Talk:J. Scott Armstrong

From Wikipedia, the free encyclopedia

This article is within the scope of WikiProject Biography. For more information, visit the project page.
Biography Portal
Start This article has been rated as Start-Class on the project's quality scale. [FAQ]
This article is supported by the Science and academia work group.

This page looks like its a WP:AUTO, and probably a copyvio of http://www.jscottarmstrong.com/ William M. Connolley 21:41, 19 June 2007 (UTC)

I, as the initial creator, of the page know Professor Armstrong - it was not written by him. All the information put onto the page was done with his permission, not violating copyrights to the best of my knowledge. Publishers have agreed to allow for his papers to be placed online in their original formats (usually not published versions). I encourage others to edit the page in the spirit of Wikipedia. Kxjtaz 22:29, 19 June 2007 (UTC)

Putting aside auto for the moment, there needs to be some evidence that you have his permission to copy stuff into wiki - just your assertion here isn't good enough William M. Connolley 08:52, 20 June 2007 (UTC)
Would you mind letting me know how to do so? Thanks! Kxjtaz 13:08, 20 June 2007 (UTC)
Try Wikipedia:Copyrights perhaps William M. Connolley 13:25, 20 June 2007 (UTC)
Thanks, will make a note on the page 130.91.25.68 14:01, 20 June 2007 (UTC)

Please read WP:AUTO. Writing an autobiographical article is poor practice. Writing one on your behalf of your boss is scarcely better.Craticula 08:51, 21 June 2007 (UTC)

The awards look non notable. Are the publications notable ? --BozMo talk 06:23, 25 June 2007 (UTC)
How is it an auto when he said "it was not written by him [Professor Armstrong]"? mike4ty4 07:25, 29 July 2007 (UTC)

May we should take note of [1] William M. Connolley 21:08, 25 June 2007 (UTC)

One assumes you meen the bits about "I just received word I’m in charge of publicity & generating some buzz". —Preceding unsigned comment added by 71.214.249.249 (talk • contribs)
Yes William M. Connolley 15:33, 26 June 2007 (UTC)
I'm not sure how generating buzz for Professor Armstrong's challenge, written on my blog, relates to this Wikipedia page? I created the Wikipedia article months before I even heard about the paper or the challenge. As I have stated previously, I am only a Wikipedia newcomer who was not familiar with all of the rules, and in light of that would like to see the article go through the community editing process. Kxjtaz 19:59, 26 June 2007 (UTC)
May the 11th, to be precise, is when you started, which isn't quite "months". It looks suspicious. But we tend to WP:AGF if you're lucky. OTOH Prof A is aware of this page and uses it as a ref William M. Connolley 20:29, 26 June 2007 (UTC)

Contents

[edit] Do not delete links to Armstrong's papers

The notability requirement WP:Notability (people) is a guideline to determine is a person should have a web page or not. Armstrong definitely meets the criteria for notability. However, some have deleted a link to one of his peer reviewed papers claiming the paper is not notable because it is only cited one time. I cannot agree with this deletion. If the article was overly long at this point and had to be trimmed back, I might agree. Length is not a problem and readers should have access to this paper, one of the first in the area of scientific forecasting. RonCram 23:57, 27 June 2007 (UTC)

There are too many papers there spammed in by his student. NN papers from way back with 1 cite aren't worth mentioning William M. Connolley 08:32, 28 June 2007 (UTC)
William, the paper discusses how physical scientists fail to consider information that is contrary to their theories. It appears you are deleting this paper only because you dislike the content or find it embarrassing. Since your deletion appears to driven by your POV, please do not delete it again. RonCram 15:44, 2 July 2007 (UTC)
Also, annotated bibliographies are acceptable. In this case, I think it is helpful. RonCram 15:49, 2 July 2007 (UTC)
Ron its an encyclopedia not a reading list, nor a CV or bibliography. We report what is notable about people/things etc. The current list is severely bloated - and should be cut down to around 5-10 of the most cited works. --Kim D. Petersen 16:20, 2 July 2007 (UTC)
As far as I can the the Technology Review is not a peer-reviewed journal, but more of a popular technology magazine. Including this with real publications is at least misleading. --Stephan Schulz 16:29, 2 July 2007 (UTC)

[edit] Sources

For someone with JSA's claims, online refs to him that aren't directly sourced back to him are very thin on the ground: I've managed to find one only. Thus I am concerned that this article isn't well sourced.

