2. Are bears smarter than we think?
3. Rant in defense of political science, and cities are selling naming rights.
4. New claims about winning strategies in the iterated prisoner’s dilemma.
5. Gabriel Roth does not like the Loudoun County rail extension.
2. I believe the measure of a bear’s intelligence is if he wears a hat and a tie.
Interesting. Why lump the rant in defense of political science and the cities selling naming rights together?
Perhaps if I can no longer get NSF funding I’ll start selling naming rights for my papers/
An anonymous Yosemite ranger, quoted on the problem of designing bear-proof trash containers: “There is considerable overlap between the intelligence of the smartest bears and the dumbest tourists.”
Quantitative political science, like “macro-economics”, is a big joke — careerism over true knowledge or insight
Even a black bear has no qualms about letting itself into a house and helping itself to the contents of a kitchen, especially if there is only one person living on a rural property.
Why do the Swedish measure their strawberries in litres?
Why has nobody commented that it’s Freeman Dyson (of “Dyson Sphere” fame) that’s a coauthor on the prisoner’s dilemma paper?
When I was a kid in Minnesota, strawberries were sold by the quart. Volume measures such as the bushel are pretty common in agriculture. How are strawberries sold where you are?
By the punnet.
Every few years someone makes the claim, using some narrow measure of intelligence such as “numerosity” or “GPSA”, that bears may be smarter than we think, yet these researchers seem to ignore the larger spectrum of bear behavior which has been addressed over and over elsewhere, namely that bears rarely, if ever, apply any of these so-called “abilities” in a natural setting. Bears, of course, are not stupid, but they are about as dumb as you think.
I think #4 is extremely interesting, and want to hear Tyler’s thoughts on it.
Anywhere I can get an ungated copy of #4?
It appears Jacqueline Stevens has made the very important discovery that political science–indeed any science concentration–is a probabilistic endeavor, often initiated for less than ideal reasons (that’s where the funding is, &etc), whose value is based more on its deliberative qualities within the particular area of study rather than the accuracy of its (tacit) predictions or descriptions of reality. I believe Jim Manzi’s treats with this topic in a much more constructive way in his excellent book “Uncontrolled.” It would be one thing if Ms. Stevens was articulating an argument concerning the lack of epistemic modesty in political science, particularly in the quantitative analyses. However, this is not the source of her discomfort. Further, Saideman notes that there are very few political scientists that are in the prediction business.
Can anyone figure out from the article what the winning prisoner’s dilemma strategy actually is?
2. Well, they have devised some very clever means of obtaining YouTube hits.
I’ve not seen the paper, but the description seems clear enough: Let A and B be players, with a for defect & A for cooperate & etc. Then there are four possible outcomes, ab, aB, Ab, and AB. The payout is symmetric, and the payout for A in AB is less than the payout for aB, with the payout for ab usually the worst of all. If we normalize the payouts from 0-1, as B, we can constrain A’s payout from ab to Ab if we wish to punish, and from aB to AB if we wish to reward. If we want to force the average payout to some range, we simply defect if A has accumulated “to much” and cooperate if too little. It is in this sense that B determines A’s payout.
This methodology works so long as the cost of implementing any given strategy is fixed. If more complicated strategies are more expensive, then there is no stable strategy. (Tit-for-tat costs more to implement than always cooperate. Always cooperate loses to random defectors. Random defectors lose to Tit-for-tat.)
It seems to me that the indirect game, where there is some probability (say 10%) that the player’s play is NOT what they intend, would promote a sophisticated player. It seems to me that the sophisticated player goes in expecting to be cheated, and is pre-prepared to handle that case.
Comments on this entry are closed.