Round Numbers And The Netflix Prize

The Netflix prize was just won, making a post I had sitting in the back of my head a lot more timely.  In case you’re unaware, the competition was to improve by 10% on Netflix’s own algorithm for recommending titles to customers based on their previous preferences.

But what’s special about 10%? I ask because improvement on Netflix’s algorithm was extremely quick.  5 % took less than 2 months, and multiple teams had reached 8% in a year.  It’s been 18 months since then to reach 10%.

On the fact of it, that’s very strange.  It’s true that a well designed prize would take long time to beat, and that it should take increasing quantities of time to make improvements as time goes on.  But then it should be possible to overshoot and set a target that won’t be met.  The timeline of the Netflix competition suggests that they did a very good job setting a target–not too ambitious, but not too easy.

But what’s the prior probability that any given target would be right? Not wonderful.  So here are three possibilities: 1) It was luck.  10% was a nice round number, Netflix chose it, and the competition worked out right.  2) There’s only so much possible improvement to be had.  More than 10% was unrealistic, and Netflix knew that, because a) there are theoretical results out there that suggest the systems can only be so good or b) the Netflix algorithm team had some know-how. Perhaps they already knew they could improve their system by n percentage points, and gambled $1 million on competition producing a better result.  3) More than 10% was achievable, but the way the problem was framed limited experimentation.  Note that even well into the competition, there was a lot of collaborative discussion of ideas between the teams.  Basic ideas would be shared, and then different teams would implement them, perhaps sharing tricks discovered on the way.  Presumably this died down as teams approached the 10% goal (though the winner fused the two teams that had previously been closest).  This kind of setup has the potential to make everyone take the approach of trying to eke out a few tenths of a percentage point from essentially the same approach as they’d been using.  If so, that has the potential to create an artificial ceiling, as teams approach a local but not global maximum of fitness.

All of this is pure speculation by an outsider, but the phenomenon surrounding the 10% target caught my eye, and I couldn’t resist.  Bottom line question: could Netflix benefit from a second round of competition, or have they tapped out the resources available for $1 million?

Ty Lawson

I just found out that the draft is in two hours, so…good timing!

So Matthew Yglesias has a post about Ty Lawson, arguing that he should go higher in the draft than he’s projected to go (21st by nbadraft.net).  He was criticized for being short (6′0″), even for a point guard, and not fast enough.  I just don’t get the speed complaint, since Lawson was possibly the fastest player currently in college.

The height issue is possibly even weirder.  As for the height issue, the weirdest thing is that Yglesias’ comment thread didn’t mention the most relevant comparison.  Namely, three years ago, UNC sent Raymond Felton, an extremely fast 6′1″ guard to the NBA, and that’s…working out nicely.  They have a similar style and played the same system in college.  And Lawson has at least one thing going on Felton–his three point shooting was substantially better than Felton’s.  Actually, Felton had bad 3-pt shooting in college, and continues to be bad in the NBA, while Lawson is an outstanding three point shooter.  There are differences–Felton is actually a mutant who can just throw the ball through defenders to make a pass, and he’s probably faster than Lawson by a bit.  Still, Felton fully justified his status as the 5th pick, so it feels safe to say Lawson would justify a lottery pick.

Last fun stat, Lawson’s assist-turnover ratio? 3.48, which is just sick.  And it’s doubly important, because he always looks out of control, like he’s about to throw the ball in one direction and then careen into the stands.  But the stat shows that’s just an illusion.

Tacked on at the last minute: I see that Felton’s PER isn’t great, so maybe I was followed by superficially impressive stats.  Hollinger points out that the Bobcats made an obvious mistake playing Felton at the 2.  Or maybe I’m just out of my league trying to comment here–I don’t really follow the NBA, except sometimes during playoffs and to check on players I liked in college.

