Apollo's Arrow

by David Orrell

From seers to scientists, mystics to meteorologists, there have always been peoplewho claim to know what will happen in the future. The Oracle at Delphi, Pythagoras, Newton and the stock analyst on a business report have all endeavoured to look forward in time. But even with recent technological advances and the help of computers and satellites, are we any better at predicting the future now than we were in the distant past? How can scientists claim to foresee future climate events when even three-day forecasts prove a serious challenge?

In Apollo’s Arrow, David Orrell looks at the history of prognostication to show how scientists (and charlatans) have tried to forecast the future. He then breaks down the mathematics of what really goes into apredictive model. Orrell has created a compelling, elegantly written history of our future that addresses some of the most important issues of our time.

Genres: Science, Nonfiction, Futurism

464 pages, Paperback

First published January 7, 2006

About the author

David Orrell, Ph.D. is a scientist and author of popular science books. He studied mathematics at the University of Alberta, and obtained his Ph.D. from Oxford University on the prediction of nonlinear systems.

His work in mathematical modeling and complex systems research has led him to diverse areas such as weather forecasting, particle accelerator design, economics, and cancer biology. He has authored or coauthored research papers for journals including Journal of the Atmospheric Sciences, Nature Genetics, the International Journal of Bifurcations and Chaos, and Physica D.

He is the author of Economyths and The Other Side of the Coin: The Emerging Vision of Economics and Our Place in The World about new economic theories; and The Future of Everything: The Science of Prediction about prediction in weather, genetics, and economics, which was a national bestseller and finalist for the 2007 Canadian Science Writers' Association book award. Foresight called it "An engaging, as well as deeply insightful, discussion on the difficult task of prediction ... it can change the way you view forecasting."

David has been a guest on radio shows including Coast to Coast AM, NPR, and BBC, and his work has been featured in print media such as New Scientist and the Financial Times. He has spoken at many conferences and events including the Art Center Global Dialogues on Disruptive Thinking. He currently lives in Oxford, UK, where he runs a mathematical consultancy Systems Forecasting.

Awards
Finalist: Canadian Science Writers' Association book award (2007)
Finalist: National Business Book Award (2011)

Reviews

[Stephie Williams · Rating 4 out of 5]

As the subtitle says, this is a book about the science of prediction. Very little of the book (the last chapter) is about specific predictions of the future. The book is divided into three parts. Part one talks about predicting in the past, a major focus of which is in mechanics. This includes Ptolmey’s model, the Copernican model, Galileo and Newton’s mechanics. Part two covers the science of prediction in the present. It focuses on the weather, biology (mainly genetics and its influence on health), and economics. David Orrell’s major claim in this part of the book is that prediction goes astray not because of chaos theory (sensitive dependence on initial conditions—the famous butterfly effect), but on model error because no model is exact enough to capture all the necessary elements. Part three then looks at the future of predicting with a dismal appraisal. But, he is not claiming that all prediction is worthless. In the final chapter he gives two sets of prediction. One in which we do nothing about carbon emissions and one where we do. There are three appendixes, which explain some of the mathematics (simplified) used in prediction models.

Here are some of the notes and comments I made on the text:

[page 9] The whole thing of “an insect flapping its wings [the butterfly effect]” and affecting a storm at a great distance seems pretty far-fetched to me. Orrell tends to agree with me, but his view comes from his work on model error [also see comment for page 145].

[page 70] He and his editor made a big mistake: “[Galileo] also discovered four moons around Saturn.” They were Jupiter’s moons. Galileo did discover Saturn’s rings, not its moons;

[page 95] “When Napoleon asked Laplace why God did not play a role in his calculations for the solar system, he is said to have replied, ‘I have no need of that hypothesis.’” This is an excellent summation of methodological naturalism, which most practicing scientists adhere to, at least in their scientific work. Some people do consider god a mathematician, but most of these people I feel are usually speaking figuratively.

[page 105] Speaking on non-local phenomena in quantum mechanics, he claims that “Einstein . . . was religious . . .” Einstein was not religious in the sense of practicing a religion. He might have had some nonstandard belief in god, or some sense of spirituality, but that is a far cry from being religious.

[page 109] He seems to defend free will by reference to unpredictability, which he is saying all our prediction models of complex systems exhibit. I claim unpredictability does not eliminate cause and effect. The only known area of the universe that is known to be noncausal is in the quantum sphere, but from there on up it is cause and effect. If cause and effect were not occurring, the world would be a pretty strange place as the quantum realm appears to be. He says that autonomy comes from this unpredictability. Autonomy should only be considered as being violated when someone does not determine his or hers actions, but is force to or prevented from doing something.

