William M. Briggs

Statistician to the Stars!

Page 149 of 729

George Gilder’s Information Theory Of Money


This is the second part of a review of George Gilder’s monograph The 21st Century Case for Gold: A New Information Theory of Money. The first part, about the return of a gold standard for money, is at The Stream. You need to read that part before continuing. We return to SAMT in two weeks.

Gilder doesn’t have a discussion about how a gold standard can be implemented. He steers mostly clear of politics and sticks with theory. His monograph’s subtitle is “A New Information Theory of Money”, but perhaps it’s better classed as an intriguing new notion. In it’s current form, the theory is a thin skeleton that badly needs flesh. And a few corrections.

What’s right is his search for a definition of what money is. Time, he says. Time is “irreversible, inexorably scarce, impossible to hoard or steal, distributed with remorseless equality to rich and poor alike. As an index of time, gold imparts the accurate price signals needed for sustained economic growth and expanded opportunity.”

This makes sense when it is considered early humans “had every natural resource we have”, and the only difference between them and us is knowledge, or information purchased by time. While this is true, or true enough, it’s not clear what function of time equals wealth. Whatever this function is, it almost certainly can’t be linear; there must be some threshold effect; wealth must “kick in” only after some bare minimum of human needs are met. And that minimum, given our biology, must be food. It’s only the well-fed who ponder wealth.

Gilder is keen to map time to wealth and wealth to knowledge, or information. It is accepted by everybody that information is a key to the economy. Ideas drive the economy forward—and backward. Information is a value-neutral term. The government, for instance, is continuously presenting the economy with new “information” in the form of regulations, laws, mandates, and so forth.

Gilder’s enthusiasm for this fecund line of thought leads him astray. He says things like Claude Shannon, a prime inventor of information theory, “resolved that all information is basically surprise. Unless messages are unexpected, they do not convey new information.” And “Surprisal—what Shannon called ‘entropy’—is both a measure of freedom and criterion of creativity.”

This is muddled. Shannon did not define information as surprise, a claim Gilder makes several times. Shannon was interested in how information is communicated, not in information per se. Information transmitted via some channel is, by Shannon’s rule, “surprising” to the extent it is improbable. But to be improbable means the fundament of the information must already be known, and that is because all probability is conditional on what is known or assumed.

Consider transmitting English letters (in sentences, say). The letter z is rarer than e, so it is more “surprising” to see a z. The probability of seeing a z, and all the other letters, has a bearing on how to best encode the information so that it arrives at its destination in the most efficient way possible. And we know the alphabet in advance. You cannot transmit an unknown letter, without also including its definition.

To some extent, information transfer is how the economy works, which is plain. But surprise isn’t everything. Information that is known can be more valuable than new information which is surprising. After all, it is already-known information that lets factories produce their goods. Information theory does not, and can not, account for the creation of information, which is by definition infinitely surprising. Much more work needs to be done in figuring how to best encourage the creation of new, good, and true information.

Gilder also says the “second law of thermodynamics ordains that entropy as disorder always increases and cannot be reversed” and that this, somehow, ties to the “time-based entropy of [gold] extraction.” Now it’s true that entropy for closed systems, like the universe as a whole, increases through time, but it is false that everywhere entropy increases. Every new good idea produces new information, as Gilder rightly emphasizes, acts which decrease entropy, or rather, increase order. Entropy in the information-theory sense increases when unpredictability increases. The economy is not a closed, or zero-sum system.

The difficulties multiply because we can’t say we’d unconditionally like to have an predictable economy. A Soviet five-year plan was highly predictable, but produced an economy only New York Times readers would enjoy. So we’re not chasing entropy, high or low, but good information, knowledge. Knowledge is information which is true.

How Many Lousy Predictions Until Error Is Admitted?

Update Breitbart kindly asked to reprint this, and just did.

How long does it take for an expert who has, year upon year after year, made predictions of unrealized doom, an expert hailed heeded and hearkened to by the whole world, to admit error? Answer: forever. He never will, and neither will most of his admirers.