Watching the draft: When I saw that Minnesota had drafted Ricky Rubio, Jonny Flynn and Ty Lawson, I thought of the scene in The Dark Knight where the Joker faces three of a dead mobster’s henchmen, breaks a pool cue in half, and tells them he’s taking auditions, which will conclude in 5 minutes.  I think that’s a much better storyline than trading Lawson.

Iran

This is a sort of collection of what I’ve thought in the past week:

First of all, what I don’t understand.  Taking as a starting point Charles Krauthammer’s assertion that

This revolution will end either as a Tiananmen (a hot Tiananmen with massive and bloody repression or a cold Tiananmen with a finer mix of brutality and co-optation) or as a true revolution that brings down the Islamic Republic.

what I want to know is, why do we assume that any sort of sharp resolution has to occur? So long as the military stays loyal to the current regime, can’t the regime just stonewall, whether or not they resort to violence? The protestors aren’t themselves going to resort to large scale violence, and they have no direct way to force the regime to listen to them.  Now, I’m not sure I can think of any analogous situations where a repressive regime just waited out protests.  But repressive regimes do all sorts of unnecessary things, such as repressing people in the first place.

In any case, I don’t mean to suggest that this is likely to happen.  I’m asking a question: what’s your model of popular protest that rules it out? I don’t have such a model, and my interest is to stress skepticism: I don’t think we collectively understand the dynamics of this sort of situation.  Several commentators have suggested that the ultimate determinant is the military’s stance: will they eventually be willing to open fire on civilians and protestors? I think that’s right, but my question is prior: what causes the situation to develop to a point where the military faces that choice? (Since I started composing this paragraph Friday night, we’ve seen violence, and a cessation of protests, but not the full confrontation that was threatened.  Sadly, it’s good news that at most a few dozen have died.  So I think the question remains interesting).

Note that the Tiananmen protests lasted some seven weeks before the massacre.

Second, much of the debate within the United States has concerned how Obama should react to developments.  During the campaign, various conservative voices said that Obama could not change our relationship to the world merely by being conciliatory.  In fact, this argument had substantial merit, as some people really did have a mental picture where it was just President Bush who had spoiled our relationship with the world, despite myriad reasons why we will remain antagonistic to many countries.  What I don’t understand is how that’s compatible with the current calls for Obama to denounce the Iranian regime.  Does America have influence only when we offer context-free condemnations, but no influence when we negotiate?

But I oscillate between that very realist assessment that nothing Obama says can help, but could hurt, by offering Khomeini a soundbite, and the thought that he could have said more.  Of course a direct statement to say that Mousavi ought to be prime minister is out of the question (and undesirable for other reasons as well).  All comments should be framed in universal terms, in terms of the values that make a society democratic.  However, on top of condemning violence, Obama could have commented on the underlying circumstances.  Violence or no, the suppression of protests is illegitimate, and incompatible with the ideals of free elections.  The same is true of internet censorship as well as restricting the presidential candidates.   Would it have been impossible to frame a statement that spoke up for those values without playing into the hands of Khomeini? That’s not obvious to me, and I don’t know why someone would be confident that it is.  Bear in mind that we can afford to offend the Iranian regime as much as we want to.  It’s only if we say something that angers the Iranian people that there could be negative consequences.

Lastly, while I understand Will Wilkinson’s hesitation, I have to disagree with him about our investment in the welfare of the Iranian people.  He writes:

When people feel pressure to signal, and it’s free, they’ll signal. But sending the signal creates a small emotional investment in the overt message of the signal — solidarity with opponents of the ruling Iranian regime. As every salesman knows, getting someone to make a big, costly commitment is best achieved by getting them to first make a tiny, costless commitment. The tiny, costless commitment of turning Twitter avatars green is thin edge of the persuasive edge for the neocons who would like to sell the public a war in Iran. Since I would rather not be Bill Kristol’s useful idiot, I will conspicuously leave my avatar as is, and continue hoping for the best.

