Statistical Modeling Quotes

Essentially, all models are wrong, but some are useful

Indeed, statistical modelling based on these results even suggests that one of the effects of the plague was a substantial improvement in life expectancy.

It is critical to recognize the limitations of LLMs from a consumer perspective. LLMs only possess statistical knowledge about word patterns, not true comprehension of ideas, facts, or emotions. Their fluency can create an illusion of human-like understanding, but rigorous testing reveals brittleness. Just because a LLM can generate coherent text about medicine or law doesn’t mean it grasps those professional domains. It does not. Responsible evaluation is essential to avoid overestimating capabilities.

Models are the mothers of invention.

Machine learning takes many different forms and goes by many different names: pattern recognition, statistical modeling, data mining, knowledge discovery, predictive analytics, data science, adaptive systems, self-organizing systems, and more.

You end up with a machine which knows that by its mildest estimate it must have terrible enemies all around and within it, but it can't find them. It therefore deduces that they are well-concealed and expert, likely professional agitators and terrorists. Thus, more stringent and probing methods are called for. Those who transgress in the slightest, or of whom even small suspicions are harboured, must be treated as terrible foes. A lot of rather ordinary people will get repeatedly investigated with increasing severity until the Government Machine either finds enemies or someone very high up indeed personally turns the tide... And these people under the microscope are in fact just taking up space in the machine's numerical model. In short, innocent people are treated as hellish fiends of ingenuity and bile because there's a gap in the numbers.

In high school algebra, someone had already worked out the formulas. The teacher knew them or could find them in the teacher’s manual for the textbook. Imagine a word problem where nobody knows how to turn it into a formula, where some of the information is redundant and should not be used, where crucial information is often missing, and where there is no similar example worked out earlier in the textbook. This is what happens when one tries to apply statistical models to real-life problems.

In this country, it is not the highest virtue, nor the heroic act, that achieves fame, but the uncommon nature of the least significant destiny. There is plenty for everyone, then, since the more conformist the system as a whole becomes, the more millions of individuals there are who are set apart by some tiny peculiarity. The slightest vibration in a statistical model, the tiniest whim of a computer are enough to bathe some piece of abnormal behaviour, however banal, in a fleeting glow of fame.

Most statistical models are built on the notion that there are independent variables and dependent variables, inputs and outputs, and they can be kept pretty much separate from one another.39 When it comes to the economy, they are all lumped together in one hot mess.

One reason why many statistical models are incomplete is that they do not specify the sources of randomness generating variability among agents, i.e., they do not specify why otherwise observationally identical people make different choices and have different outcomes given the same choice.

All models are wrong, but some are useful.’ CHAPTER 6 Algorithms, Analytics and Prediction

You no longer watch TV, it is TV that watches you (live),” or again: “You are no longer listening to Don’t Panic, it is Don’t Panic that is listening to you”—a switch from the panoptic mechanism of surveillance (Discipline and Punish [Surveiller et punir]) to a system of deterrence, in which the distinction between the passive and the active is abolished. There is no longer any imperative of submission to the model, or to the gaze “YOU are the model!” “YOU are the majority!” Such is the watershed of a hyperreal sociality, in which the real is confused with the model, as in the statistical operation, or with the medium. …Such is the last stage of the social relation, ours, which is no longer one of persuasion (the classical age of propaganda, of ideology, of publicity, etc.) but one of deterrence: “YOU are information, you are the social, you are the event, you are involved, you have the word, etc.” An about-face through which it becomes impossible to locate one instance of the model, of power, of the gaze, of the medium itself, because you are always already on the other side.

This book is an essay in what is derogatorily called "literary economics," as opposed to mathematical economics, econometrics, or (embracing them both) the "new economic history." A man does what he can, and in the more elegant - one is tempted to say "fancier" - techniques I am, as one who received his formation in the 1930s, untutored. A colleague has offered to provide a mathematical model to decorate the work. It might be useful to some readers, but not to me. Catastrophe mathematics, dealing with such events as falling off a height, is a new branch of the discipline, I am told, which has yet to demonstrate its rigor or usefulness. I had better wait. Econometricians among my friends tell me that rare events such as panics cannot be dealt with by the normal techniques of regression, but have to be introduced exogenously as "dummy variables." The real choice open to me was whether to follow relatively simple statistical procedures, with an abundance of charts and tables, or not. In the event, I decided against it. For those who yearn for numbers, standard series on bank reserves, foreign trade, commodity prices, money supply, security prices, rate of interest, and the like are fairly readily available in the historical statistics.