OTOH, having looked up the papers, I'm quite happy that JSA *has* written a pile of papers many of them with quite high cite counts.

William M. Connolley 18:57, 2 July 2007 (UTC)

I've removed , the most frequently cited book on forecasting methods {International Journal of Forecasting (see International Journal of Forecasting, 1, 1985, p.1)} since a cite from 1985 cannot possibly establish it as the most-cited *now* William M. Connolley 16:04, 16 July 2007 (UTC)

[edit] External links

What's with all the links to papers and primary sources? This was so written by one of his students. (At least, I hope it wasn't Dr. Armstrong himself. I would think if he did it, having a PhD and everything, it would be less obvious.) Anyway, this article needs some serious clean-up. Rocket000 17:20, 15 October 2007 (UTC)

[edit] Fair use rationale for Image:PrinciplesOfForecasting.jpg

Image:PrinciplesOfForecasting.jpg is being used on this article. I notice the image page specifies that the image is being used under fair use but there is no explanation or rationale as to why its use in this Wikipedia article constitutes fair use. In addition to the boilerplate fair use template, you must also write out on the image description page a specific explanation or rationale for why using this image in each article is consistent with fair use.

Please go to the image description page and edit it to include a fair use rationale. Using one of the templates at Wikipedia:Fair use rationale guideline is an easy way to insure that your image is in compliance with Wikipedia policy, but remember that you must complete the template. Do not simply insert a blank template on an image page.

If there is other fair use media, consider checking that you have specified the fair use rationale on the other images used on this page. Note that any fair use images uploaded after 4 May, 2006, and lacking such an explanation will be deleted one week after they have been uploaded, as described on criteria for speedy deletion. If you have any questions please ask them at the Media copyright questions page. Thank you.

BetacommandBot 21:19, 24 October 2007 (UTC)

[edit] Fair use rationale for Image:LongRangeForecasting.jpg

Image:LongRangeForecasting.jpg is being used on this article. I notice the image page specifies that the image is being used under fair use but there is no explanation or rationale as to why its use in this Wikipedia article constitutes fair use. In addition to the boilerplate fair use template, you must also write out on the image description page a specific explanation or rationale for why using this image in each article is consistent with fair use.

Please go to the image description page and edit it to include a fair use rationale. Using one of the templates at Wikipedia:Fair use rationale guideline is an easy way to insure that your image is in compliance with Wikipedia policy, but remember that you must complete the template. Do not simply insert a blank template on an image page.

If there is other fair use media, consider checking that you have specified the fair use rationale on the other images used on this page. Note that any fair use images uploaded after 4 May, 2006, and lacking such an explanation will be deleted one week after they have been uploaded, as described on criteria for speedy deletion. If you have any questions please ask them at the Media copyright questions page. Thank you.

BetacommandBot 10:41, 7 November 2007 (UTC)

[edit] POV wording

The article currently reads: "Most recently, Armstrong appears to be unhappy that those interested in global warming, including the IPCC, completely ignore the literature on forecasting principles that Armstrong writes." This is clearly POV wording. Armstrong is not unhappy the IPCC has ignored his own writings as much as he is upset the IPCC has ignored the entire body of scientific literature on the subject. While Armstrong is a leader in this field, anyone can see the Journal of Forecasting has been publishing since 1982. Dozens of issues with close to 100 scientific papers have been published by dozens of different authors in that journal alone. The International Institute of Forecasting publishes three different journals: The International Journal of Forecasting, Foresight and The Oracle. Again, dozens and dozens of researchers and practicioners of the science have contributed to these journals. The IPCC has ignored all of this literature and is making "projections" on the basis of computer models which are untrustworthy. RonCram (talk) 18:47, 16 April 2008 (UTC)