This is important to understand—let this sink into your bones—because this sad but true fact about the human condition tells us how long global warming will be with us. Answer: forever. It will last until the last of it proponents die. Which, given improvements in medical practice, will be at least several more decades.

No one said this pithier than Max Planck: Science advances one funeral at a time. The increased and increasing longevity of our species thus partly accounts for the deceleration in scientific knowledge which many have remarked.

Ladies and gentlemen, I bring you the New York Times. The New York Times! No publication is more progressive or as detached from reality, no group of writers more convinced of their secularly divine destiny. Yet, somehow, they brought themselves to publish the piece “The Unrealized Horrors of Population Explosion“. Watch the video at the top. I’ll wait here.

Back? Fascinating, ain’t it?

Paul Ehrlich has been as wrong, wronger even, than Jim Hansen or Al Gore, yet no number of failed predictions has so much as put a crease in the man’s unwavering support of himself. And wasn’t he as a younger man convincing! The sheer authority in his voice, the utter believability! How many times did he say he was on The Tonight Show?

Dr. Ehrlich was so sure of himself that he warned in 1970 that “sometime in the next 15 years, the end will come.” By “the end,” he meant “an utter breakdown of the capacity of the planet to support humanity.”

As far as I can tell, though I haven’t done a systematic count, not one of Ehrlich’s predictions has come true. Does it matter? Not to him, and not to the many, many grant- and award-awarding bodies who, to this very day, fete the man. The Weekly Standard:

In 1990—the same year he lost his bet with Julian Simon—Ehrlich was awarded a million dollar MacArthur “genius” grant and was simultaneously feted across the Atlantic with Sweden’s Crafoord Prize, which was worth just about half a million. In 1993 the Heinz Family Foundation bestowed on him its first Heinz Award. This little trinket came with $100,000 in cash and the most delusional praise possible, claiming that Ehrlich’s “perspective, uncommon among scientists, has made [him and his wife] the target of often harsh criticism—criticism they accept with grace as the price of their forthrightness.” Which is a peculiar way of explaining that Ehrlich was completely wrong and that he responded to all such evidence with ad hominem attacks. Five years later, in 1998, he was awarded the Tyler Prize, which comes with $200,000. The money train kept on rolling.

Now don’t make the same mistake this magazine did and call Ehrlich “insanely dangerous”. He is not. He is a harmless old man, who was wrong about, as far as I can tell, absolutely everything. It is a world class blunder to focus on Ehrlich and not on the political scum who “leveraged” Ehrlich’s preposterous predictions to push their anti-human agenda. It was the one-worlders who deserve all the discredit, not poor Ehrlich.

Forced sterilizations! They happened. Forced population “planning”. It happened. Not because Ehrlich said they must, and not even because this deluded and delusional man desired them, but because the powerful progressives who used Ehrlich as a front-man wanted them. They were content to let Ehrlich and others create panic and then to use that panic to their benefit. Headlines shouted “Science Says…!”, “Scientists Agree…!”, “Science Science Science!”

Sound familiar?

Increased population was going to bring only ills. Every bad thing that happened was because of increased population. There were no possible benefits to increased population. Why, increased population was going to cause so much pollution that the globe was going to be plunged into another ice age! The only solution that any right-thinking person could conceive of was to cede more power to governments. Only governments, and preferably world government, could get us out of the mess.

Sound familiar?

It is long past the point where anybody, scientist or not, can seriously consider climate models worthy of attention. They are wrong as Ehrlich was, or wronger. But it doesn’t matter. Their inaccuracy is not the root malady.

Update Bill McKibben is not today’s Ehrlich, at the least because he is not nearly as intelligent, but this essay of his shows certain parallels.

Heartland Conference, Statistics Classes, Statistics Books, & Statistical Cups


Heartland Conference

I’ll be there, speaking Friday on “The Need To Believe In ‘The Solution’ To Global Warming”. This is a paper which I’ll try to find a home for (2,000 words).

