's statistics.
But yeah, it'd be pretty close.
in other words you can interpret the confidence in the case of the election just as in the case of the football game, by stating that:
1. if your mathematical model of the population is accurate (eg. for instance your population really is voting using a binomial law of a certain probability p_democrat of voting for the democrat candidate) 2. and if the measurements (polls or player states) you have made give you that at least k voters out of N polled would have voted democrat
then you can consider all the (hypothetical) instances of a population which would have resulted in a same or higher number of democrat voters for this poll, and you can compute the confidence interval that gives you the probability that such a population would indeed elect a democrat president
So you don't necessarily have to view it in terms of "not changing the population vote, but changing the precision of the data", you can also cast it in just the framework of the football prediction
This is a distinction without a difference. There is "internal chaos" in the game because we don't know its initial conditions exactly, just like in the elections. If we had precise knowledge about the particles in the ball, the field, and the players, we could theoretically predict the outcome with high certainty. Except for quantum mechanical effects, statistics isn't about true randomness, it's about modeling lack of knowledge.
The greater point though is that with our current capability of determining state we can eliminate a huge amount of chaotic factors from elections, but very, very few chaotic factors from sports and lotto drawings.
We're not anywhere near the ability to determine state to that degree, and even if we were there's the certitude that unless the system was completely closed before the measurements were taken, the minute changes that result from the prediction (a team getting a sense of security, a fan cheering louder, a teammate changing a stride into the locker room because he read a tweet spawned by the prediction) throw the prediction out the window.
So yes, on a scientific basis you're right and that's the grist for another blog post, but on a practical level it's irrelevant.
Probability is still in the mind. Not in the territory.
http://lesswrong.com/lw/oj/probability_is_in_the_mind/ (In case you didn't already spot me.)
That said, even though there is good experimental evidence for this, it is not 100% certain that the world actually works this way.
But yeah, as long as we didn't find complete laws of physics, we're bound to ignore something.
Let's see with an example. A half sievered mirror, a photon going through (or bouncing), and 2 detectors. You say there is no hidden variable that would allow us to predict which detector we will see go off.
I agree.
It doesn't mean however that the laws of physics have true randomness in them. Under the Everett interpretation, two "worlds" (blobs of amplitude) are created: one in which detector 1 goes off, and one in which detector 2 goes off. from the outside, this is purely deterministic. The only randomness lies in our subjective experience. The only question remaining is why do we experience the Born statistics (based on square moduli), rather than something else (like, based on moduli)?
We would have true randomness at the fundamental level if some form of the Copenhagen interpretation where true. But believing the blob of amplitude we don't see would mysteriously disappear, violating the equations in the process? That's just nuts. If this idea appeared after the Everett interpretation, it would have been ridiculed.
I say largely irrelevant because there is a minority of people (including, if I'm not mistaken, Nobel prize winner 't Hooft) who say that neither Many Worlds nor Copenhagen is right, because there are hidden assumptions in EPR that are overlooked. Since you seem to be a fan of lesswrong.com, this should appeal to you ;-).
I think for the rest we are in agreement, I misinterpreted "as long as we didn't find complete laws of physics, we're bound to ignore something" and "Probability is still in the mind. Not in the territory.". The latter statement becomes rather tautological and its meaning rather weak if you include QM randomness though. In the context of the frequentist vs Bayesian debate it means something far stronger.
As for how this is metaphysics… Sure, but no more that wondering if a photon (or a colonization ship) you just send into deep space, outside the observable universe, still exists. I mean, I don't care if it's meta or not. I just trust Occam's razor (to the extent it gives prior probabilities, not prior certainties), and leave it at that.
Nitpick: You didn't even need Occam's Razor: Many World implies that the QM predictions are true. of course the truth of those predictions are more probable than Many Worlds. I, on the other hand, used Occam's Razor to compare Many Worlds and Copenhagen, which are mutually exclusive.
I agree agnosticism is good. I'd rather be aware of my own ignorance. I like it better however when I'm also aware of its magnitude. Probability theory does just that. I don't know if Many Worlds is true, nor do I know if Copenhagen is false. But my current state of ignorance is more precise than that.
But if you want to use polling data, Nate Silver is demonstrating the least-bad way to do it.
The fact that agents use polling information to make decisions doesn't necessarily mean the system won't reach equilibrium.
But I do think it's conceivable that one candidate's predicted overwhelming chance of victory might have a way of becoming a self-fulfilling prophecy in a close election.
Ron Paul supporters presumably prefer Romney to Obama. In an election where it was reported that the race was close, they'd vote for Romney. But if it was reported that an Obama win was basically guaranteed, they might be willing to vote for the Libertarian Party or write-in Ron Paul (who isn't even running) as a protest vote.
I really think our democracy would be healthier if we switched to an instant runoff system [1], but the Democrats and Republicans have a very strong incentive to unite against anything that could strengthen third parties.
It also has the possibility for the opposite by lowering turnout for the "winning" candidate. Why vote? Nate Silver already called the election, etc.
There are other ways in which polling data itself can influence an election. It appears that Nate Silver did an great job early on at predicting the effect of changing demographics in swing states like Florida--which ended up giving the election to Obama. Hindsight being 20/20, if the Romney team had paid more attention to that (and perhaps compromised on things like immigration reform), things could have gone very differently.