Last summer, if I remember it correctly, the boyfriend of a friend of my wife's died in a Florida swamp; it appears he didn't bring enough water, and didn't stay on the path, and he got lost and died of sunstroke or something similar. I spent a few months in 2008 hiking, and what kept going through my head last August was that he died because he made a couple of mistakes — and I had probably made mistakes that were, at least by some measure, just as bad. (There was in fact a day in which I got lost wandering off the trail and ran out of water; fortunately it was in a more forgiving environment.)
Friday, July 08, 2016 :::
In 2004, when I first heard a no-down-payment, interest-only mortgage loan advertised, I termed it "the foreclosure loan". There were a lot of mortgage loans made in that era that were affordable if the borrower never got sick or lost a job or had a car break down and got regular raises; that is to say, they were not affordable. If something really big and really surprising makes your financial set-up unsustainable, you can blame that, but if your financial set-up can't tolerate any relatively normal bumps in the road, the problem isn't the bump. "Human error" in general only makes sense as a target for blame when it's somewhat beyond the normal, run-of-the-mill human error; if a system produces a disaster because of normal human error, it's a bad system.
I'm not sure I've ever done anything as dumb, in some sense, as resisting arrest, though I've done things that should be more likely, in the natural run of things, to be fatal. I don't know how often people resist arrest, so I have no particularly good estimate of what fraction of the time that is fatal. When someone gets killed resisting arrest, there are people who note that you shouldn't resist arrest, and (often other) people who note that doing so should not typically be fatal, and these people are, in fact, correct. We aren't going to have a perfect system, but if we make things such that it is less likely that people's mistakes turn unnecessarily fatal, that would be, you know, super.
What's more troubling, though, is when the person who dies is not even someone who made a mistake. Even if I don't think resisting arrest for selling bootleg CDs or cigarettes should be fatal, it's more troubling when someone in a car that is pulled over for a minor equipment violation is killed because he was reaching for his ID in response to having been told by a police officer to reach for his ID. People die in car accidents in which they aren't at fault, and otherwise for other people's mistakes, and we're never going to get rid of all of that, but it's still very frightening when it happens.
As I noted along the way, there is some extent to which the world can never be fully tamed, and we should keep statistics in mind when assessing the scale of the problem when there are particularly salient episodes, and it's even worth keeping in mind the possibility that the best response might actually be to do nothing, but at least investigating and looking around for something better to do is surely worth the trouble.
::: posted by dWj at 12:54 PM
Wednesday, June 15, 2016 :::
It's dangerous to change public policy in response to a single event, especially shortly after the event, when its importance is especially inflated. This is partially because any rushed policy change is likely to be poorly thought out, but also because the stuff of life doesn't grab headlines.
If we let media coverage drive policy, attempts to reduce murders even more than they have already fallen in the last several decades will focus too much on mass killings that dominate news cycles nationally and too little on other murders which are more frequent but which get less attention. Gun rights activists like to say that, when the bad guys have guns, we are better off when good guys have guns and are trained to use them. This story is plausible, but it seems strongest in the case of a mass shooting - if a shooter has one victim in mind, he may be done by the time anyone knows he's a threat, but in a mass shooter situation, it's the good guy with the gun who has the element of surprise. So we might be tempted to respond to a mass shooting by making concealed carry easier, reducing the likelihood and magnitude of a large-scale event, but whether that would increase or reduce the overall murder rate is an open question.
My understanding is that the empirical results of well-done studies are mixed (the poorly done studies all seem to find a positive correlation between gun ownership and gun crime or gun deaths and infer that guns cause crime), and that the best conclusion is that any effect of gun ownership on crime is small, whatever its sign. Unless something useful can be said about subsets of guns or gun owners or crime, gun policy should probably be made based on other considerations.
::: posted by Steven at 9:36 AM
Friday, May 20, 2016 :::
The census recently released estimates of populations of cities over 50,000 residents, and I've seen (a few different places) lists of the ten cities over 50,000 or over 100,000 with the largest year-over-year growth. They all use percent growth, and the leaders are naturally all close to the inclusion threshold; conversely, if you use absolute growth, your leaders are NYC, Houston, and Los Angeles, which are also the three largest cities that showed positive growth. Below is a list of the top 25 cities ranked by absolute change divided by latest population to the power of 5/8, a number whose selection was based on my own aesthetic sense that it gave a good mix of cities at different ends of the population range.
One of the reasons I've seen this pop up is that the town in which my brother and I lived from 6–12 grade is near the top of the percentage lists; it shows up as 11th here.
|San Antonio, Texas||1469845||1440309||29536||4.13|
|New Braunfels, Texas||70543||66204||4339||4.05|
|Fort Worth, Texas||833319||813425||19894||3.96|
|South Jordan, Utah||66648||62851||3797||3.67|
|Charlotte, North Carolina||827097||809402||17695||3.54|
|Raleigh, North Carolina||451066||440399||10667||3.12|
|Mount Pleasant, South Carolina||81317||77667||3650||3.11|
 1/2 seems natural, but still seemed to have too many large cities at the top. At that point a power of 5/8 pretty well fit the distribution of the top cities, i.e. their absolute changes tended to go as city size to the 5/8. So it was driven by my own aesthetic sense, but justified after the fact by Science.
::: posted by dWj at 11:34 AM
Wednesday, April 20, 2016 :::
Bernie Sanders seems at this point to have sewn up second place in the Democratic race, but let's ponder the least implausible path for him to still win the nomination. He would have to win the remaining primaries by close to 20 percent of the vote to catch up with her total of elected delegates, and a big chunk of those remaining delegates are elected next week, when he'll have "momentum" (insofar as that's a thing) working against him.