I see Kim has reverted my edit. Not only is Kim's preferred version POV, it is clearly wrong. Armstrong does not complain in his paper about his writings being ignored so much as he complains that ALL scientific papers on forecasting have been ignored:
We also examined the 535 references in Chapter 9. Of these, 17 had titles that suggested the article might be concerned at least in part with forecasting methods. When we inspected the 17 articles, we found that none of them referred to the scientific literature on forecasting methods.
It is difficult to understand how scientific forecasting could be conducted without reference to the research literature on how to make forecasts. One would expect to see empirical justification for the forecasting methods that were used. We concluded that climate forecasts are informed by the modelers’ experience and by their models—but that they are unaided by the application of forecasting principles. (page 1015) [2]
I have restored my edit along with this evidence. RonCram (talk) 20:34, 16 April 2008 (UTC)
Nah, A is self-promoting William M. Connolley (talk) 21:13, 16 April 2008 (UTC)
Ron. A simple question here: What is the scientific literature on forecasting? How does Armstrong define this? Why is this not forecasting literature?
And now a bit more complex: If you read the paper you find that Armstrong is referring singularly to his own research on forecasting as being authoritative. Its an "audit" against Armstrongs list of variables. --Kim D. Petersen (talk) 21:28, 16 April 2008 (UTC)
Kim, fair question. I do not know how Armstrong would reply to your question but I will take a shot at it. At first glance, the authors appear to be attempting to move toward a scientific forecast - which is commendable. Unfortunately, the authors show absolutely no acquaintance with the scientific literature on forecasting. 95 papers are cited but none published by any of the four leading scientific journals on forecasting. Does it not seem odd to you that authors wanting to move toward a scientific climate forecast would not be familiar with the literature? Look also at the "Acknowledgements" section. They thank a number of climatologists, but not a single expert in scientific forecasting. They appear to be completely unaware the field has been in existence for more than 25 years. For it to qualify as part of the scientific literature on forecasting, the authors would have to show some knowledge of the literature in their paper. RonCram (talk) 23:08, 16 April 2008 (UTC)
Again: Who defines the literature? Who defines the "four leading scientific journals on forecasting"? (hint: the ones that Armstrong is citing is primarily on economics). How do you know that someone is unaware if they do not cite specific papers? Does every paper on gravity cite Newton or Einstein? Why not? Are they unaware of the field? --Kim D. Petersen (talk) 23:25, 16 April 2008 (UTC)
Kim, I answered the last question only. Different people may define the scientific literature on forecasting somewhat differently. I would say the main body would be the four journals we have been discussing: the three published by the International Institute of Forecasting and Journal of Forecasting. Certainly other articles may be published in more general journals and still make a contribution to the literature, but the approach would have to be evidence-based. Climate modeling is not. Also, if an author in a general journal does not cite articles from these four journals... well how can they be expected to add to the scientific discussion or even be taken seriously by the experts in the field? RonCram (talk) 18:11, 18 April 2008 (UTC)
Ron, there is a huge "scientific literature" on forecasting in general, Armstrong writes specifically about one sub-set (business studies) which has only the most tenuous links to climate science. The idea that his work somehow defines or is central to the field is risible. There is a massive body of theoretical and practical work on Bayesian estimation and prediction, which is directly relevant to the climate science issue and is well-known to researchers in this area. FWIW, I work in this field (the climate science bit, not business studies).Jdannan (talk) 02:22, 17 April 2008 (UTC)
Jdannan, I am glad to hear you work in climate science. However, you are wrong when you write that "Armstrong writes specifically about one sub-set (business studies)..." Obviously your conclusion has more to do with your knowledge that Armstrong is a Professor of Marketing than it does with actual knowledge of Armstrong's writings. Armstrong has compiled a list of forecasting principles, some he has discovered and described himself and others have been described by others, that have selective applicability to all fields in which forecasts are made. Armstrong also conducts audits of forecasts made by others. One of the steps in his audit is to decide which principles are applicable to a particular field or forecast. He then determines if the forecasters followed all of the applicable principles. Forecasters can generally be divided into theorists and practicioners. Armstrong is both. Forecasters also join "Special Interest Groups" for different types of forecasts. Special Interest Groups exist for forecasts in Conflict & Terror, Crime, Health, Neural Networks, Politics, Public Policy, Rules Based and Software Estimation. Armstrong has written in most of these areas. He is obviously very involved in Rules Based, Public Policy and Politics. All of this information is available on the website www.forecastingprinciples.com and I think you would find the page very helpful. RonCram (talk) 18:11, 18 April 2008 (UTC)
