2011 04 03
Small Sample Size Theater

Posted by in: Baseball, logic, Math

The baseball season has begun! Each MLB baseball team has played two games. There isn’t a lot of solid trend data to report on, yet articles must be written – and so, quoth S – “it’s time for another edition of Small Sample Size Theater”.

In baseball, of course, this means things like:
The Mariners, predicted to be terrible this season, are tied for first in the league!

(Also there are nineteen pitchers tied with an unbelievable 0.00 ERA. This season looks set to turn a lot of conventional wisdom on its head.)

We see Small Sample Size Theater in other domains as well; no surprise that most trend reporting is of this type. I wanted to post this today because I think the term is so apt. And of course, if my posting this year keeps up at this rate, I’ll post well over 300 entries, which would more than double my previous record. In year seven, anything is possible.

Nada (0)

2010 01 17
Recently read: Clearing out the Backlog Edition

Posted by in: Books, Brooklyn, Math, Programming

Peter Siebel. Coders at Work: Reflections on the Craft of Programming

This superb book is a collection of fifteen interviews with well-known and highly-regarded programmers (Norvig, Armstrong, Knuth, etc). Siebel (author of Practical Common Lisp) is a professional programmer with a keen sense of the (brief) history of the profession. This gives the interviews a depth and a richness that even a clever journalist could never have matched. Siebel is a consistently thoughtful interviewer who asks just the right mix of questions. In any one interview, the questions range from practical ones concerned with how the subjects debug code to more general questions about whether the nature of programming has changed over time. Across interviews, Siebel asks enough of the same questions that we can start to view the answers in comparative perspective, while also allowing what is special about the careers and interests of the subjects to emerge.

In short, if you’re interested in programming, this book is wildly engrossing. A word of warning: If you don’t have any experience programming, and some background knowledge of the field, you’re probably not going to be able to get much out of the book. Some passages were certainly over my head, as I’ve only been a professional programmer since June, when I got my green card, and if I recall correctly, only really got started teaching myself Python about a year and a half ago. But most of it was accessible and inspiring to this junior programmer.

Amy Sohn. Prospect Park West

We lived briefly in (very South) Park Slope when we first moved to Brooklyn, and although we’ve since moved out to Flatbush, we’re back in the Slope all the time. We eat at Al Di La whenever we can afford to. We’ve been members of the infamous Park Slope Food Coop for several years now, and we’re set to have a baby in the Spring. So although my expectations weren’t all that high, I pretty much had to check Prospect Park West out of the Brooklyn Public library, after waiting patiently for my turn in a queue that was over 250 holds long. Prospect Park West is set against this familiar background. The plot follows the ill-considered affair of a Park Slope mother, whose life is connected to a few other characters by a string of coincidences that I would have found far-fetched ten years ago, before I started to notice equally striking coincidences in my own life. (Always remember that odds are that life will be filled with the improbable, since there are an enormous number of possible improbable events—so many that it would be highly improbable for us to go long without another improbable event occurring. This is one reason, among several, that life is filled with strangeness and magic, if you keep an eye out for it.)

Prospect Park West is not a great work of literature, but it’s readable enough. The book’s basic outlook is misanthropic without much in the way of compensating insight. I get that some Park Slope mothers can be a bit much, but so can the author when she (in the mouths of her characters) gets going about them. The author gets points, though, for her depiction of the strange, confusing, prickly racial tension you run across in Brooklyn all the time, and which I struggle to explain to my friends back in Canada. This too was perhaps also a bit overdone, but unfortunately not by much.

One correction: A check out line at the Coop that stretches back to the bread section does not count as long. I don’t know when Sohn shops, but that’s pretty routine in my experience. Long is when it goes all the way along the produce aisle as far back as the milk section.

Charles Dickens. Oliver Twist

This is only the second Dickens novel I’ve read, the other being A Tale of Two Cities. I found A Tale of Two Cities pretty silly, but against my better judgment found the ending weirdly sublime. I didn’t have as much luck with Oliver Twist, which I read for the sole reason that we’re naming our kid “Oliver” and I figured I should at least read the book that helped make his name famous. (On my to do list: Who the hell is Oliver Cromwell?) I found the social commentary in the first part of the book entertaining enough, if heavy-handed. But as the plot advanced, the melodrama and the general absurdity of it all started to suck the fun out of it. Also, I know the book is a product of the early nineteenth century, but the fact that one of the characters is usually referred to simply as “the Jew” and even gets to be the butt of a big nose joke was driving me nuts. What’s that? Dickens was a child of his era, so cut him some slack? Well, I’m a child of my era, so take your own advice and cut me some slack while you’re at it.