It seems to me one of his least unlikely paths to the nomination hinges on a dramatic development in May, perhaps related to deepening legal troubles or demonstrations of Clinton's casual assumption that she's above the law, followed by a large shift toward Sanders that lets him win California (where he currently looks likely to lose by about 20 percent of the vote) by a sizeable margin, and lets him narrow Clinton's lead in bound delegates to perhaps under 100.
At that point, the pitch to the superdelegates is obvious: I may be senile and out of touch with reality, but I'm the current choice of the Democratic voters, and you owe it to them to outvote the delegates they elected — some of them when they were misinformed — and let me bear their standard in the general election.
It's quite possible I underestimate how much the superdelegates fear or respect the Clintons or fear or, you know, really fear Sanders, but it seems at least arguable that, as of today, the presence of unbound superdelegates marginally improves his chances from what they would otherwise be.
"But"? Let's go with "but".
Labels: Bernie Sanders, psephology
::: posted by dWj at 8:35 AM
Tuesday, April 19, 2016 :::
Bernie Sanders has released his tax returns, which show him taking deductions that he wants to eliminate and Jim Geraghty is calling him a hypocrite. For the reasons I laid out four months ago, I disagree. As I noted on Facebook:
I don't see hypocrisy here, or when libertarian politicians accept money from programs they would scale back, or when the Koch brothers decry corporate subsidies yet fail to calculate how much they benefit and send it back to the feds. It would be more troubling if he were calling for higher taxes on income above an amount that just happens to approximate his own income, or the elimination of all tax deductions except those that benefit him. He says that rich people should pay more; his tax returns indicate that he doesn't mean to exempt himself. Good for him.
Note that if Sanders has criticized his fellow rich people for taking the deductions - if he has said that, until the law is changed to eliminate the deductions, taking them is morally wrong - the hypocrisy charge sticks. The assertion you are making when you charge someone with hypocrisy is that their actions suggest that their purported beliefs are insincere, or at least that they don't believe they should be held to those same standards, and there is no reason to believe that someone who supports higher taxes to pay for more government programs but doesn't pay higher taxes than required is either insincere or looking for a loophole for himself.
Labels: Bernie Sanders, hypocrisy
::: posted by Steven at 11:41 PM
Monday, March 14, 2016 :::
Warning: sports post.
The NCAA announced last night which 68 teams get to compete in a tournament for the men's college basketball championship, and will announce tonight which 64 teams compete for the women's championship. One sometimes hears from casual sports fans that a team's wins in the tournament justify their inclusion and validate the decision of the committee, while a good performance in one of the consolation tournaments by a team that might have been included in the big one refutes the decision. One sometimes hears from the experts that this is wrong, that selection is on the basis of "the merits", and that it is not a prediction. Oddly enough, I side to a significant extent with the populists on this one.
The committee claims to have as its goal the selection of the "best" teams in the country. I'm not sure they really do, even by their own assessment; I think, in particular, that South Carolina is among the 34 best teams that didn't get in automatically, and was left out because they scheduled all of their out-of-conference games against weak opponents, so while it's quite possible the committee simply disagrees with me, I can easily tell a story wherein they are at least creating small incentives rather than going strictly with the best teams. If they are trying to pick the "best" teams, I can't imagine an argument that says that a team's performance in games in the next week or two isn't evidence as to how good a team is. In 2011, Butler and Virginia Commonwealth (which were both, to the committee's credit, included in the tournament) were both better than we thought — though not, most likely, among the four best teams in the country.
Perhaps the argument the experts are trying to make is that there is a degree of unpredictability, that the race isn't always to the swift or the battle to the strong, and that is certainly true. There is also an argument as to what is reasonably knowable by the committee and what isn't; the committee cannot be expected to know exactly which the best teams are, even if that's more reasonable than knowing which teams are going to win which games in the future. In a tournament that is so large that it easily accommodates any team with a reasonable claim to be the best in the country, it seems perfectly reasonable to me to use some of the extra slots to reward teams for testing themselves so that it is easier to tell in the future which teams are better than which; a team that regularly builds a schedule that makes it hard for outsiders to be confident that this is one of the best teams in the country doesn't need the benefits of any doubt.
The best team they played before the conference schedule was selected by the committee, but is required to play a play-in game on Wednesday.
Institutionally, of course; some of the actions of the "team" are taken by the school's athletic director, some by the coach, and others by the players.
::: posted by dWj at 10:36 AM
Thursday, March 03, 2016 :::
Okay, here's a beaut from the New York Times; apparently the state law of New Hampshire specifies that
The secretary of state shall apportion delegates to the national party conventions among the candidates voted for at the presidential primary by determining the proportion of the number of votes cast for each presidential candidate to the total votes cast for all presidential candidates of the same political party, rounded to the nearest whole number.
It seems to me that the obvious extremely literal way to read this is that, well, the proportion of the number of votes cast for each Republican candidate to the total votes cast for Republican candidates was less than one half, and therefore rounds to 0. This would therefore allocate no delegates; later clauses would essentially make New Hampshire a winner-takes-all state in all circumstances. This would be a convoluted way of writing that; it looks like they were trying instead to write the rule that I thought a month ago that they were using, in which what is being rounded to the nearest whole number is the number of delegates apportioned, rather than the fraction of the vote received. (This is what the Secretary of State actually did at this election.) Apparently, though, the media and its experts had (as far as I can tell) substituted the word "percentage" for "number" in the above passage to make it read differently from a convoluted "winner-take-all" rule but an even stranger rule in which candidates were awarded fractional delegates based on rounded percentages — which they then presumed would be rounded off again, even though the law makes no provision for this. (Presumably rounding off the number of delegates at the end seems more natural than some other means of getting an integer.)
::: posted by dWj at 9:55 AM