The ten worst predictions in history: learning from past mistakes

  Ugo Bardi experiments with new predictive methods.

This post was inspired mainly by the shock I had with the various failed attempts to predict the outcome of the Covid-19 epidemic. It was truly a sobering experience: bad predictions, clueless politicians, arrogant scientists, idiotic journalists, and more. It made me doubt of the usefulness of models in general. I think we are doing several (too many) things wrong with the way we use models and (sometimes) we trust them. I'll be discussing more on this subject in future posts, for the time being, here is a list of failed predictions that I think can teach us something.

1. Coronavirus Deaths. In 2020, the model developed in large part by Neil Ferguson at the Imperial College in London was the main element that led the British government to engage in a strict "lockdown" policy to avoid the hundreds of thousands (perhaps millions) of deaths that the model predicted as a result of the COVID-19 disease. Most European States followed suit. It is still early to evaluate how the real world followed the model but, if we look at the result proposed in the "Report n. 9", we see that the model was clearly overly pessimistic. The authors of the model defended their work saying that their prediction of doom was just one of several scenarios, which is correct, but weak as a defense. In the future, we'll be able to say if Europeans truly wrecked their economies for nothing but, for the time being, the coronavirus experience can be seen as a sobering experience on the limits of the models as predictive tools.

2. The Population Bomb. In 1968, Paul Erlich and his wife Anne wrote a book titled "The Population Bomb." To say that it was catastrophistic is a little of an understatement. It is known to have contained the sentence  In the 1970s hundreds of millions of people will starve to death in spite of any crash programs embarked upon now. As we all know, that didn't happen. Instead, the 1970s ushered an era of apparent abundance in the food supply that led to a rapid increase in the human population. Yet, The Erlichs' prediction may not have been so bad, at least in qualitative terms. In reality, the current population is so large that if a new famine comes, it will cause the death of not just "hundreds of millions of people" but, likely, billions of them. You might argue that the present situation is much more dangerous in terms of risks of famines than it was half a century ago, when the world's food supply was not so critically linked to the supply of fossil fuels as it is today. This story illustrates once again how difficult it is to use models for quantitative predictions, even though they may be qualitatively good. 

3. Julian Simon's 7 billion years of prosperity. In 1994, Julian Simon wrote, "We now have in our hands in our libraries, really the technology to feed, clothe, and supply energy to an ever-growing population for the next 7 billion years... " (Myers, N., and Julian Simon. (1994): Scarcity or Abundance, New York: Norton.). In reality, we are reaching the limits of our capabilities of doing that just now. We may conclude that Simon was off of some 7 billion years. It was a ridiculous idea from the beginning but, at least, it shows to us that there is no limit to how wrong you can be with your predictions!

4. Too much oil. In 1999, the respectable (so to say) "Economist" magazine published a leader titled "Drowning in Oil" where it was argued that the prices of oil were going to become so low that they would damage the economy, literally "drowning it in oil." The authors went on proposing that the oil price could soon go down to 10 dollars per barrel and perhaps even to 5 dollars per barrel. Needless to say, the trend was exactly the opposite, with oil prices starting a sustained rally that took them all the way up to some 150 dollars per barrel in 2008. Curiously, oil prices did plunge, briefly, to "negative" values in 2020, but if this is what the Economist authors were thinking, they were 20 years off with their prediction! This story is just one of the many disasters that ensued when someone tried to predict the vagaries of prices, even in a not so remote future. Prices are so strictly linked to human whim and perception that it seems to be simply impossible to predict them.

5. The Great Horse Manure Crisis. It is said that, in 1894, The Times newspaper predicted… “In 50 years, every street in London will be buried under nine feet of horse manure.” This prediction was obviously wrong, as reported and it is often used to poke fun at those silly catastrophists. But, in reality, there is no trace in the archives of "The Times" that it was ever uttered, it is a piece of fake news.  This story illustrates that you can be wrong even with a non-existing prediction!

6. The Wargame that lost the war for Japan.  It is known that, before the start of WW2, the Japanese admiralty engaged in a series of simulation games destined to determine the possible outcome of a war against the United States. It is reported that when one of these simulations was leading to a crushing defeat for the Japanese forces, the umpires (or perhaps Admiral Yamamoto himself) ordered two Japanese carriers "refloated," so that the game ended with a Japanese victory. The story of the refloated carriers is probably a legend, at least in part, but it is true that the Japanese wouldn't have engaged in the war against the US hadn't they had some kind of evidence that they could have won. It shows how dangerous it is to trust models for systems that have a binary outcome.

