Talk:VAN method

128.112.22.28 was me. Ulixes 23:59, 4 May 2005 (UTC)

Unbalanced

I am tagging this article as "unbalanced" in that it gives VAN more status than warranted. An indication of this is that the eight items under "Further reading" are all by Varotsos and colleagues (and the usefulness of some these doubtful). Recent edits by 134.139.204.63 suggest an attempt to rehabilitate VAN based on a new book; there is no comment on the book, or why it changes the criticisms made of VAN in the 1990s. ~ J. Johnson (JJ) (talk) 19:17, 9 May 2013 (UTC)

As I am neither a supporter nor a critic of VAN, I don't think the article is unbalanced, especially as the "criticisms" section is about the same size as the description of the method. It seems that there are no opposite views expressed recently for a number of reasons. BTW I object to the unqualified statement Their prediction methodology is now in the mainstream, unless 2-3 citations are provided. If none is provided, this statement must be deleted. SV1XV (talk) 05:44, 10 May 2013 (UTC)

When P. Varotsos retires in 2014, the VAN method shall be abandoned, saving a lot of taxpayers' money. His main opponents (Stavrakakis, Papazachos), who also were fighting over government funding, have left seismology and moved into politics, so all discussion about VAN has died out. 149.210.91.19 (talk) 05:57, 10 May 2013 (UTC)

I removed all "further reading" references which duplicated in-line citations. I left the historical references on the 1980s and one recent publication, because it is available on-line and can help the reader understand the concept of natural time. SV1XV (talk) 09:58, 10 May 2013 (UTC)

A couple of citations in the section "natural time" where clearly out of place and I moved them to "further reading". From its title, it appears that "Determination of the epicentral distance of an impending earthquake from the rise time of Seismic Electric Signals" (2003) is also non relevant there and should be moved to further reading as well. SV1XV (talk) 13:29, 10 May 2013 (UTC)

  An alternative was to tag this article for undue WP:WEIGHT. What I hear from seismologists is that VAN is has no role nor credibility in the mainstream; to say otherwise is thus incorrect. (Yes, remove it.) Unfortunately, though this is a "known truth" among seismologists, and there are many articles criticizing it in detail, I have yet to see any source clearly stating that. In addition to the Lighthill book there are at least two special issues of journals on this; I may add some refs when I get an hour or two to spare.

  The idea of "further reading" is for something that goes beyond what is covered in the article, and even better if it is some kind of review by an outside party. But here there don't seem to be any review articles, and all (?) of the "pro" articles seem to be coauthored by Varotsos.

  I don't know that "natural time" really ties in with the VAN method, so perhaps it can be removed entirely. The bit about the Japanese Chinese satellite ought to be deleted as well, as there is no showing that is is related to VAN. ~ J. Johnson (JJ) (talk) 21:20, 10 May 2013 (UTC)

It seems that there is some truth in the anonymous posting by 149.210.91.19. I am surprised that Springer published two books on the subject recently, both written by members of the VAN team. I removed two unsources and out-of-place statements and I think that the article is balanced now. It is not very long either, it just covers the main points of the method without any details. I would like to read the new book by Mary Lazaridou, but it is quite expensive. SV1XV (talk) 02:39, 11 May 2013 (UTC)

  Sorry, still not balanced. Note that "balance" does not mean equal representation of opposing views (criticism), but proportional representation in the weight of the sources. And VAN is repudiated by practically all notable sources. It is discredited, and I think that needs to be in the lead. I'll make a proposal in a little while.

I have deleted the section on "International use" as it was mainly about the prospective launch of a Chinese satellite not connected with VAN.

  The "natural time" bit seems to be an effort to rehabilitate VAN (VAN 2.0?); I wonder it should be dealt with separately. ~ J. Johnson (JJ) (talk) 20:32, 11 May 2013 (UTC)

Proposed replacement of lead

As WP:BALANCE requires treatment in proportion to the weight of the sources, and most sources challenge the validity of the VAN method, I propose the lead be rewritten as folloows:

The VAN method — named after P. Varotsos, K. Alexopoulos and K. Nomicos, authors of a 1981 paper describing it — purportedly measures "seismic electric signals" (SES), by which Varotsos and various colleagues claim to have successfully predicted earthquakes in Greece. Both the method itself and the manner by which successful predictions are claimed have been severely criticized; the International Commission on Earthquake Forecasting for Civil Protection declared: "subsequent testing has failed to validate the optimistic SES prediction capability claimed by the authors".

