No biased coins
Andrew Gelman eviscerates that favorite tool of statisticians everywhere, the biased coin. From the abstract:
Dice can be loaded—that is, one can easily alter a die so that the probabilities of landing on the six sides are dramatically unequal. However, it is not possible to bias a coin flip—that is, one cannot, for example, weight a coin so that it is substantially more likely to land “heads” than “tails” when flipped and caught in the hand in the usual manner.
Too bad nobody told Professor Mitzenmacher.
Syriza and the French indemnity of 1871-73
Michael Pettis does some neat economic history, comparing today’s Eurozone crisis to a massive capital outflow from France to Germany after the Franco-Prussian War (in the form of a massive reparations bill, 20% of either country’s GDP). Surprisingly, this destabilized Germany much more than France. Pettis argues that various distortionary mechanisms within the Eurozone today have had similar effects, and points out that the conflict should not really be about national moralizing (“wasteful Greeks vs. industrious Germans”) but about who absorbs the costs of a (basically inevitable) debt restructuring:
Most currency and sovereign debt crises in modern history ultimately represent a conflict over how the costs are to be assigned among two different groups. On the one hand are creditors, owners of real estate and other assets, and the businesses who benefit from the existing currency distortions. One the other hand are workers who pay in the form of low wages and unemployment and, eventually, middle class household savers and taxpayers who pay in the form of a gradual erosion of their income or of the value of their savings. Historically during currency and sovereign debt crises political parties have come to represent one or the other of these groups, and whether they are of the left or the right, they are able to capture the allegiance of these groups.
Pettis is great at lucid yet dense explanations of interesting takes on macroeconomics. I recommend the rest of his blog as well.
Effective Altruism Books
The Effective Altruism Handbook, a project of Ryan Carey, compiles a bunch of foundational EA essays into one convenient package. In terms of intellectual bang for the buck, this may be the densest compilation I know of. I’m hoping they’ll put together a bulk order of nice printed copies sometime soon.
Peter Singer’s new book, The Most Good You Can Do, is out! Not much new for fans of effective altruism, but it’s very clearly written, concise, and hits all the major points. (Ironically, Amazon’s recommender system told me to buy David Brooks’s new book after I looked at it. Eww.)
Finally, Nick Cooney has written How to be Great at Doing Good, which appears to be much the same material written from the perspective of people who already buy into the “altruism” part of effective altruism. I haven’t read it yet, but I’m looking forward to it!
More EA news
Instagram cofounder becomes the second extremely rich tech person to start working closely with GiveWell. Hopefully we’ll see more of this in the future!
Math and Mystery
Aeon asks: “Is there something mysterious about math?” Scott Aaronson answers yes: the mystery is why it isn’t more mysterious!
Comments
Trivially a coin with two faces marked heads is biased, but that’s not what they’re asking.
Their argument against the possibility of a biased coin has two unrelated parts: (a) students can’t make them out of checkers and clay and (b) a theoretical argument, that a spinning disc will spend half its time with each side up. The real argument is (b); the (a) seems to be just for internalizing the pattern and is very weak evidence.
With (b), however, one thing that’s missing is an analysis of the mechanics of the catch. For example, consider their pickle jar lid. The shape of the lip seems to me like as the fingers curved in they’d be more likely to catch the inside of the lid, affecting when it stopped spinning. The main thing here is depending on the shape of the coin you’re not guaranteed for your catch to cut the space of angular rotations neatly in half.
A strong candidate for a biased coin would be one that was slightly bent. At the extreme, a fully taco’d coin is 100% biased, so one a bit of the way there should be a bit biased.