Green forecasting

Kesten Green forecasting, that is. We had a fascinating presentation today at lunch by Dr Kesten Green, a world authority on forecasting technique. This is how his presentation was advertised within the School:

Title: Are global warming forecasts scientific? Evidence from a forecasting audit and a validation study

Abstract:
With J. S. Armstrong, I conducted a systematic analysis of the processes used by those who make alarming forecasts of dangerous manmade global warming. We found that the processes violated basic scientific procedures such as full disclosure of data and methods, assessment of reasonable alternative hypotheses, reporting of potential conflicts of interest, and ensuring that conclusions do not go beyond the findings. The processes also violate principles of scientific forecasting by using methods that have not been empirically validated and by making dramatic forecasts when there is great uncertainty about the situation.

With Armstrong and W. Soon, I investigated whether long-term prediction of global mean temperatures is possible by testing a naïve ‘no change’ benchmark model. We found that the mean absolute error of 108 50-year-ahead forecasts from 1900 to 2007, for example, was 0.24°C. We then conducted validation tests of the U.N. Intergovernmental Panel on Climate Change’s projection of cumulative warming of 0.03°C per year. We found that the IPCC forecast absolute errors were 6.8 times larger than the naïve benchmark errors for horizons 41 through 50, and 12.6 times larger for horizons 91 through 100.

If there is time, I will describe ongoing research on forecasting the persuasiveness and outcomes of the global warming alarmist movement.

It may also be worth looking at his background and personal history. If the global warming issue were about reasoned discussion and open debate, Dr Green would be amongst the first whose views would be called upon:

Dr Kesten Green, International Graduate School of Business and Ehrenberg-Bass Institute for Marketing Science, University of South Australia. Dr Green is an international recognized expert on forecasting. He is a Director of the International Institute of Forecasters and is co-owner and co-director of the Forecasting Principles public service Internet site (ForPrin.com), the leading online resource for scientific forecasting.

Dr Green has developed two new forecasting methods that have led to improvements in predicting the decisions people make in conflicts such as occur in business competition, supply chains, mergers and acquisitions, between customers and businesses, and in warfare.

Dr Green has also conducted research on forecasting for public policy making, including climate forecasting. He has established the PublicPolicyForecasting.com Internet pages to encourage a scientific approach, and to disseminate findings.

Dr Green’s research findings have been published in the International Journal of Forecasting, International Journal of Public Opinion Research, Interfaces, and the International Journal of Business. His research has been covered in the Australian Financial Review, the London Financial Times, the New Yorker, the Washington Times, and the Wall Street Journal.

He has advised the Alaska Department of Natural Resources, the U.S. Department of Defense (The Pentagon), the Defense Threat Reduction Agency, the National Security Agency (NSA), and more than 50 other business and government clients.

If you are interested in following his work, his website is found here. As an interesting aside, if you type the word “forecasting” into google, you get 39 million hits of which his is the third.

This entry was posted in Uncategorized. Bookmark the permalink.

19 Responses to Green forecasting

  1. rog

    [take a hint. You’re a thread wrecker. Sinc]

  2. Louis Hissink

    Steve,

    I continue to be frustrated with the ignorance of the scientific method by some here – it’s used to explain something that has been persistently observed during the past but without a satisfactory explanation in terms of extant knowledge.

    Science is always concerned with explaining “history”, because “history” can be subjected to testing.

    However positing something to happen in the future, with no previous basis in observation, is nothing other than wishful thinking. Applying the scientific method in this case might seem useful but because no one can replicate the experiment immediately, because it is in the future, it isn’t scientific.

    It’s technically sophisticated advocacy, or pseudoscience.

  3. Rafe Champion

    Everyone needs to understand that when you have a regression model with a lot of variables you can get any damn result you want depending on (a) the variables that you put in and (b) how you manipulate them, for example lagging (to pick up a time effect) or a log transformation (to scale down the numerical value of a variable). Also, in some systems a lot of that data are so lousy that you have to use inadequate substitutes (proxies).

    A sophisticated soil scientist gave out a warning on the abuse of models several decades ago.

    This is clearly a major problem with climate models and also with econometrics and especially the study of third world economies. Remember, Garbage In -> Garbage Out!

    In the unlikely event that anyone wants to know more about regression models there is a very simple introduction at page 10 of this Regression Model of NSW Hospital Costs which dates from 1978 when I got into and out of econometrics.

