An FAQ for the 2012 US Presidential Election(norvig.com) |
An FAQ for the 2012 US Presidential Election(norvig.com) |
I love how he concludes that voting is rational. I'd never thought to multiply the, potentially massive, cost of the candidate I disfavor getting elected, multiply it by the tiny chance my vote makes the difference, and compare the result to the amount of money I make in an hour. He admits his cost figures are fudged, but in principle, it works out.
Since I spotted this on the front page, it seems to have been nuked by mods. Perhaps that's for the better, but I'd be much more interested in discussing politics if everyone could handle it with the restrained style and tone in this article. Interesting link.
If we assume (using numbers from the article) that a vote costs $20 (an hour of our time), the odds are 1 to 10 million, and the payoff is $7.7 trillion (even though it's the country getting the benefit, not the voter) then the Kelly criterion* implies that it's only rational to vote if your net worth is more than $200 million. Somebody making $20 an hour probably isn't worth $200 million, and so even if voting did have a positive expected value, it's not a rational strategy.
TL,DR: There is more to betting strategy than expected valve, and no rational gambler takes a bet with 1/10 million odds for $20.
(It's also worth considering a la Hofstadter/Yudkowsky that something like your decision process is instantiated in other people like you -- if you decide not to vote, many of them won't either, even with no causal link.)
-- Its astonishing that there is nothing in this regarding the US Debt. eg http://www.kpcb.com/insights/usa-inc-full-report [1]
______________
[1] Summary> http://www.businessinsider.com/mary-meeker-usa-inc-february-...
Having said that, I find Washington state initiative 502 (and to a lesser extent colorado amendment 64) to be the most interesting and important elections this year.
Recent polls have shown support for marijuana legalization of more than 50% and a state voting for legalization will be the start of the end of prohibition.
Also, remember that Bush presided over two terms, winning two elections, and the majority of the "cost" was in the second term and thus not a direct result of the close 2000 presidential election he's analyzing.
Conclusion: Norvig's analysis in this case, in contrast to his technical work, is not worth the paper it is (not) printed on.
Moral: Political speech is usually politically motivated, and is generally not as strongly based in rational analysis as the speaker would have you believe.
This is one of the most bilious elections that I remember and although I'm interested to hear the spectrum of views on HN as a general matter I'm depressed by the likelihood that the discussion would just degenerate into a flamefest. I'd rather not see friends and fellow HNers slinging mud at each other, because heaven knows there's enough of it on the rest of the internet. Website comments on political matters make me question the whole premise of democracy :-(
1. The majority of the cost (the war in Iraq) was initiated in the first term; a better president with better advisors (IMO, Iraq II was all Cheney and Rumsfeld) would not have gone to war in Iraq with the flimsy justifications presented. Thus, the continuing cost of that war would not have happened during the second term. We may have seen an expanded conflict in Afghanistan, but there was substantial international support for this, which would have spread the costs around. It's a fair cop to place the whole cost of the Iraq war on the election in 2000.
2. The financial crisis was enabled by four primary factors: loosened mortgage rules from before 2000 (e.g., under Clinton), irresponsible tax cuts (reducing real stimulus available to the government), the war in Iraq (reducing real stimulus available to the government), and increased and accelerated deregulation of financial players at a time when it was known that there were some systematically bad players (see Enron and others). Given that all four factors happened in the first Bush term, it's also fair to place at least a large chunk of this on Bush.
I would fault Norvig's evaluation mostly on his view of a non-Bush presidency. It is highly unlikely that Gore would have been revenue neutral. On some aspects, it is almost certain that he would have been revenue neutral (no war in Iraq, but possibly expanded action in Afghanistan). Even so, the cost of 9/11 outside of the Bush context would be impossible to determine. I suspect that the TSA as it exists today would not exist; I fear that it could have been replaced with something more expensive (although hopefully less disdainful of civil liberties and the Americans they supposedly protect).
There are other things about the cost-benefit analysis that don't sit right with me as I'm reading them (the way of thinking of a vote as worth 600,000 vs pennies; it seems that more rational math would still be 600,000 vs 20,000), but it's the weakest part of the article to that point. I know what he was trying to do—but I don't think he made the case that it's patriotic to vote based on a financial analysis. (I think it is patriotic to vote, which is why I do vote.)
Everything before is a good summary and pointers to other people doing solid analysis.
The remainder of the article is fully political and can be treated as such (even though I personally agree with him).