Be wary, though, of the way news media use the word “significant,” because to statisticians it doesn’t mean “noteworthy.” In statistics, the word “significant” means that the results passed mathematical tests such as t-tests, chi-square tests, regression, and principal components analysis (there are hundreds). Statistical significance tests quantify how easily pure chance can explain the results. With a very large number of observations, even small differences that are trivial in magnitude can be beyond what our models of change and randomness can explain. These tests don’t know what’s noteworthy and what’s not—that’s a human judgment.

Thus, they do not need to understand the statistical and mathematical models in depth. However, marketers need to understand the fundamental ideas behind a predictive model so that they can guide the technical teams to select data to use and which patterns to find.

The point that apocalyptic makes is not only that people who wear crowns and who claim to foster justice by the sword are not as strong as they think--true as that is: we still sing, 'O where are Kings and Empires now of old that went and came?' It is that people who bear crosses are working with the grain of the universe. One does not come to that belief by reducing social processes to mechanical and statistical models, nor by winning some of one's battles for the control of one's own corner of the fallen world. One comes to it by sharing the life of those who sing about the Resurrection of the slain Lamb.

Avoid succumbing to the gambler’s fallacy or the base rate fallacy. Anecdotal evidence and correlations you see in data are good hypothesis generators, but correlation does not imply causation—you still need to rely on well-designed experiments to draw strong conclusions. Look for tried-and-true experimental designs, such as randomized controlled experiments or A/B testing, that show statistical significance. The normal distribution is particularly useful in experimental analysis due to the central limit theorem. Recall that in a normal distribution, about 68 percent of values fall within one standard deviation, and 95 percent within two. Any isolated experiment can result in a false positive or a false negative and can also be biased by myriad factors, most commonly selection bias, response bias, and survivorship bias. Replication increases confidence in results, so start by looking for a systematic review and/or meta-analysis when researching an area.

There are two primary strains in the Conservative Party: grocers, and grandees. … By ‘grandees’ and ‘grocers’, I am not referring to social class or any of that; nor do I refer to the Worshipful Company of Grocers, all cloves and camels. I refer rather to two fundamental positions within the Conservative Party, regardless of one’s antecedents. … A grandee Conservative sees the country as a village: a village of which he and his party, when in government, act the Squire. As the Squire, the grandee moves jovially amongst his tenants in their tied cottages, dispensing largesse and reproof…. There are two problems with this model. The first is that HMG is not the Squire and the subjects of the Crown are not the smocked tenantry of the government of the day. The second is that these principles – or instincts, as one can hardly call them principles – however different they may be to the fiercely held maxims of Labour old and new, lead in the end to the same statist solutions as those the Left proposes, and to accepting and ‘managing’ statism when a Conservative government succeeds a Labour one. It is the grocers who will always and rightly attempt to roll back the State and its reach in favour of liberty.

I don’t mean to compare myself to a couple of artists I unreservedly admire—Miles Davis and Ray Charles—but I would like to think that some of the people who liked my book responded to it in a way similar to the way they respond when Miles and Ray are blowing. These artists, in their very different ways, sing a kind of universal blues, they speak of something far beyond their charts, graphs, statistics, they are telling us something about what it is like to be alive. It is not self-pity which one hears in them, but compassion. And perhaps this is the place for me to say that I really do not, at the very bottom of my own mind, compare myself to other writers. I think I really helplessly model myself on jazz musicians and try to write the way they sound. I am not an intellectual, not in the dreary sense that word is used today, and do not want to be: I am aiming at what Henry James called “perception at the pitch of passion.