  4. PSC

    I’ve read this paper. It also claims that the following alarming forecasts were incorrect:

    – Lead in petrol causes brain damage
    – Asbestos causes lung damage
    – Industrial production causes acid rain
    – Radon in homes causes lung cancer
    – Mercury in fish affects human development

  5. JC

    At the top of my head the radon scare was nonsense, PSC.

  6. PSC

    Why not type “radon lung cancer pubmed” into Google?

  7. JC

    PSC

    There was a radon scare going around in the US in the 90’s. Every house was endangered and you had to have it cleansed by a radon expert in order to remove the contaminant.

    The studies show that homes can at times accumulate radon equal to that exposed to by miners and may then pose a risk.

    No shit sherlock. And how often does that happen?

    In other words it was one big fucking huge scare about nothing.

  8. This is clearly a major problem with climate models and also with econometrics and especially the study of third world economies.

    It is much more widespread than that Rafe. I’ve just finished reading “The Alzheimer Myth”, written by a highly respected neurologist who argues that if you look closely at the data the concept of Alzheimers as this single entity disease is a gross simplification. Researchers have their hobby horses which typically require mountains of funding to investigate. So they promote their hypotheses and ignore data that doesn’t suit their ideas, let alone regarding dementia as a pathology that arises from the whole of the person, not just the brain. So all too often the data is tortured to give a desired result.

    What is now emerging in relation to dementia is the recognition that by the time symptoms present there is typically too much damage. The concept of “cure” is silly. Prevention is the key but that means people must change their lives and doctors are not in a position to make people change their lives. People do that. And thankfully there is sufficient research to provide many valuable strategies towards that end.

    This suggests that if we are looking for those who can help us better understand dementias we may be better off looking at the ideas of those whose careers do not depend on their narrow focus. Thus …

    Upton Sinclair: “It is difficult to get a man to understand something when his salary depends on his not understanding it.”

  9. PSC

    And how often does that happen?

    Often enough to make radon a small causing factor for lung cancer – small but still the second causative factor behind smoking.

    Given it’s about 1.3-1.5 million people worldwide who die of lung cancer each year, and estimates vary around the 3%ish mark for radon, that gives us around 50,000 deaths a year.

    So to answer your question, I’d guess it happens something in the order of 50,000 times a year.

  10. PSC

    Actually now I think a moment longer, I would guess the cancer is caused by multiple exposures. So let me revise that to: A lot more than 50,000 times a year.

  11. Jim Rose

    Rafe,

    It was only at honours level was I exposed to truly dangerous essays such as:
    1. Leamer’s Let’s take the con out of econometrics, and
    2. David Hendry’s Econometrics: alchemy or science?

    My methodology of econometrics lecturer was fresh from Simon Fraser.

    At ANU at the master’s level, Hendry’s PC-give was al the rage.

    The background debate where applied economists were despairing for their futures because of data mining and publication bias was passed over.

  12. JC

    PSC

    No it happens to (55,000/6,500,000,00).

    Okay, granted it would only really happen to adults so it’s around lets say .0000016 of that.

    As I said it’s a beat up.

  13. JC

    Actually now I think a moment longer, I would guess the cancer is caused by multiple exposures. So let me revise that to: A lot more than 50,000 times a year.

    nonsense.

  14. JC

    Quodger thinks he’s now goign to get brownie points by pimping as a self appointed troll at the Trop against the CAT.

    I guess that chip on his shoulder about not finishing high school really hurts. (9th grade)

  15. PSC

    JC – 50,000 deaths a year is a beat up? About 60m people die each year. So we’re (ballpark) a bit under 1/1000 deaths.

  16. JC

    I think you’re looking at it the wrong way, PSC. You need to look at it from the number of people exposed to radon which is adults 3,250,000,000. and out of those 55,000 die as a result of radon exposure.

    I presume we all live in houses, yea?

  17. JC

    PSC

    How many people are exposed to radon? You’d agree we all are as we all live in houses.

    what’s the death rate of people living in homes exposed to radon.

    60 million people dying each year doesn’t mean squat to be honest. People dying of car accidents, plane crashes or sickness. It has no real relevance.

  18. PSC

    You’d agree we all are as we all live in houses.

    Nope.

    It depends on the rocks the house is built on.

Comments are closed.