Doesn't it tell you anything that none of these SIGs deal with physical sciences? And sorry, but I think your tone is somewhat patronizing, considering whom you are speaking to. --Stephan Schulz (talk) 19:38, 18 April 2008 (UTC)
Stephan, actually "Public Policy" touches on the physical sciences because physical science (like global warming and other environmental) issues must be assessed by policymakers. Also, (I am not certain about this but I believe) "Software Estimation" may be applied to physical sciences. And Stephan, you know me. I treat everyone the same. James is used to me being patronizing. I have posted on his blog. RonCram (talk) 05:48, 19 April 2008 (UTC)
Ron, there is not one single mention of the term "data assimilation" on Armstrong's site. This topic (both theory and practice, of which there is a lot) underpins the entire subject of model-based forecasting in the physical sciences. That includes, but it not limited to, the entirety of meteorology and weather prediction, which I hope you would accept is a large area of research (far larger than climate science, for example). Viewed in this context, his grandiose claim that his site "summarizes all useful knowledge about forecasting" is simply a joke. It's a classic case of someone sticking their nose somewhere where they don't even know how little they know. Your not knowing this isn't a crime, but persisting in pushing it on wikipedia simply discredits the resource.Jdannan (talk) 22:36, 18 April 2008 (UTC)
James, your concern over "data assimilation" shows you are wedded to computer models for forecasts. And yes, I understand computer models are used in weather forecasting .... and they are good for about a week. The problem, James, is that climate scientists assume they cannot learn anything from the scientific forecasters. It reminds me of Michael Mann thinking he could innovate statistical methods without checking with any statisticians. Computer models have their place and we can learn from them... mainly they us what we don't know about climate. But computer models have limited to zero predictive power. This is A&G's point and the point Orrin Pilkey makes in his book "Useless Arithmetic." Computer models inform the opinions of the experts but research shows that experts are no better at predicting the future than non-experts. I know the experts find this offensive, but it is well established in the literature. My "pushing" the A&G studies does not discredit Wikipedia, it makes interesting and desired information available to readers. I'm mainly hoping more climate scientists will read A&G's papers. They need to deal with the issues raised there. RonCram (talk) 05:48, 19 April 2008 (UTC)
I am not "concerned" over data assimilation, I merely pointed out that the complete absence of any material on this completely destroys any claim by Armstrong (and you on his behalf) to speak for "scientific forecasting" in general. As for using computer models, reasonable predictions of climate change can be made with what are essentially back-of-the-envelope calculations using known physical laws - the fact that much more complex and detailed models support with these calculations (to a good approximation) is supporting evidence but hardly fundamental to the whole field.Jdannan (talk) 08:03, 19 April 2008 (UTC)
James, Armstrong's website may not discuss data assimilation but that hardly means it is not discussed in the literature. Journal of Forecasting has been publishing since 1982. Read Armstrong's book and see what it says about the subject. Or, if you haven't time for that, email Armstrong yourself and ask him. I have much more confidence in back of the envelope calculations than I do the GCMs. I just have more confidence in Schwartz's envelop and in Chylek's envelope than in yours. Scientific forecasting is evidence-based. Armstrong has the evidence on his side. For years, no forecast verifications have been done on climate projections. Just recently Pielke Jr. did one and that spawned a number of others. All of them showed the GCMs had significantly overpredicted global warming. Now that the PDO has turned to its cool phase, the difference between projections and reality will grow even greater over then next few years. You need to ask yourself, what will it take for me to be convinced global warming will not be catastrophic? When that happens you can join the Light Side. RonCram (talk) 15:02, 19 April 2008 (UTC)
Ron, you should have saved yourself the embarrassment and simply searched for "data assimilation" in the journal. There is precisely one hit in more than 25 years of issues, and that is a review article which looks at a range of other journals and also discusses neglected areas (I don't know the exact context of the reference). I don't see what else I can do to convince you that their use of the term "forecasting" is simply not general and comprehensive. I have literally never heard of the Lancaster Centre for Forecasting (home of the author of that one article) which claims to be "the pre-eminent focus of forecasting research in Europe". I bet there are two order of magnitude more people working at the UK Met Office, the main goal of which is actually doing forecasting, and none of them will have heard of it either. All you demonstrate by your faith in Schwartz and Chylek is that you don't understand any of the science.Jdannan (talk) 22:16, 19 April 2008 (UTC)