7. The Peace Dividend. In the 1990s, with the Soviet Union defeated and gone, some people thought that there was no more justification for the gigantic and cumbersome military apparatus that the Western Empire (officially) maintained as a defense against Communism. So, they thought it could be dismantled and the money saved distributed to the people. That was called the "peace dividend." Alas, that idea now looks as remote as the fall of the Roman Empire. Soon, there came the 9/11 attacks and the military apparatus was beefed up even more. Those who had proposed the idea of a peace dividend never realized that nobody ever made money on peace. And, as you see, it is rarely a good idea to think that people will behave in ways that make them lose money.

8. Nibiru. The story of the planet Niburu that should impact on Earth appeared for the first time in 1995, diffused by Nancy Lieder. The first predictions said the cataclysm should have been in 2003, but the best-known prediction indicated 2012 as the fatal year, something that gained a certain notoriety and a number of followers. Of course, nothing happened in 2012, so the date of the impact has been progressively moved onward, but the whole story seems to have lost interest -- as it should have. At least, this story tells us that wrong predictions may be a lot of fun and cause little damage. 

9. The Solow-Swan model. In the 1950s, Robert Solow proposed a model for the US economy based on a simple formula that generated exponential growth. The model came to be known as the "Solow-Swan" model.  There was no term in the model that accounted for the possibility of a decline or a collapse so that it was generally interpreted by economists as implying that there were no limits to growth and that the economy could be growing forever. Today, most of the choices made by politicians and decision-makers all over the world seem to be based on this model, trying, for instance, to "restart the economy" after the SARS-Cov2 epidemic by means of stimulating more consumption. That, of course, is going to make things worse if the crisis is caused by a dearth of natural resources, as it is the case. This model shows us how you can do tremendous damage to humankind even by means of just a one-line equation.

10. The second coming (or the parusia). If you read the texts written by the early Christians, Paulus, and others, it is rather clear that they were expecting the "second coming" of Jesus Christ during their lives. That, of course, didn't happen and Christians have been waiting for this dramatic event for a long time, and they are still waiting for it, sometimes under different names such as the "rapture." We cannot exclude that one of these days we'll hear a booming voice coming from the sky telling us that "The Day of Judgement has come" -- but so far that didn't happen. So, this prediction turns out to have been wrong of at least 2000 years (so far). But Christians are still waiting for the second coming and so we can conclude that predictions with no well-defined time frame can never be disproved!

The View From Les Houches: What Are Models For?

Sandra Bouneau, researcher and lecturer at the university of Paris-Sud, shows her model at the School of Physics in Les Houches, France, in March 2018. As you can see from the image, her model is complex and detailed. It is one of the several models presented at the school which attempt to describe the trajectory of the transition.

Overall, all the models based on physics (including Bouneau's one, as far as I understood it) arrived to similar conclusions, confirming the calculations that myself, Denes Csala, and Sgouris Sgouridis published in 2016. In practice, the transition is possible, but it won't happen all by itself. The economic system needs to be pushed in the right direction, in such a way that it will be able to provide the necessary investments.

The problem is that the system is not being pushed hard enough. Some parts of it, including the US governments, are pushing in the wrong direction, dreaming of an impossible "energy dominance" (and even if it were possible, what good would it be for America?).

At the bottom of the whole problem, it is the fact that policy-makers don't believe in models, although they may declare the opposite. There have been many models developed during the past century or so which would have created a different world if the powers that be had acted on the advice provided - first and foremost "The Limits to Growth" of 1972. But that model was not only disbelieved but positively demonized.

In the end, All models are made to search for trajectories which avoid collapse, so ignoring models ensures collapse. And that's what we are doing!

Where is the proof that CO2 warms the Earth?

A persistent element of the climate debate is the claim that "there is no proof" that CO2 and other greenhouse gases warm the atmosphere. This has generated a number of amateurish demonstrations of how the greenhouse effect works  A good example of how NOT to carry out such a demonstration is shown above. Nobody told this poor kid how to perform a scientific measurement. If he had switched his jars he would have discovered that the effects he observes are only the result of one of the two jars being closer than the other to the light source. And they didn't even explain to him how the greenhouse effect actually works: you cannot see any warming in this set-up unless you place a light adsorber (e.g. a piece of black cloth) in the jar. I  discuss this and other disasters in a post of mine (in Italian), see also here.

Imagine that someone asks you to prove that the Moon orbits around the Earth because it is pulled by the force of gravity. Your first reaction would be to say something like "huh?" But - assuming you are in a good mood - you might try to explain how Newton's law of universal gravitation works and how it can be used to describe the moon's trajectory.

"So," your opponent could say, "is it correct to say that nobody ever measured the gravity pull of the Earth at the Earth-Moon distance?"

"Aw... No. What for? It would be terribly expensive. And useless, too."