[page 114] “It is important that models not be confused with this [complex systems] far richer reality.” A good example of the basic principle of “the word is not the thing” or “the map is not the territory.”

[page 124] “[Aristotle’s] own view [of meteorology] was that thunder, lightening, and hurricanes were all caused by ‘windy exhalations,’ perhaps from quarreling philosophers.” Orrell shows this sense of humor in places throughout the book.

[page 145] In response to the butterfly effect as a perceived source of prediction error, not model error, he states: “But is the weather really so delicate and finely poised a system that an insect can stir up a hurricane—or knock it off track—with a beat of its tiny wings?” My sentiments actually.

As far as the second half of the book, I found nothing worthy of comment, except I admire his willingness to see metaphor as an integral part of science, and I say as it is with life.

I thought the book to be a fair appraisal of the state of the science of prediction, and what can be expected from it in the future. I am not sure how other scientists working on prediction models would rebut his arguments, that the main problem with these models is model error (for instance, not having enough parameters), not sensitive dependence on initial conditions (ala chaos theory); although, I do not think he believes that chaos theory does not play any role at all. Orrell is a good explainer and easily understood by someone conversant enough with science. The book was enjoyable to read, and I did learn some new stuff, not just stuff I already new explained in a different way, that I often run into
in reading popular science books.

If you are looking for what is going to happen in the future, you maybe disappointed. However, if you would like to know or know more about models used in prediction science, especially those involved with the weather, health, or economics, the book should be of interest to you. I will offer one word of caution. If you are looking for actual mathematics (except for the appendixes) and not just description, I feel you would be as disappointed as those that wish for predictions for what is going to occur in the future.

[David · Rating 2 out of 5]

A lot of words to reach an underwhelming conclusion.

[Gumble's Yard - Golden Reviewer · Rating 2 out of 5]

The book considers three main areas:

The introduction to the book is an interesting history of forecasting - starting with ancient history (Delphi, Greek attempts to understand astronomy, astrology) then Renaissance developments (Kepler, Galileo, Newton, Descartes) then through Darwin, Malthus and Freud and into quantum theory and chaos and more importantly complexity). This is probably the strongest part of the book. A key premise of the book is that many models are not just chaotic; they are extremely sensitive to parameters. Further some systems are simply not predictable; even if a model can be built there is no way to predict outcomes other than by running the models. Another strong view is that biologists (Dawkins is explicitly mentioned) still have an over strong belief in determinism and in a mechanical world model of the type discarded by physicists a long time ago.

The next section considers three areas in turn and explores three areas of forecasting - weather/climate, genetic (the most complex to follow but most original section and most thought provoking - showing the complexities of the genetic process like those of the weather and economy), economic. The conclusions are similar - unlike say planetary motion, the systems are complex, model and parameterisation error dwarfs the butterfly effect, models can easily be used to fit past data but are unreliable going forwards and also cannot be predicted without running them (often in not far off real time).

The third section brings some of the thoughts together including discussing the psychology of the climate change debate. This is the weakest section including a rather over long Gaia section and incongruously making some future long term predictions.

A mixed "curate's egg" of a book - definitely with some interesting ideas and worth retaining but not one to recommend to others.

[Lavender · Rating 4 out of 5]

A knowledgable book that covers the study of prediction since the ancient times to now. Discusses the error of current models and provides counterarguments for common arguments, etc. I can definitely say I learned (a lot) from the book, though I couldn't hold on word-for-word and towards the end only read the summaries, bullet points, asides/case studies, etc

[Ford Miller · Rating 4 out of 5]

read this as a wall street journal copy so not sure if this is the same book.

that said, at the time it was great insight to future tech. is outdated now.

[Shourov Hasonov Pre · Rating 3 out of 5]

🐶🐺🐱🦈🐠🐧🐜

[Dianne · Rating 4 out of 5]

The title of this book is a little over-reaching, since it mostly just discusses the prediction of weather, the economy, and an individual's health over the course of his lifetime. That said, I thought this book was so interesting. It starts with a historical overview of the tools the ancients used to predict events in their world. For example, how the model of the earth as the center of the universe was so entrenched in their mind-set, and how that prevented them from moving forward even when observations were inconsistent with the model. When Isaac Newton invented calculus and was actually able to predict the movement of the planets with real accuracy, it caused scientists in all disciplines to believe that all science should be tools of this kind of clean and precise prediction. But, physics is an anomaly in that way. Systems like the weather, organisms and the economy can't be predicted from given starting conditions, because, as the system moves forward, new conditions emerge as the result of the system moving forward, and that can't be captured by mathematical equations.