It sounds as if he’s saying “don’t care about the fate of the Iranians because there are people who’d abuse that concern to push bad ideas.”  But that’s a problem with the abuse, and a reason to reject militarism.  Nor does it make sense to say that we shouldn’t be concerned because we can’t act.  If there were a way to prevent violence and repression against the Iranian people, we would have to consider it.  The green twitter icons were always meant to be symbolic.

The reason I wouldn’t “go green” is that I’m not so sure I’m willing to endorse the message.  Green is the color of Mousavi’s movement, and it symbolizes the entire movement, not just the desire to protest peacefully.  So far as that’s what they’re about, we should be in solidarity with them.  His label of ‘reformist’ does not mean that we should be interested in fully supporting him.  Even ignorance would make me wary of the symbolic import of “going green”, and what’s worse, Mousavi’s history within Iran doesn’t make me confident that he’s a model of liberal values (Will has a followup that mentions of these concerns, but I don’t think it justifies his original stance).

17th Century Productivity Blogging

You know that Munday is Sundayes brother
Tuesday is such another
Wednesday you must go to Church and pray
Thursday is half-holiday
On Friday it is too late to begin to spin
The Saturday is half-holiday agen.

That’s by Merlin Mann some moralist, writing in 1639, which I found in “Time, Work-Discipline and Industrial Capitalism,” the E.P. Thompson paper that Julian recommended. Thompson soon offers as an aside:

The work pattern was one of alternate bouts of intense labour and of idleness, wherever men were in control of their own working lives. (The pattern persists among some self-employed — artists, writers, small farmers, and perhaps also with students — today, and provokes the question whether it is not a “natural” human work-rhythm.)

I am so glad to hear that.

Thought Of The Day

Richard Nixon makes me wonder how stable our democratic institutions really are.

The Weekend

This 1991 Atlantic Monthly article “Waiting For The Weekend” by Witold Rybczynski is interesting throughout, even if it doesn’t fully substantiate its most ambitious claims.  It’s worth reading if just for the depiction of work culture before and after the Industrial Revolution, which is quite fascinating.  I’d known that drunkenness on the job was common before the Industrial Revolution and remained a problem throughout it, but I’d never heard of Saint Monday, for instance.

I found the article via the Atlantic’s new ideas blog, which is being written by Conor Friedersdorf–more about which later.

Update: In a comment to this post, my friend Julian suggests a very nice piece of academic history on changing conceptions of time and work for those who want to read more.

What Are They Thinking?

I have a long Iran related post that’s mostly done, but we’ll have to wait until at least tomorrow morning (at which time I’ll pray that it hasn’t become outdated.  That’s not because my blogging is so important, but because my post will be outdated if the violence escalates).

However, I wanted to highlight a NYTimes report: Iran Admits Discrepancies In 3 Million Votes.  Specifically, the votes concern polling areas where more votes were cast than there were voters.  The Guardian council has admitted that this occurred in at least 50 cities, though Moussavi has alleged it was 80-170.

What I do not understand is how they can maintain their stance in light of this admission.  In reading the article, I found that they have not strictly admitted fraud–they gave the explanation that individuals were legally allowed to vote other than where they were registered.

Still, to admit such a large discrepancy without calling for major investigations beggars belief.  For more than a week now, the Iranian regime has been acting unjustly, but this is the first time I have been baffled by their actions.

Quote Of The Day

So there’s a characteristic pseudo-self-sufficiency to concessive theory – a characteristic pseudo-completeness. (A feeling that when you’ve said all there is to say about how to be an effective leader, or an effective person, in a rather instrumental way, there isn’t anything left to be said.) There’s also a characteristic bait-and-switch quality to it, in that it is normative, so it is easier to miss that it may not be normative in all the ways that we need it to be. (John Holbo)

A Small Gift To The Iranian People

It’s a very small intervention, but I think that this story is damn cool. A State Department official emailed Twitter co-founder Jack Dorsey on behalf of the Obama administration to request that they change the timing of scheduled maintenance so that it would fall during nighttime in Tehran.