Vivant Denon. Introduction by Peter BrooksNo Tomorrow

Vivant Denon was, among other things, the first director of the Louvre Museum, in charge of sorting and cataloging all the goodies that Napoleon stole from the Egyptians. A wing of the Louvre bears his name to this day. Denon was also “maybe, probably,” in the words of Peter Brooks, the author of No Tomorrow a thirty odd page long erotic masterpiece. The New York Review of Books has recently published a fine bilingual edition of the story with an introduction by Peter Brooks. The intellectual imprimatur provided by the publisher and the scholarly introduction makes it totally not skeevy that I’m writing about erotica on my blog.

There’s a lot to admire in Denon’s story and the way he tells it. As for the tale, a woman seduces a man, for pleasure, without negative consequence for either. As for the telling, Denon is delicate without ever being prudish, erotic without ever being explicit. It’s good clean fun for the adults in the family.

Surendra Verma. The Little Book of Maths, Theorems, Theories, and Things

This book covers a very wide variety of mathematical and logical puzzles and problems and more. The author even throws in a discussion of the Body-Mass Index*, presumably because it’s . . . expressed in numbers? Because it’s a little book, and because it’s trying to get to so many subjects, and because the author also likes to throw in limericks and factoids and anecdotes willy-nilly, this book treats each of its subjects in an extremely superficial way. I like limericks and factoids and anecdotes as much as the next guy, but there really wasn’t room for a lot of math in this book, or much opportunity for the author to make the case that mathematics is intrinsically interesting.

Let me also take a moment to plead with the publisher to fix the typos in this book before reprinting, if the book ever gets another shot at life. You know you’re in bad hands when you read the sentence: “No one has ever found an even number that can be expressed as the sum of two prime numbers” (p. 76). Oh, really? Cause I think I might be about to make mathematical history!

* Verma tells us that knowing your BMI “can give you an idea of how healthy your weight is.” He doesn’t note that a lot of researchers think the BMI is misleading or useless.

Howls of outrage (6)

2008 09 04
Lies and Damn Lies

Posted by in: Math, Pundits, U.S. politics

(Note: I apologize for breaking Explananda with this post last night — I let my enthusiasm override my remembering-how-to-post.)

I suppose it’s fitting since my last post was right before the 2004 election, but I never thought my return would be about politics. Well, kind of.

For various reasons, including the interesting Obama/Clinton delegate math, I’ve been following this year’s election in greater detail than any in the past. Which, unfortunately, means I’ve been reading a lot of political articles. In the course of my travails I came across a particularly egregious example of the mis-use of statistics that got me worked up enough that I had to write about it somewhere. So here you go.

I found this in a post yesterday by Peggy Noonan at

I’m bumping into a lot of critics who do not buy the legitimacy of small town mayorship (Palin had two terms in Wasilla, Alaska, population 9,000 or so) and executive as opposed to legislative experience. But executives, even of small towns, run something. There are 262 cities in this country with a population of 100,000 or more. But there are close to a hundred thousand small towns with ten thousand people or less. “You do the math,” the conservative pollster Kellyanne Conway told me. “We are a nation of Wasillas, not Chicagos.”

The worst thing is that this even passes the sniff test; 262 times 100,000 is way, way less than 100,000 times 10,000, so it does seem like there are more people living in small cities than big cities.

Except: the first alarm should go off when you look at the numbers for small cities. 100,000 times 10,000 is 1 billion, and the population of the US is just over 0.3 billion. And of course, when you actually get to the facts, you find that over 58% of the US lives in cities with over 200,000 people.

The trick is using a floor for the number you want to minimize, and a ceiling for the number you want to maximize. The counting in the quote above counts New York City as a city of 100,000, and counts Eastport, Maine (my ancestral home, population 1640) and many other towns with population under 1000 as cities of 10,000.

Another example of this fun statistical manipulation: only 1% of the US population has a household income of over $400,000, but over 50% of the population has a household income of under $50,000. Clearly, the evidence show that most of the wealth in the US lies in the hands of working families.

Howls of outrage (5)

2008 07 27
The Monty Hall Problem

Posted by in: Math, Programming, Python

Thanks to a friend, this morning I learned about the Monty Hall Problem. As she remarked, it is counter-intuitive in the extreme. But I see from the Wikipedia article that even Paul Erdos got it wrong the first time, so I don’t feel too bad about being initially stumped. (If you’re having trouble getting it, I found it very helpful to step back and think about the related N doors puzzle discussed in the Wikipedia article.)