There is no freedom or justice in exchanging the female role for the male role. There is, no doubt about it, equality. There is no freedom or justice in using male language, the language of your oppressor, to describe sexuality. There is no freedom or justice or even common sense in developing a male sexual sensibility—a sexual sensibility which is aggressive, competitive, objectifying, quantity oriented. There is only equality. To believe that freedom or justice for women, or for any individual woman, can be found in mimicry of male sexuality is to delude oneself and to contribute to the oppression of one’s sisters. Many of us would like to think that in the last four years, or ten years, we have reversed, or at least impeded, those habits and customs of the thousands of years which went before—the habits and customs of male dominance. There is no fact or figure to bear that out. You may feel better, or you may not, but statistics show that women are poorer than ever, that women are raped more and murdered more. I want to suggest to you that a commitment to sexual equality with males, that is, to uniform character as of motion or surface, is a commitment to becoming the rich instead of the poor, the rapist instead of the raped, the murderer instead of the murdered. I want to ask you to make a different commitment—a commitment to the abolition of poverty, rape, and murder; that is, a commitment to ending the system of oppression called patriarchy; to ending the male sexual model itself.

System 1 is generally very good at what it does: its models of familiar situations are accurate, its short-term predictions are usually accurate as well, and its initial reactions to challenges are swift and generally appropriate. System 1 has biases, however, systematic errors that it is prone to make in specified circumstances. As we shall see, it sometimes answers easier questions than the one it was asked, and it has little understanding of logic and statistics. One further limitation of System 1 is that it cannot be turned off.

It is a positive sign that a growing number of social movements are recognizing that indigenous self-determination must become the foundation for all our broader social justice mobilizing. Indigenous peoples are the most impacted by the pillage of lands, experience disproportionate poverty and homelessness, and overrepresented in statistics of missing an murdered women, and are the primary targets of repressive policing and prosecutions in the criminal injustice system. Rather than being treated as a single issue within a laundry list of demands, indigenous self-determination is increasingly understood as intertwined with struggles against racism, poverty, police violence, war and occupation, violence against women, and environmental justice. ... We have to be cautious to avoid replicating the state's assimilationist model of liberal pluralism, whereby indigenous identities are forced to fit within our existing groups and narratives. ... Indigenous struggle cannot simply be accommodated within other struggles; it demands solidarity on its own terms. Original blog post: Unsettling America: Decolonization in Theory and Practice. Quoted In: Decolonize Together: Moving beyond a Politics of Solidarity toward a Practice of Decolonization. Taking Sides.

Price mostly meanders around recent price until a big shift in opinion occurs, causing price to jump up or down. This is crudely modeled by quants using something called a jump-diffusion process model. Again, what does this have to do with an asset’s true intrinsic value? Not much. Fortunately, the value-focused investor doesn’t have to worry about these statistical methods and jargon. Stochastic calculus, information theory, GARCH variants, statistics, or time-series analysis is interesting if you’re into it, but for the value investor, it is mostly noise and not worth pursuing. The value investor needs to accept that often price can be wrong for long periods and occasionally offers interesting discounts to value.

Marriage is inefficient!” she proclaims. “The whole construct is a model of wasted resources. The wife often stays home to care for the children, or even a single child, abandoning the career she worked so hard for, losing years of creative output. Beyond the wasting of talent, think of the physical waste. For every home, there are so many redundancies. How many toasters do you think there are in the world?” “I have no idea.” “Seriously, just guess.” “Ten million?” I say impatiently. “More than two hundred million! And how often do you think the average household uses its toaster?” Once again, she doesn’t wait for my answer. “Just 2.6 hours per year. Two hundred million toasters are sitting unused, statistically speaking, more than 99.97 percent of their active lives.

VaR has been called “potentially catastrophic,” “a fraud,” and many other things not fit for a family book about statistics like this one. In particular, the model has been blamed for the onset and severity of the financial crisis. The primary critique of VaR is that the underlying risks associated with financial markets are not as predictable as a coin flip or even a blind taste test between two beers. The false precision embedded in the models created a false sense of security. The VaR was like a faulty speedometer, which is arguably worse than no speedometer at all. If you place too much faith in the broken speedometer, you will be oblivious to other signs that your speed is unsafe. In contrast, if there is no speedometer at all, you have no choice but to look around for clues as to how fast you are really going.