Reading Deficit Tea Leaves

This is old news, but it particularly bothers me:

The administration’s tax proposals call for hiking the top two tax rates from 33 and 35 percent to 36 and 39.6 percent and raising the threshold to get into the new 36 percent bracket. For couples, that bracket would start at $231,300 in 2009, up from $208,850; the starting point for singles would climb from $171,550 to $190,650. (Via Atlantic Politics)

This is not a huge quantity of revenue, relative to the deficits that we’re facing, but it seems particularly senseless.  The cut goes beyond Obama’s campaign promises, which were already too generous, and it is a cut–individuals in between the $208k and $231k have their top marginal rates reduced from 33% to 28%.

Part of what is bothersome is the nature of deficit politics in 2009.  Direct responses, like cutting spending or raising taxes are likely to have a procyclical impact which would deepen the present recession.  But the deficits, both current and projected, are huge, and will become an increasingly large problem unless dealt with.  Of course the administration says it had a plan, but so did Bush, and we know how that went (Megan McArdle made this point here).  Small points like this one indicate something about how the administration sees the issue and their level of commitment.  This particular decision says “we are going to pander to absolutely everyone.  We can’t see past tax ads in 2012.”

Followup: Here’s a Matthew Yglesias piece for the American Prospect detailing some of the tax problems that the Obama administration faces.

Rematch

Iyama Yuta will once again play Cho U for the Meijin in the fall. The league hasn’t actually finished, but Iyama is 6-0, with two games to play.  Yamada Kimio, Takao Shinji and Sakai Hideyuki are 5-2, 4-2 and 4-2 respectively, but they’ve each lost to Iyama head-to-head.

It’s potentially  something of a weird year for Iyama.  It’s his second year with a really high level performance, and he leads the Japanese win-loss standings, with a 22-7 record.  And yet the Meijin will be the only title he’ll challenge for this year, after narrowly missing the chance to challege for the Kisei and Gosei titles.  Whether or not he wins the Meijin he’ll be the top of his age group, but it would be strange for 2009 to essentially be a repeat of 2008.

Also interesting, Ri Ishu, a 20 year old player, is third in the win-loss standings, with a 17-2 (!) record.  He just earned a 7-dan promotion for entering the Kisei league, making him one of the young players to watch.

July 9th: Iyama ended up sweeping the league 7-0.

Who Is Responsible For Doing “Something Else”?

In his essay on the passing of many great figures of post-WWII analytic philosophy, the professionalization and specialization of philosophy, and the shift towards philosophy as a vocation, Brian Leiter discusses and rejects Rorty’s view of the field:

…Richard Rorty (1931-2007), who felt so confident that the philosophical problems of millennia –about the nature of truth and knowledge – were misguided and pointless (all cases of chmess?), that he wanted philosophers to give up on the traditions running from Plato to Hume to Hegel, to do “something else”. What the “something else” amounted to was, alas, never very clear – and why philosophers should have been thought especially competent to do it even less so. One doesn’t need to be trained in the history of philosophy, after all, to produce pablum about “liberal” patriotism, as Rorty, sadly, ended up doing. Rorty may have garnered a substantial following among those who knew little about philosophy and its traditions, but it is a safe bet that, in the era of intellectual specialisation, his actual influence on philosophy will be short-lived.

My view of philosophy doesn’t follow Rorty’s anymore, if it ever did.  I’m at most a common traveler, as far as I suspect that the field may have an inaccurate image of what it can hope to accomplish.  And do pay attention to that “suspect…may have…” in the previous sentence.  There’s far too little that I understand to be confident about what type of philosophy needs to be done.  So when I speak for Rorty, in what follows, remember that I’m not speaking for myself, and I may be distorting what Rorty would’ve thought.

Those caveats aside, I’m doubt that Leiter’s charge is particularly accurate here.  Actually there are three charges: that it was never too clear what “something else was” (I agree with this–the closest I ever saw were hints in Contingency, Irony, Solidarity); that it wasn’t clear why philosophers were the ones who should be doing that particular “something else”; that Rorty’s own later writings often looked like liberal pablum (this hits close to the bone).  It’s the second charge, that current philosophers might just be ill-suited to doing work that interests me.  I’m not sure it’s as damning as Leiter thinks it is.

