prior probability

Missing from this infographic: What is the average (median) sentence imposed on the 83 convicted war criminals?

Remember the Y2K Problem? Looks like we can add “the lack of net neutrality” to the list of terrible worries that were overblown or never materialized. Back in 2006, for example, our libertarian friend and colleague Tyler Cowen supported the policy of net neutrality: “Without neutrality, Comcast and Verizon would use differential pricing schemes to extract more revenue and thus diminish some forms of Net output, including Google, Amazon, ebay, and possibly blogs.” (See here, for a full list of Prof Cowen’s reasons in support of net neutrality.) Today (2017), however, after examining some evidence, Prof Cowen has now concluded that net neutrality is no longer a necessary evil: “we’re at the point where we’ll do just fine without it.” For our part, our position is that antitrust law, not telecommunications law, should be used to police Internet markets.

Source: @nationaljournal

Note: This is a follow-up to our Nov. 14 post titled “A Bayesian defense of the Hadley rule.”

In our 2014 paper Trolley Problems, we wrote (footnotes omitted): “Consider the [standard] version of the trolley problem [pictured below]. There are seven parties to this conflict. On the one side of this moral equation is the person at the switch, who must make the difficult decision whether to leave the trolley on the main track where five workers will be harmed or divert it to a side track where only one worker will be harmed. On the other side of this moral balance are a total of six workers who are all potentially at risk from the runaway trolley, depending on whether the switch is pulled. But because of the veil of ignorance, none of the players knows their role ahead of time. Given this counterfactual world, imagine what would occur if these seven unfortunate souls could call a Coasean time-out to take part in a Coasean auction behind a veil of ignorance. In truth, since this a second-order thought experiment, the outcome of such a hypothetical auction is not obvious, unless a probabilistic approach to the trolley problem is taken.”

In place of a hypothetical Coasean auction, however, what if we imagined a “diabolical trolley lottery” instead? In the original position, there is a 0.714 probability that you will be one of the five workers trapped on the main track (5 ÷ 7), a 0.143 probability that you will be the worker on the side track (1 ÷ 7), and a 0.143 probability that you will be the person at the switch (1 ÷ 7). Accordingly, we can now imagine a negative lottery corresponding with these same probabilities and ask, How much would we pay to avoid having to play this diabolical lottery? Or in the alternative, would we prefer to pay nothing and take our chances? After all, in the original position there is a high probability (71.4%) that we will be one of the five workers trapped on the main track, and likewise, isn’t there a high (but unknown) probability that the person at the switch is a Humean consequentialist who will divert the trolley to the side track to save the five lives?

Continue reading →

Florida has more forest cover than Colorado! (For an explanation, check out this thread on Reddit; hat tip: danyellan.)

According to various reports (just google the words “Jameis” and “Uber”), the ride-sharing firm Uber listened to its driver and banned NFL quarterback and accused rapist Jameis Winston from using their service. The driver, however, did not file a police report. Can you blame her for not doing so, considering the way victims are treated by our public systems of justice? File under: “private-sector justice is swifter and more certain than public-sector justice.”