In Bohr’s model of the atom, electrons could change their orbits (or, more precisely, their stable standing wave patterns) only by certain quantum leaps. De Broglie’s thesis helped explain this by conceiving of electrons not just as particles but also as waves. Those waves are strung out over the circular path around the nucleus. This works only if the circle accommodates a whole number—such as 2 or 3 or 4—of the particle’s wavelengths; it won’t neatly fit in the prescribed circle if there’s a fraction of a wavelength left over. De Broglie made three typed copies of his thesis and sent one to his adviser, Paul Langevin, who was Einstein’s friend (and Madame Curie’s). Langevin, somewhat baffled, asked for another copy to send along to Einstein, who praised the work effusively. It had, Einstein said, “lifted a corner of the great veil.” As de Broglie proudly noted, “This made Langevin accept my work.”47 Einstein made his own contribution when he received in June of that year a paper in English from a young physicist from India named Satyendra Nath Bose. It derived Planck’s blackbody radiation law by treating radiation as if it were a cloud of gas and then applying a statistical method of analyzing it. But there was a twist: Bose said that any two photons that had the same energy state were absolutely indistinguishable, in theory as well as fact, and should not be treated separately in the statistical calculations.

How I Got That Name Marilyn Chin an essay on assimilation I am Marilyn Mei Ling Chin Oh, how I love the resoluteness of that first person singular followed by that stalwart indicative of “be," without the uncertain i-n-g of “becoming.” Of course, the name had been changed somewhere between Angel Island and the sea, when my father the paperson in the late 1950s obsessed with a bombshell blond transliterated “Mei Ling” to “Marilyn.” And nobody dared question his initial impulse—for we all know lust drove men to greatness, not goodness, not decency. And there I was, a wayward pink baby, named after some tragic white woman swollen with gin and Nembutal. My mother couldn’t pronounce the “r.” She dubbed me “Numba one female offshoot” for brevity: henceforth, she will live and die in sublime ignorance, flanked by loving children and the “kitchen deity.” While my father dithers, a tomcat in Hong Kong trash— a gambler, a petty thug, who bought a chain of chopsuey joints in Piss River, Oregon, with bootlegged Gucci cash. Nobody dared question his integrity given his nice, devout daughters and his bright, industrious sons as if filial piety were the standard by which all earthly men are measured. * Oh, how trustworthy our daughters, how thrifty our sons! How we’ve managed to fool the experts in education, statistic and demography— We’re not very creative but not adverse to rote-learning. Indeed, they can use us. But the “Model Minority” is a tease. We know you are watching now, so we refuse to give you any! Oh, bamboo shoots, bamboo shoots! The further west we go, we’ll hit east; the deeper down we dig, we’ll find China. History has turned its stomach on a black polluted beach— where life doesn’t hinge on that red, red wheelbarrow, but whether or not our new lover in the final episode of “Santa Barbara” will lean over a scented candle and call us a “bitch.” Oh God, where have we gone wrong? We have no inner resources! * Then, one redolent spring morning the Great Patriarch Chin peered down from his kiosk in heaven and saw that his descendants were ugly. One had a squarish head and a nose without a bridge Another’s profile—long and knobbed as a gourd. A third, the sad, brutish one may never, never marry. And I, his least favorite— “not quite boiled, not quite cooked," a plump pomfret simmering in my juices— too listless to fight for my people’s destiny. “To kill without resistance is not slaughter” says the proverb. So, I wait for imminent death. The fact that this death is also metaphorical is testament to my lethargy. * So here lies Marilyn Mei Ling Chin, married once, twice to so-and-so, a Lee and a Wong, granddaughter of Jack “the patriarch” and the brooding Suilin Fong, daughter of the virtuous Yuet Kuen Wong and G.G. Chin the infamous, sister of a dozen, cousin of a million, survived by everbody and forgotten by all. She was neither black nor white, neither cherished nor vanquished, just another squatter in her own bamboo grove minding her poetry— when one day heaven was unmerciful, and a chasm opened where she stood. Like the jowls of a mighty white whale, or the jaws of a metaphysical Godzilla, it swallowed her whole. She did not flinch nor writhe, nor fret about the afterlife, but stayed! Solid as wood, happily a little gnawed, tattered, mesmerized by all that was lavished upon her and all that was taken away!