From this book I learned (1) a precipitation forecast is only accurate for 24 hours, and (2) invest in index funds and forget the rest of the stock market. The author was able to explain concepts in such a way that I could understand even though I am a product of the public school system and my math skills are shameful. This book wasn't a quick, easy read. There were several parts where I had to read the same paragraph about 40 times before what he was saying came together, and I finally grasped where he was coming from. This book made me want to learn more about chaos and emergence theory.

[Mike Smith · Rating 4 out of 5]

This book is about why it's so hard to predict certain things. In particular, it focuses on models and methods for predicting weather, health and wealth. The book begins with a historical overview of major scientific advances, from the Greeks to the 20th century, that improved our ability to predict things (like the motions of the planets) and shaped our expectations for our ability to control the future. Then it looks at methods for making short-term predictions and long-term predictions of its three main topics. Weather covers tomorrow's weather as well as climate change. Health looks at individual genetic profiles and global pandemics. Wealth looks at predicting stock and commodity prices over a few weeks or months and longer-term effects like recessions and bubbles.

In all these cases, the author argues that the systems are so inherently complex that they cannot be modeled. That is, they cannot be reduced to equations that can be solved to predict future states. They are uncomputable. The best we can do is create models that simulate these systems. But our models have so many parameters that even minor adjustments to a few of them can cause huge changes in the predictions the models make. All of which means that the future is inherently unpredicatble. Faster computers or better models won't help.

The author concludes with some thoughts about the usefulness of models and how to prepare for an unpredictable future. He believes that, even if we can't be sure about major events like climate change, it seems foolhardy not to prepare for them, just in case. The cost of preparing for catastrophe and being wrong is less than the cost of NOT preparing for catastrophe and being wrong.

[E · Rating 5 out of 5]

Thought-provoking study of the limits of forecasting

This book is a fascinating, very readable look at the accuracy of modern forecasting. David Orrell begins with an overview of the history of telling the future, including humanity’s inherent need to try to decipher what tomorrow holds. He covers the current state of forecasting in fascinating detail, dwelling on weather and climate, economics and medicine. He points out the shortcomings in experts’ current ability – or lack thereof – to predict the future accurately in any of these realms. Finally, he discusses the future of forecasting, and makes a case for using the limited models available to become better prepared for future events, particularly climate-related ones, even those that are impossible to forecast with precision. The fiscal and commercial relevance of his advice is startlingly clear in the light of recent natural disasters. getAbstract highly recommends this well-structured overview of forecasting and the author’s cautionary message that the planet’s health will govern much of what lies ahead.

[Bernd · Rating 4 out of 5]

This book is a fascinating, very readable look at the accuracy of modern forecasting. David Orrell begins with an overview of the history of telling the future, including humanity’s inherent need to try to decipher what tomorrow holds. He covers the current state of forecasting in fascinating detail, dwelling on weather and climate, economics and medicine. He points out the shortcomings in experts’ current ability – or lack thereof – to predict the future accurately in any of these realms. Finally, he discusses the future of forecasting, and makes a case for using the limited models available to become better prepared for future events, particularly climate-related ones, even those that are impossible to forecast with precision. The fiscal and commercial relevance of his advice is startlingly clear in light of recent natural disasters

[Steve · Rating 4 out of 5]

This is definitely something that you should read before you start putting a lot of trust into any long-term forecasting involving weather, genetic outcomes, and economics. As you would expect, lots of variables and assumption errors in models can - and do - produce wildly different results.

The only thing that I didn't care for was all the quotes that he used. Many of them were unnecessary for him to get his point across.

[Charm]

Literally just started this book. It was the only thing I asked for for Christmas (nerd I know).

I received a forward of the abstract from a Sr. Director from work who handles our analytics and promotions team. It explores the history & evolution of forecasting as well as the intriguiing limits of prediction, so far anyway. Looking forward to the rest of the read.

[John Boettcher · Rating 3 out of 5]

I thought the book was going to have more substance to it than what it did. I guess my expectations were a little higher than they should have been, but I was wooed in by the title and guess I was expecting some profound revelations revealed in the book. I guess in the long run, we're all dead! lol

[Tina · Rating 5 out of 5]

Excellent Book! Combines the knowledge of the ancients with the science of the present and the future.

[Robert · Rating 4 out of 5]

The bottom line - there are and always will be limits in making predictions about future outcomes. This is applicable to weather, human events, and medicine.

[Alex Pang · Rating 4 out of 5]

A vastly underrated book about the limits of what we can know about the future.

[Raghavendra Vadali · Rating 4 out of 5]

A thoughtful book that gives an its own unique insight into current questions of the day. There is quite a bit of historical data and lots of models at the end of the book.

[Cassandra · Rating 3 out of 5]

really enjoy the topic and I got some good history, but it was a lot of history. it didn't feel very pointed.