One of the wonderful things about picking up even the slightest bit of programming is that you can test and play around with things like this. The Python programming language makes it especially easy for a beginner to muddle through to a test very quickly:

import random

remainingchoice = []
carcount = 0.0
trials = 100000

for i in range(1, trials):
    possibilities = ['goat', 'goat', 'car']
    if possibilities[1] == 'goat':
for item in remainingchoice:
    if item == 'goat':
        carcount +=1    

print (carcount * 100) / trials

Somehow makes it all seem more solid. Except when I changed the trials variable to 100000000 and my computer was all like “What the fuuuu?” and then Python crashed hard.

Howls of outrage (24)

2008 07 25
G. Polya’s How to Solve It

Posted by in: Books, Math

G. Polya. How to Solve It: A New Aspect of Mathematical Method

This is a book about heuristic, the study “of the methods and rules of discovery and invention,” in which most (but not all) of the examples are drawn from mathematics. Polya is interested in the question of how we go about solving puzzles in general, and, having acquired a facility with problem-solving, how we then go on to teach others the same skill. There’s no straightforward algorithm for problem-solving, but there are general patterns. As Polya never tires of reminding us, we typically need to ask ourselves: What is the unknown? What are the data? Do we know a related problem? Can we use this problem in the solution of our current problem? And so on. These might sound obvious, but there’s a value in having them stated clearly, and significant value in some of Polya’s imaginary discussions with students who are walked through the solutions to puzzles by a teacher making intelligent use of Socratic questioning.

Since it’s about thinking in general, and not just mathematics, I think this book might be read with profit by most people. I imagine it would be especially useful for mathematics teachers, especially because the author clearly has a keen sense of pedagogy. Unfortunately, many of the mathematical examples were a bit over my head, since I’m awfully rusty these days. And the book is marred, in my opinion, by an unconscionable amount of repetition. Still, on the whole a decent book.

Nada (0)

2008 04 28
Recently read

Posted by in: Books, History, Math

Tom Slee. No One Makes You Shop At Walmart: The Surprising Deceptions of Individual Choice

This book is not really about shopping at Walmart, but I’ll start there anyway. Suppose you’re grousing about Walmart. It wants to move into your town and you’re worried about the effects on the local economy. Or it already has, and you think the effects you warned about are already becoming apparent. There’s a good chance that the person you’re grousing to is going to point out that no one makes anyone shop at Walmart. Indeed, if your interlocutor impolite enough she may even raise the awkward point that she saw you just last week emerging from that very store. If everyone, including you, shops at Walmart, what better evidence, then, that the community has collectively decided to welcome Walmart by making a series of individual choices to support it?

The answer, according to Slee, is no, and his book is a very careful and methodical cataloging of some of the most important ways in which choice is more complicated than the view sketched above suggests, which Slee calls MarketThink. Very briefly, our choices are not made in a vacuum. We make the choices we do while responding to agents who are making choices of their own, which choices themselves are in part responses to our own choices or what they anticipate will be our choices. And in choice situations of this sort, it is often the case that every individual agent makes choices which are perfectly reasonable from where she is situated, but which lead to outcomes which no one involved would prefer all things considered.

Preference, then, turns out to be more complicated than it first appears. Sometimes “preference” refers to what an agent chooses from the options available to her, given the choices that other agents are making or intending to make. Sometimes, by contrast, it attaches to the outcome which the agent would prefer. A great deal of Slee’s book is taken up with explaining how and why these two notions of preference frequently come apart.

Readers who know even the most basic game theory will know that Slee is not just being modest when he claims in the preface that there is little original in his book. I’m not sure I actually learned anything new from this book, since I’ve already read my Rapaport and my Axelrod and my Sen and so on (and I really haven’t read much beyond that). Even so, I found it well worth my time. There’s real value in having the inadequacies of MarketThink detailed for two hundred pages of marvelously clear prose. I think that I was already predisposed to agree with virtually everything Slee said, but reading him made me aware that I do sometimes slip into versions of MarketThink when I clearly shouldn’t. So, the book was edifying and entertaining at the same time.

It’s worth noting, though I don’t intend this as a criticism, that the book is almost entirely negative. That is, it’s about the inadequacies of MarketThink, and not about how to correct for those inadequacies. That’s just fine with me, but I found myself wondering from time to time how an intelligent libertarian would respond to Slee’s main line of argument. I suspect that an intelligent libertarian would have to concede that MarketThink, as Slee depicts it, is crude and inadequate. But a libertarian version of Slee might just as easily write a whole book, also drawing on economics and game theory, to show how regulation and intervention in the market often leads to unintended and frequently unwanted results. The failure of MarketThink does not automatically establish the soundness of any alternative, of course. Now, I think the response to this point is to try to get more specific about exactly what interventions are warranted and how we propose to avoid unwanted consequences. But that just means that the argument goes on (and Slee would surely agree). But thanks to Slee (and the people whose work he draws on) the argument ought to go on without silly appeals to MarketThink. And that’s an advance worth celebrating.