I’ve been told registrations have maxed out. So if you want to get autographs and pictures or locks of hair, you’ll have to do it in the lobby of the hotel. I think presentations will eventually be on-line. Stay tuned.

On-line statistics class: sort of

It’s that time of year. My class starts on the 15th. In preparation, on the blog I’ll be doing mostly epistemology from now until the class starts, after which I’ll switch to the class for two weeks. The “on-line” part means you’re welcome to follow along. I’ll be blogging about the Heartland conference, of course.

The level is entry. But it isn’t a standard statistics class by any stretch.

As usual, I won’t do much lecturing, but that which I do sticks close to the book. I’ll be using the same book, which is badly in need of updating. But it has the right price: download a free copy.

I prefer the fancy-sounding Socratic method. Meaning I want people to figure things out for themselves by asking questions. I give some canned examples, but these are never as effective as having people make their own. Because of that, I can’t see how a strict on-line class like mine would ever work, unless everybody participated by video camera (which I understand some systems now do).

Updated book?

Speaking of books, my new one is almost finished. It’s still in review at Cambridge.

This book is not like the first. It is meant for advanced audiences, the sort who might have read (say) Howson and Urbach’s Scientific Reasoning: The Bayesian Approach or ET Jaynes’s Probability Theory: The Logic of Science.

Well, maybe only the philosophical portions of Jaynes. My book has some (advanced) math, mostly about a different interpretation of exchangeability and the origin of parameters, but as little as possible. The problem is that probability isn’t, at base, a mathematical subject. Portions of it—only portions!—can be turned into formulas. Ad the danger is applying those formulas. That’s when interpretation arises and the Deadly Sin of Reification lurks.

Everything flows from this idea: probability, the completion of logic, is the study of relations between propositions; it is thus objective; and thus all probability, like all logic, is conditional on the exact premises specified. And from that simple truth, a complete theory of probability and statistics arises. This theory looks very unlike classical statistical practice, but it resembles, or actually is, the way physics used to be done (before politics and wishful thinking took even that field over).

Maybe next week, I’ll post the Table of Contents. Once it’s “done”, I’ll be sending copies of it to some colleagues (at this point, this is only a handful of people; I’ll let blog readers know when it’s ready for publication).

There is no global warming, politics, or cultural matters in the book. I do take on bad statistics practices, but I do not name names (that way everybody can think it was the other guy that made mistakes). It is pure (applied) epistemology.


The other exception to posting will be the Pope’s encyclical. It’s been said it’ll be about 1/3 longer than this last, which already ran over 200 pages. There is some serious work ahead of us.

Drink up!

Due to underwhelming demand I have created my own swag store! Be the first one your block, and very probably the first one on your own continent, to own the Official WMBriggs.com Thirst-Quenching COFFEE CUP OF THE STARS!

It works swell with brandy! It positively sings with bourbon! But it comes into its own with rum! And I have even heard that it can be used with coffee!

Notice all the exclamation cups! That’s how you know it’s a Good Deal!


Get the option with the black handle and rim. It looks best. But you can also get a plain white (cheaper), stein or travel mug.

I get 10% from every order (before PayPal takes their cut). I figure if I can sell about 31,000 of these each year, then I can make a good living. So don’t worry about breaking yours.

I’ll have a permanent button at the top of the site for these items soon. Oh, the cups are the same. I have to figure out why the store shows it twice. Don’t ask for t-shirts. T-shirts should never be worn except as undergarments.

Why Global Warming Models Are Hot And Bothered

Everybody remembers what happened when Lord Christopher Monckton, Willie Soon, David Legates and I published our (what is physics is called) toy model of the climate. Global warming increased! Caused by apoplexy, over-heating, fuming, and ranting by the usual crowd of dimwits.

It reached the point where one incompetent or malevolent—there is no third choice—United States Senator calling on skeptics like myself to be prosecuted by the RICO act (see also this). Because why? Because tobacco, or something. The man is an ass.