"Then, you have no proof that the Moon goes around the Earth because of gravity. So, can you prove that the Moon is not being pushed by invisible angels, instead?"

That would end the civil conversation, but there is some logic in this kind of questions. You can't reproduce the Moon's motion in a lab, here on the Earth. Instead, to prove that Newton's law is valid you create a model based on the law and then use it to describe the movement of moons, planets, and stars. The model works, hence the law behind it is correct. No need of measuring the Earth's gravity pull at the Earth-Moon distance (and no need of invisible angels pushing).

Now, let's translate all this to a question often asked in the climate debate. What proof do we have that greenhouse gases, and CO2 in particular, warm the Earth's atmosphere? As a question, it is similar to the one about the Moon orbiting the Earth, in the sense that we can't reproduce the properties of a whole planetary atmosphere in the lab.

To answer the question, we can start from laboratory experiments showing that CO2 absorbs infrared radiation - they were done already by Tyndall in the 19th century. We don't need amateurs today to remake those experiments, doing the job poorly. Then, we create models that describe the Earth's atmosphere and we use them to fine-tune the parameters of the warming effect. The models follow reasonably well the warming trends of the past so that we confirm that CO2 warms the Earth.

In addition, we even have direct experiments showing that, as the atmosphere warms, more infrared radiation is radiated to space, while at the same time less infrared radiation escapes at the wavelengths where CO2 absorbs radiation. It is what we expect a greenhouse gas should do.

Case closed, then? Not really. The Moon-Earth system is relatively simple and, unless you really want to believe in invisible angels, there is no doubt that the Moon is kept in its orbit by gravity, and by gravity only. Instead, the Earth's ecosystem system is a tangle of subsystems interacting with each other in ways that - in many cases - we have troubles in quantifying exactly. Newton could never have described the Earth's atmosphere by means of a single universal equation. So, how to disentangle the various contributions to the Earth's temperature: albedo, atmospheric particulate, vegetation cover, and more?

This is the task of models, coupled with various kinds of measurements. For what we can say nowadays, the rising CO2 concentration is the main factor in the observed warming, but that's subjected to changes and refinements as we have new data. Among other examples, a very recent paper appeared in Nature quantified the effects of the vegetation cover. This is mainly a local effect and changes little to the overall picture, but it shows how dangerous it can be to say that "the science is settled." Of course, that doesn't mean jumping to the conclusion that angels are warming the Earth with their breath. And it doesn't mean to engage in silly games of data torturing. It just means recognizing the complexity of the problem.

Then, there is an even deeper problem: complex systems such as the Earth's ecosystem are just that: complex. They may react to perturbations in a strongly non-linear way, amplifying or dampening the perturbation. In a complex system, it is always difficult to say what causes what. You can say that in a complex system there are no causes and no effects, only forcings and feedbacks. So, when people say that the increase in CO2 concentration is not a cause but an effect of global warming, are they wrong? Yes, they are basically wrong, but it is also true that increasing concentrations of greenhouse gases may be both a cause and an effect of global warming. Complex systems are dominated by feedbacks. And that means they can amplify the effect of a small forcing, turning it into a disaster. This is, by the way, the origin of the "Seneca Effect"

In the end, it all means that we have to recognize that the Earth's climate could react to perturbations in ways we can't even imagine. And more than seven billion humans on the planet have been a huge perturbation, no matter how we want to see the relative importance of what we have been doing. We ignore that at our risk.

The world as a canvas: Vincent Van Gogh's models of the world

An intriguing post by Ilaria Perissi on her blog "boundaries" where she examines in depth the relation of painting and modeling - the latter in the modern sense of using mathematical tools to describe the behavior of complex systems. She goes in depth into describing how Vincent Van Gogh and how his paintings can be seen as models of the world. An excerpt from her post is below, but do read the whole thing; it is fascinating!

The image of the world around us, which we carry in our head, is just a model. Nobody in his head imagines all the world, government or country. He has only selected concepts, and relationships between them, and uses those to represent the real system – Jay Forrester (1971).

Could these sentences represent also a painting process? Is a painter a sort of modeler?  Following the rational of previous words, most of the painters are interested in reporting an image of the world and the paintings are just models of that images, including both the material world as well as the rapresentation of feelings and situations; they are not models of the whole world, but of a set of selected concepts and relationships used to represent a real material system, which could be a landscape, a still life, a portrait, a situation or event, as wars, battles, a sunset or a ‘starry night’, ‘potatoes eaters’  and representation of feelings as in the painting 'Sorrowing Old Man' or 'Two lovers'.

 

 Rread the whole post at Ilaria Perissi's "Boundaries"