Nassim Nicholas Taleb. The Black Swan: The Impact of the Highly Improbable

What a strange book! This book engages with an extensive literature on risk, probability and the psychology of risk and probability assessment, but it does so in a most unacademic way. Taleb tells stories, engages in autobiography, subverts our expectations about the relevance of the autobiographical passages, harangues, insults, scolds, relates his fantasies about humiliating rival thinkers, bullies, ends expository paragraphs with “Capiche?”, pleads, and repeats himself again and again. Taleb writes in a rough, informal, and highly idiosyncratic way. I cannot imagine that the editor of this book wanted the book as it is in its final form, and I sort of get a kick out of imagining Taleb forcing them to accept it anyway (I’m sure they made money on it nonetheless). He must be a serious pain in the ass to work with.

Anyway, Taleb’s basic idea is that we – human beings, that is – are incredibly bad at assessing risk. Our models for risk assessment tend underestimate the impact of the highly improbable. But marinate on this, brother: There are so many possible highly improbable events that it is highly probable that the highly improbable will intrude, and intrude very messily, into reality, and blow all our little models to bits. We just don’t know what they’ll be. We live in a world dominated by the highly improbable, and most of our risk models are worse than useless: because they encourage us to think we’ve got a handle on things, they make us even more vulnerable to extraordinarily improbable events when they do occur. Taleb doesn’t just confine himself to the world of finance, where he made his living dealing with risk, in order to illustrate this point, preferring to range over a much broader field of history in search of arguments and examples.

So why we do this? Much of Taleb’s book is a meditation on this very question, and, I think, a very useful one. When we look at past events, we have a tendency to slip into narratives that make events seem to follow one another in a natural and expected way. This encourages us to think that, going forward, events can be expected to follow one another in a natural and expected way. It isn’t so. And if you actually look at the track record of experts in various fields, you’ll find them regularly getting blindsided by events which were anything but predictable. And if you actually listen to the experts defending wrong predictions after the fact, you’ll regularly hear them defending their predictions in the following form, “But I was exactly right in my prediction, except that X,” where X is something highly improbable that unfortunately threw everything off. But if this happens again and again, we ought to stop and wonder about the value of such predictions in a world that serves up so many Xs.

This insight is not just valuable for people making their living predicting the future, in my opinion. I think this book should be required reading for anyone doing historical work, which, in my experience, frequently a) falls into the habit of ad hoc explanations which oversimplify reality, and b) often attempts conjectural reconstructions of the past on the basis of mere plausibility, again in a way that I think grossly oversimplifies matters. (I have a post in draft now illustrating (b).)

There’s more – much more – to Taleb’s story about the human propensity to underestimate the potential impact of the improbable, which I won’t go into. I will say that Taleb seems to me just a bit too enamored with stock evolutionary psychology explanations – which is funny in a book about how prone we are to manufacturing bogus ad hoc explanations, since the same vice is pretty common in evolutionary psychology, in my opinion. But whatever. There’s still a lot of great material here. And I found myself grateful for the repetition in the end. The first twenty times I read Taleb complaining about bad excuses for predictions gone wrong I nodded my head and thought “Yeah, sure.” But about the twenty-first time I thought “Holy fuck! That is so true. I do that too!”

Anyway, if you’re curious here’s a Malcolm Gladwell piece on the guy from the New Yorker. And here’s his (ugly!) website, where you can get a taste of how feisty and combative he is, since he appears to respond to all the reviews his books have gotten. I was amused to see his response to Gregg Easterbrook, who is seriously the dumbest fucking guy ever.

Henry Fielding. Joseph Andrews

Not nearly as good as Fielding’s Tom Jones, but then practically nothing is. I found the first fifty and the last fifty pages awesome, with some pretty plodding material in between them. Fielding’s theme is sexual desire in its various forms, its frustrations, its gratification, and so on. As I was reading it I thought I was entertaining myself with a fluffy, silly story about a man and a woman eager to get married so that they could get it on. But when I finished it and looked back I realized that for a fluffy, silly book it managed to sneak in quite a bit of interesting reflection about its theme while it was at it. Anyway, check out Chapter V, which is pretty damn funny:
Continue Reading »

Howls of outrage (4)

2008 03 23
Recently read

Brecht. Galileo

Brecht explores the moral difficulties in Gallileo’s decision to recant. Not bad.