Though Hoover conceded that some might deem him a “fanatic,” he reacted with fury to any violations of the rules. In the spring of 1925, when White was still based in Houston, Hoover expressed outrage to him that several agents in the San Francisco field office were drinking liquor. He immediately fired these agents and ordered White—who, unlike his brother Doc and many of the other Cowboys, wasn’t much of a drinker—to inform all of his personnel that they would meet a similar fate if caught using intoxicants. He told White, “I believe that when a man becomes a part of the forces of this Bureau he must so conduct himself as to remove the slightest possibility of causing criticism or attack upon the Bureau.” The new policies, which were collected into a thick manual, the bible of Hoover’s bureau, went beyond codes of conduct. They dictated how agents gathered and processed information. In the past, agents had filed reports by phone or telegram, or by briefing a superior in person. As a result, critical information, including entire case files, was often lost. Before joining the Justice Department, Hoover had been a clerk at the Library of Congress—“ I’m sure he would be the Chief Librarian if he’d stayed with us,” a co-worker said—and Hoover had mastered how to classify reams of data using its Dewey decimal–like system. Hoover adopted a similar model, with its classifications and numbered subdivisions, to organize the bureau’s Central Files and General Indices. (Hoover’s “Personal File,” which included information that could be used to blackmail politicians, would be stored separately, in his secretary’s office.) Agents were now expected to standardize the way they filed their case reports, on single sheets of paper. This cut down not only on paperwork—another statistical measurement of efficiency—but also on the time it took for a prosecutor to assess whether a case should be pursued.

Was this luck, or was it more than that? Proving skill is difficult in venture investing because, as we have seen, it hinges on subjective judgment calls rather than objective or quantifiable metrics. If a distressed-debt hedge fund hires analysts and lawyers to scrutinize a bankrupt firm, it can learn precisely which bond is backed by which piece of collateral, and it can foresee how the bankruptcy judge is likely to rule; its profits are not lucky. Likewise, if an algorithmic hedge fund hires astrophysicists to look for patterns in markets, it may discover statistical signals that are reliably profitable. But when Perkins backed Tandem and Genentech, or when Valentine backed Atari, they could not muster the same certainty. They were investing in human founders with human combinations of brilliance and weakness. They were dealing with products and manufacturing processes that were untested and complex; they faced competitors whose behaviors could not be forecast; they were investing over long horizons. In consequence, quantifiable risks were multiplied by unquantifiable uncertainties; there were known unknowns and unknown unknowns; the bracing unpredictability of life could not be masked by neat financial models. Of course, in this environment, luck played its part. Kleiner Perkins lost money on six of the fourteen investments in its first fund. Its methods were not as fail-safe as Tandem’s computers. But Perkins and Valentine were not merely lucky. Just as Arthur Rock embraced methods and attitudes that put him ahead of ARD and the Small Business Investment Companies in the 1960s, so the leading figures of the 1970s had an edge over their competitors. Perkins and Valentine had been managers at leading Valley companies; they knew how to be hands-on; and their contributions to the success of their portfolio companies were obvious. It was Perkins who brought in the early consultants to eliminate the white-hot risks at Tandem, and Perkins who pressed Swanson to contract Genentech’s research out to existing laboratories. Similarly, it was Valentine who drove Atari to focus on Home Pong and to ally itself with Sears, and Valentine who arranged for Warner Communications to buy the company. Early risk elimination plus stage-by-stage financing worked wonders for all three companies. Skeptical observers have sometimes asked whether venture capitalists create innovation or whether they merely show up for it. In the case of Don Valentine and Tom Perkins, there was not much passive showing up. By force of character and intellect, they stamped their will on their portfolio companies.