James, you are assuming I embarass easily. I don't. It appears to me that you checked only one of the four journals we have been discussing here. I could be wrong, but you have still not proven your point to me. Given, for a moment, that you are capable of proving your point in the future, what have you gained? Have you proven that climate scientists cannot learn anything from the scientific forecasters? Not at all. Regarding Schwartz and Chylek, when you say I am ignorant - you are saying the same of them. I prefer to think of Schwartz and Chylek as knowing your envelope as well as you do, but they also understand the role of feedbacks and do not overestimate the cooling impact of aerosols. After all, the temp record is more in accord with their calculations than with yours. Back to the issue at hand. I would love to see you explain why it is acceptable for the IPCC to neglect the evidence-based principles identified in the forecasting literature and compiled by Armstrong. You could start with these three:

  • Principle 1: Consider whether the events or series can be forecasted.
  • Principle 2: Keep forecasting methods simple.
  • Principle 3: Do not use fit to develop the model.

I would suggest you start by reading this [3] so you understand the evidence the principles are based on. Once you understand them, explain why these do not apply to physical sciences. I look forward to your reply. Actually, perhaps you should make your reply on your blog, because we are getting off topic now. When your blog posting is complete, just leave me a note here and I will come look. RonCram (talk) 01:05, 20 April 2008 (UTC)

Yes, I obviously gave you altogether too much credit in thinking that you might be embarrassed to see your ignorance so clearly displayed. I've learnt my lesson and will not bother trying to engage you in rational discussion again.Jdannan (talk) 04:39, 20 April 2008 (UTC)
James, please don't take your marbles and run away. Take the challenge! If you really believe you don't need to follow these principles, tell us why. RonCram (talk) 05:18, 20 April 2008 (UTC)

Some principles are so important that any forecasting process that does not adhere to them cannot produce valid forecasts. The following are three such principles, all of which are based on strong empirical evidence, and all of which were violated by the forecasting procedures described in Chapter 8 of the IPCC report.

Clearly that empirical evidence does not include physics or engineering. Prinicples 1 3 and 2 do not make sense for models based on physics. Count Iblis (talk) 14:02, 20 April 2008 (UTC)
Count, thank you for stating your viewpoint but a bald assertion does not move the conversation forward much. I take from what you said that you agree with Principle 3: "Do not use fit to develop the model." But why do you not agree with Principles 1 and 2? You cryptically write: "Clearly that empirical evidence does not include physics or engineering." Why not? RonCram (talk) 01:11, 21 April 2008 (UTC)
I actually meant Principles 2 and 3 ). I work in physics. When I was studying physics and even before that when I was in high school, I used to work with my father who was an engineer (sometimes he would bring problems home and we would work on these if he thought it was something interesting for me). So, I know how physicists and engineers solve problems in practice and not just what the books say how they should work.
Principle 3 is clearly violated in engineering. Engineers will often develop models using empirical data and fit parameters of some (phenomenological) model. E.g., engineers knew for a long time that a theoretical equation for drag in turbulent flow did not give accurate reasults and they used a better empirical equation. Only recently (perhaps 5 or ten years ago, I don't remember exactly) was it discovered that the standard theoretical derivation was flawed and when the error was corrected one obtains the empirical equation.
Principle 2 is not used at all. What matters is that you don't want to make a model too compicated is a simpler model would suffice. But there is no requirement to keep models simple. Einstein said: "Make everything as simple as possible, but not simpler" :) Count Iblis (talk) 13:35, 21 April 2008 (UTC)
Count, okay so you agree that forecasters should follow Principle 1, so we agree there. Regarding Principle 2, I think the scientific forecasters would agree with Einstein's description. In fact, I think that is exactly what they have in mind - so it appears we agree there as well. Regarding Principle 3, I think the engineering process you describe is somewhat different than fitting a model so it provided an accurate hindcast and then saying it had predictive power. In the early 90s, I bought a very powerful software program that looked at a multitude of different market conditions and relationships. I was able to tune this model by weighting different conditions and relationships for a specific stock or market index. Then I could do a hindcast and it see how well it did. With enough tuning, the hindcast would perfectly represent prior trading. The thing is, the model had no predictive power and I ended up losing a lot of money. The engineering process you describe is more akin to the approach Stephen Schwartz took. He looked at the fact the Earth did not warm as much as expected based on the equations then in vogue and he developed a new approach based on sensitivity to perturbation and relaxation time. He determined that climate sensitivity to doubled CO2 was only about one-third the IPCC estimate. It will be interesting to see how well Schwartz's climate sensitivity estimate holds up now that we have entered the cool phase of the PDO. I still think Principle 3 is valid. Using fit to develop the model is not scientific. RonCram (talk) 16:07, 21 April 2008 (UTC)

Manufacturing Consent: The Political Economy of the Mass Media William M. Connolley (talk) 17:48, 21 April 2008 (UTC)