Of course the people who are now in philosophy departments were selected because their abilities match the work now being done, and there’s no guarantee that they’d be suited to doing something different.  But if the work being done in philosophy departments is misguided, it’s no defense to say that the people in philosophy departments couldn’t do the truly important work.  It often happens that developments in a field leave its old practitioners ill-equipped to contribute.  I wouldn’t be surprised if the genuinely important developments in logic in the early 20^th century cut many people with important insights out of the philosophical conversation.  An opposite phenomenon concerns philosophers who have a great insight, but struggle to show the importance of that insight given contemporary understanding of the field (you might think of Frege’s work here).  I think it’s just uncontroversial that the skills that make a good philosopher are not always those that characterize the field at a particular moment.

Rorty certainly should’ve conceded that a lot of people who had trained as philosophers had gone into the wrong field, so that his appeals concerning what philosophers should do are only aimed at some people in academic philosophy.  Perhaps others should have been linguists, or scientists, or mathematicians, but came to believe that there were more important questions to be answered in philosophy.  Refocus philosophy on the sort of humanistic concerns that Rorty cares about, and they’d stop wanting to be philosophers (not that humanistic insight is essentially antagonistic to interest in science).  On the other hand, there are also people now in philosophy departments who have been shoehorned into doing the kind of work that has a high profile, but would be more comfortable under Rorty’s new regime.

More speculatively, I wonder how Rorty’s exact target makes a difference.  I would think that often it’s not so much the questions that philosophers are investigating per se that Rorty objects to, so much as the framing that attaches to them (but maybe Rorty rules out this reading).  There are obviously sensible questions about the word ‘know’ that we can ask, and they might be very important questions, even without philosophical motivation.  I’m confident that there are many examples of this type of question within philosophy of science.  All that Rorty should do is deny that answering these questions gives us any particular philosophical insight.

Or perhaps it’s the case that the philosophical projects attached to our investigations of the word ‘know’ bias our research so that we can’t contribute to the important non-philosophical topics at hand.  But no matter what happens to philosophy, qua topic critiqued by Rorty, there would be fruitful research that could be done, perhaps even by current philosophers.

Leiter ends on a deflationary note†:

There will, we may hope, continue to be world-historical geniuses – “seers and prophets” as it were – but there is no special reason to look for them in academic departments of philosophy anywhere. From those departments, we may quite reasonably look for incremental contributions to understanding of our sciences, of our moral and political lives, and of our language and our mental capacities.

The question about why such prophets would be found in academic departments is hard to answer, especially if you think that Rorty never managed to articulate the kind of work that his future philosophers should do.  But that does not prove that philosophy departments should continue to go on in the same way.  Perhaps there should not be philosophy departments, or perhaps they should shrink drastically.  Philosophy departments and philosophers will always need to reflect on what makes their work have philosophical value.  Rorty’s charge is that all the current answers are bad ones.

† Being fair, Leiter is not necessarily attributing the view that seers and prophets would be found in philosophy departments to Rorty.

A Cute Coincidence

I just today found out that Khoi Vin, the design director for nytimes.com since 2006, was also responsible for the 2005 redesign of The Onion’s online newspaper.  In that design project, he said the hardest task was to make the site look like an actual newspaper.

If the results look suspiciously like a green version of The New York Times Online, it’s because we spent a lot of time studying how the Gray Lady delivers news — but I like to think we were conscientious enough not to steal crassly.

Comment Spam Keeps On Giving

Start making out with the person at the terminal next to you (It helps if you know them, but this is also a great way to make new friends).

“if” was a link to an appliance website.