Paulos, John Allen. Innumeracy

A fun little book that provides a healthy dose of motivation to the non-mathematical to get their (our!) act together. Paulos provides lots of examples of fuzzy thinking that follow from a neglect of basic mathematics. At times Paulos seems to cast his net a bit more broadly than mathematics even, commenting on various fallacies in informal reasoning. But that’s ok – those mistakes matter too.

Frank. Falling Behind: How Rising Inequality Harms the Middle Class

Entertaining and reasonably well-written. Frank charts the rise of inequality in American society since WWII, and then explains why he thinks that inequality is so harmful. Some goods are absolute goods. These we care about regardless of how much other people have. Others are positional goods. These we value very differently depending on context, most importantly how others around us are doing with respect to that good. Frank argues that many more goods are positional than one might first think, and then ties this insight to his observations about rising inequality. The result is a decent critique of a lot of mainstream assumptions about inequality in American society, and more broadly of the social policies that have produced it.

Two quibbles. First, it’s ok to dumb down a bit for a popular book, but Frank’s remarks about evolutionary psychology were pretty silly at times. I’d have to read Frank’s other work on the subject to know whether I would find a more careful statement of his views silly. But anyway, I don’t really think Frank needed to introduce claims about evolutionary psychology in the first place. His motivation for doing so, if I understood it correctly, was just to point out that the psychological tendencies he’s attributing to us are fairly stubbornly entrenched. But a) you don’t need to point to evolutionary considerations to do that; and b) you shouldn’t point to evolutionary considerations to do that (just for starters, innateness and malleability are completely distinct issues).

Second quibble: Frank talks throughout about the middle class. He even put the middle class in the subtitle of his book. But the book really seems to be about how just about everyone gets screwed by rising inequality, even very well-off people. So perhaps the subtitle to his book ought to have been “How Rising Inequality Harms Us All.”

Tufte, Edward R.The Visual Display of Quantitative Information

Superb. Tufte wrote the book in the last seventies and early eighties; it changed the way many people think about how to display quantitative information in a clear, engaging and helpful way. Tufte’s book is part polemic against a dumbing down of statistical charts on the grounds that no one finds them interesting, and part analysis of what considerations go into getting it right. Good stuff.

Nada (0)

2006 10 07

Posted by in: Math

Steve Laniel writes up two little mathematical proofs that even math-dunces like myself can follow. They’re both very pleasing, the first so much so that I actually laughed with pleasure as I read it.

Nada (0)

2004 08 07

Posted by in: Math

Neat write-up on the graph theory of the US and Canada at Isabel’s math blog, asking questions like: how many way can you split up the 6 New England states? Of possible interest to those trying to split up the US into red and blue.

My parents live in Maine, so I’m well aware that you have to pass through New Hampshire to get there. #%@*&! tolls and highway liquor stores.

Nada (0)

2004 05 15
41st Mersenne prime found?

Posted by in: Math

MathWorld is reporting that a computer participating in GIMPS, the Great Internet Mersenne Prime search, claims to have found the 41st Mersenne prime this morning.

This would be the largest known prime number, with roughly 7 million digits. There’s a $100,000 prize for the discovery of the first prime with 10 million digits, which GIMPS is likely to claim within a couple years.

The discovery of the 41st Mersenne prime would also mean the discovery of the 41st perfect number.

Update: After finding the message boards where the discovery is being discussed, it looks quite possible that there has been a new Mersenne prime found, but that it might be smaller that a previously known Mersenne prime.

Nada (0)

2004 04 08
Computers will eat your brain

Posted by in: Baseball, Math

Nearly every mathematician I know uses computers in their work in one form or another – at the very least they’re an easy way to do calculation that would otherwise take up too much time and energy. But an article like this seems to miss the point entirely.

There are two issues here: the first is that using a computer to check cases or run lengthy computations doesn’t affect the worthiness of the math; you’re not going to get shunted off to a computational journal for that. The second is that a proof that is actually found by a computer is not that interesting from a math standpoint. It’s nice to know a theorem is true, but mathematicians have been proving theorems based on unknown facts for thousands of years.

Stephen Goldman, possibly the best baseball writer of the last 25 years not from Kansas, wrote about similar issues in the baseball community. Computers in math, and the new (mainly statistical) methodologies used in baseball are just tools, and correspondingly can be used for good or evil. But dismissing them out of hand is foolish.

Nada (0)