In the frenzy, the world’s mean IQ dropped a full ten points. Only one group of actual scientists tried to take on the paper. Mark Richardson, Zeke Hausfather, Dana Nuccitelli, Ken Rice, and John Abraham wrote the response “Misdiagnosis of Earth climate sensitivity based on energy balance model results” and published it in the same journal our original paper appeared. We were allowed a rebuttal, which is the usual process, and which was (again, as usual) supposed to be published simultaneously with Richardson et al.

Well, you know how these things go. The Science Bulletin—the journal took a lot of heat for publishing us—put up Richardson et al. but misplaced our rebuttal, which will now show later. (Or you can download it here, in draft form, now; but, shhhh, don’t tell anybody where you got it. It’s a, the Lord be merciful, Word doc.)

Anyway the publishing, um, mishap allowed The Guardian to pee its pants: “Research downplaying impending global warming is overturned: A new study finds Monckton et al. (2015) riddled with errors.”

The author, Nuccitelli, like many, fails to distinguish between model fit and model predictions. (To be fair, Nuccitelli likely did not write the hyperbolic headline.)

Sure, GCMs can fit past data, more or less well. Any model can be made to fit past data! But you don’t show that fit and then claim the model works, that would be foolish. Right!?

Oh, wait. You do. That’s what classical statistics is all about: showing how well some data can be squeezed into a model, much like how the bodies of certain women in June (after a long winter) are squeezed into last year’s bathing suits. The thing can be done, but it doesn’t mean the results are good.

Thousands upon thousands—oh, some damnably large number—of papers showing how data can be squeezed into models are published yearly. How many of these models would make good predictions? Ask it another way: how many models are good, since the only test of true goodness is how well models make predictions of data never seen before?

Not too many, that’s how many.

Particularly compare the picture Nuccitelli has in the Guardian with the one displayed in the tweet above (or at this link). One shows fit, the other forecast. See how easy it is to fool yourself? (We discuss Nuccitelli’s picture in our rebuttal.)

Tell the truth, I’m sick of this whole business. I’m thisclose to never saying another word about climatology. What a dismal science, filled with untrained civilians (which includes sociologists, economists, etc.), all of whom have stronger opinions on the subject than any scientist, cowardly scientists, many of whom know damn well what is happening but who keep their mouths shut, bottom-feeding immoral politicians, whose only concern is self-aggrandizement, activists, unhinged, every mother-loving one of them, reporters, the worst of the bunch, because each thinks himself a crusading genius. The damage done to thought by this preposterous situation is incalculable.

Over twenty years I’ve made maybe a couple of thousand dollars from this field, but I’ve lost much, much more (my retirement “plan” consists in investing in lottery tickets). The work I’ve done has given me only grief. Try finding a job once you’re labeled a “denier”. The ignoramus who thought that one up needs to be first in line for the series of blanket parties which are long overdue our nation’s intellectuals.

Us skeptics are supposed to be awash in oil money. The next nitwit that says that to me better have a good dental plan. For decades I’ve had my hand out to oil companies—I have no compunction taking their money—but never a cent have I received. Not from them, and not from any company affiliated with them.

The monks—the only group which has a chance of surviving—who write the history of this period will never stop laughing.

Update We heard from the reviewers why our rebuttal wasn’t printed. There are “errors” we have to fix. Here is a sample they noticed.

Page 2 Line 46: “Appropriate caveats about the limitations ” is undefined and vague language. The crux of [13] is that the limitations reduce the accuracy beyond applicability, while [6] maintains strong conclusions drawn from this model. “Appropriate” is a matter of opinion in this case, and should not be included without merited justification.

Page 2 Line 50-52: The claim that global temperature measurements are only accurate after 1979 is unsubstantiated. Please provide a citation.

Page 3 Line 11: “Perform satisfactorily” is vague. Please refrain from subjective language.

All niggling (are we still allowed to use this word?) details, mostly copy editing. Because of the back-and-forth, don’t look to see official version of the paper above for a least a month.

« Older posts Newer posts »

© 2017 William M. Briggs

Theme by Anders NorenUp ↑