Time Moves In A Circle On Cable News

During a long layover in Philadelphia Tuesday, I had the chance to watch a lot of Wolf Blitzer.  This is a surreal experience, since he apparently pitched a four hour show with the idea that no one would watch more than half an hour at a time.  Viewers come and go, so Blitzer just repeats segments of the show.  We get Sonia Sotomayor’s nomination, the kid who ran away from chemo, prop 8, and occasionally Obama’s snub from Nevada (or is it vice versa?), and then a repeat.  Each segment has different talking heads phone in, but it’s pretty hard to tell them apart.

By all objective standards, Blitzer is a privileged member of American society, but this sounds like a horrible life.  He’s get’s a ton of money and in exchange acts like a political version of Didi and Estragon.

Nor is he exactly rescued by the content of his show.  Another way he might have pitched the show is “no argument too stupid to consider.”  Yesterday, the subject of Sotomayor’s diabetes was mentioned during each segment–the thought being that it might impact her longevity on the court, but her diabetes is apparently well-controlled.  And as far as I see, the only people who’ve even mentioned her diabetes are mainstream media figures like Blitzer who’ve decided, on who knows what grounds, that it’s newsworthy.

Dum Dum Da Da

This Saturday, May 16th, I’m going to be getting married. I don’t know if I have a good, or at least short, way to describe what this means to me.  I think it’s enough for now to say that this will be the most important day of my life so far.  Perhaps the most important day ever.

So I’ll confine myself to commenting on how this event affects my blogging. I’ve set this entry to appear as I’m leaving for Raleigh, and leaving without my computer.  Instead I’ve settled on a few books and my kindle, all of which is pleasure reading–Murakami’s Kafka on the Shore being the highlight.  I’ll be gone until the 25^th and I won’t be posting during that time. On the other hand, after May 16th, I’ll consider myself, based on personal experience, to be an expert on any marriage related topic.

Opinion Polls For Philosophers

People within philosophy will probably have already heard of this, but my non-philosopher readers might not.  There has been a series of polls run by Brian Leiter concerning the top philosophers of various eras.  For short results, see the master list of modern philosophers. There have also been polls on the early modern period and the past two hundred years.  The lists is likely to be of merely sociological interest, reflecting what professional philosophers think, but they’re still a nice curiosity.  Note: so far as the term makes sense, those voting on the polls are likely to be ‘analytic’ philosophers.

Obama’s Five Day Rule

During the campaign, Obama promised that he would wait to sign legislation until the text of the bills had been placed online for at least five days time.  Since taking office, he has been 1 for 11 on the promise.  It’s easy to think that the promise was mere window dressing, so that the real fault was that Obama had made such a frivolous promise (bullshitting the electorate being a venial sin).  But Cato’s Jim Harper makes a compelling case to the contrary (h/t Radley Balko):

It is easy to dismiss the five-day promise as an idea that would not have changed much anyway. Bills coming out of Congress are faits accomplis, aren’t they? They are not.

Members of Congress are highly skilled political risk balancers, and the president’s firm insistence on leaving bills sitting out there, unsigned, after they pass Congress would have a significant effect on congressional behavior. It would threaten to reveal excesses in parochial amendments and earmarks, which could bring down otherwise good bills. Recognizing the negative attention they could draw to themselves, representatives and senators would act with more circumspection, and last-minute add-ons to big bills would recede. A firm five-day rule at the White House would also inspire the House and Senate to implement more transparent and careful processes themselves.

Yet Another Reason Not To Make Bad Arguments

Megan McArdle replies to Conor Friedersdorf’s criticism by arguing that the problem with arguing that torture doesn’t work is that it undermines your credibility when you turn out to be wrong:

You are interested in proving that torture doesn’t work as a way of forcing them towards your moral conclusion.  They will rightly suspect that your investigation of the factual question is not likely to be of a high quality.  And indeed, that is what I find in these arguments: people wildly overstating an at best modest case that torture rarely produces all that much usable intelligence. [my emphasis]

While I can understand what might tempt you to write that passage, it just flat out begs the question.  Of course one shouldn’t present an wildly overstated arguments.  If you do, then your credibility should take a hit.  But that point only matters if the arguments that torture don’t work are obvious crap.  And no one making those arguments is going to think they are.

Now, many anti-torture commentators are being careless in their argumentation, and they might be letting their moral position confuse their factual judgment.  In fact, I’d be astonished if that weren’t the case–it’s not as if human rights advocates are free from ordinary human frailties.  What I don’t get is how the answer to that fact is to cede the field of debate.  Then torture advocates get to assure us that torture is extremely effective, and the only way to safeguard America, but surprise–I don’t think they’re dispassionately evaluating the evidence either.

There some kind of extra scrutiny that you have to apply to yourself when you find that you’re arguing for something counter-intuitive that also would be very convenient to believe.  Read charitably, that’s McArdle’s point, and she’s overstated it.  But that doesn’t excuse you from exercising your judgment–you have to actually evaluate the evidence.  If, upon reflection, you find that you’re unable to make a confident judgment, then you bow out of that particular debate.  If you find that the evidence is completely equivocal, so that you suspect people will just read their biases into it, then you say that fact, and push back against anyone who is being confident, one way or another.

More Torture Revelations

Thanks to the release of the torture memos, debates about torture have gotten quite interesting and urgent.  And to someone like me, who believes that the United States has crossed a bright line since September 11th, there is very disconcerting news: Admiral Blair, Obama’s director of national intelligence, believes that the use of torture produced significant information about Al Qaeda (based on an informative visit to the corner [!!]). It’s all the more noteworthy, as Blair had argued that torture was not just immoral but also ineffective during his confirmation hearing. That suggests that he has had an about face during two months of service, based on classified information that he has seen.

Now, Blair is one man, and we can’t know the basis of his judgment.  We are not obligated to take his opinion as gospel. We could find someone even more credible than Blair who’s full of crap.  Or someone could change Blair’s mind, or provide a knock down argument that torture doesn’t work.  Still, this should give us pause–Blair knows more than us, and he’s not part of the raving pro-torture crowd.

How much does this matter? Putting my cards on the table, I’m one of the people Megan McArdle is describing (but see Conor Friedersdorf demolishing the conclusion she draws based on that characterization).

Most people who make this argument do not, in fact, care whether torture works.  They would still be every bit as much against it if waterboarding worked perfectly.

That’s me.  I’ve never felt in my gut that there was any reason to torture people to try and get intelligence.  Maybe if I bought into the idea that this was WW3, and the United States faced an existential threat, I might be conflicted. As it is, I’m not.

Returning to a pair of purely analytical points, without trying to say whether or not torture is effective, I think McArdle makes a framing mistake in her opening:

the evidence for those arguments seems empirically shaky, especially since many people employing them insist on arguing that torture basically never works, rather than that it doesn’t work very often and therefore has a bad cost-benefit ratio.

This seems to just assume that they payoff of torture is non-negative, as if it might be zero, but we couldn’t end up worse off.  Even from a strictly intelligence gathering perspective, I think that’s wrong.  Torture will sometimes produce useful information, but interrogators would have to recognize it as such.  That means a potentially very complicated process of confirming or disconfirming the evidence provided.  If the process of confirmation is itself subject to noise or error, it’s no guarantee that we won’t end up trusting bad information produced by torture (can I assume that those interrogators and units that choose to torture are the ones most likely to trust the technique too much?)  Lastly, I think that even under the most extreme kangaroo-courts instituted during the War on Terror, I believe information derived from torture was excluded from trials.  So someone who wants to argue that torture works still has to face the real possibility that it’s just useless, instead of assuming that the question is how much it works.

And this is all before we touch Jim Manzi’s point that torture could be tactically wonderful,but strategically awful.  I’m less convinced by his historical examples–wasn’t World War II decided by military strength and industrial production? But today, the most important question is how many people want to join terrorist movements and how many countries are willing to cooperate with us, and surely those are two areas where torture can only hurt us.