Notes:

---

If there are no objections I will dig out the missing reference(s), and make this revision. Caveat: I prefer all the references in a References section, linked by {{harv}} short-cites, as it really is easier. But I haven't decided if I want to pull out all the citations and convert to harv, or just dump my references into the refs. Anyone else have any strong preferences here? ~ J. Johnson (JJ) (talk) 21:59, 12 May 2013 (UTC)

Hmmm, I'm afraid that your proposal is also unbalanced for a lead-in, especially by including all those old anti-VAN citations, which are OK further on, in the criticism section. The ICEF one, being issued recently by an organization, is enough to make the point. What is needed is a brief description of the method. Otherwise your text in mostly ok. See this version:

The VAN method — named after P. Varotsos, K. Alexopoulos and K. Nomicos, authors of a 1981 paper describing it (only the 1981 paper, no need for Kagan and Mulgaria here), purportedly measures "seismic electric signals" (SES), by which Varotsos and various colleagues claim to have successfully predicted earthquakes in Greece. (no need for papers here, there are many further on, or perhaps only the new book when accessible) Both tThe method itself, its theoretical foundation, and the manner by which successful predictions are claimed have been severely criticized; the current view of the International Commission on Earthquake Forecasting for Civil Protection is that: "subsequent testing has failed to validate the optimistic SES prediction capability claimed by the authors" (ICEF reference here). Currently (2013) the VAN team is part of the Solid Earth Physics Institute of the University of Athens, Greece.

Also I don't like the section about Natural Time. It is too technical and not at the same level as the rest of the article. I would like to shorten it by removing the maths.

Agree re "natural time" (see next section). As to your version: the main difference in the text is addition of the last line as to their current location, and changing "the ICEF declared" to "current view of the ICEF". For the former, I think it could be worked into the first sentence, but don't have any strong opinion on this. As to the latter: if "declared" seems too strident then it could be replaced with "stated". But this "current view" is weak, suggests that perhaps their future view might change. Which is quite incorrect: the ICEF issued their report, and dissolved; there is no current view, nor future possibility.

Your main changes are with the references, and I am much inclined to keep them in so that the statements don't become isolated from the sources. And Mulargia & Gasperini and Kagan are quite necessary, as the 1981 paper is not readily available, so we need the indirect description of it. These papers are no less important for being "old" (but 15 years younger than the original VAN paper?) or "anti-VAN" (a bit strident, no?).

I would agree that a description of the method would be good, but that kind of detail would not suitable for the lead. And I haven't seen a good description yet. Similar for the supposed underlying physical basis: there seems to have been four or five formulations of it, none amounting to much more than speculation. We can't do better than the sources, and those are wanting. ~ J. Johnson (JJ) (talk) 03:08, 15 May 2013 (UTC)

If we include all these anti-VAN reference in the lead-in, then the proper name of the article would be "Criticism on the VAN method". It is better to keep the lead in as simple as possible, there is enough space for all this further on, unless we want to carry over the debates of 1990s here. BTW, I have been watching from a safe distance and my personal view is that both sides at the time were bending the truth and using personal connections and publicity tricks, each for their own reasons. In 2013 it is best to keep the article short, neutral and balanced and let VAN fade into obscurity within the next couple of years, and then we can rework the article into the past tense. SV1XV (talk) 04:44, 15 May 2013 (UTC)

I am a bit dismayed by your repeated reference to "anti-VAN" sources; it seems to me that you do have a bit of inclination here. That the "VAN method" is controverted is a fact; to state that Varotsos et al. claim successes without mentioning that those claims are controverted would amount to promotion of VAN. My proposed text is very simple: it states 1) the claim that the VAN method successfully predicts earthquakes, 2) that this has been criticised, and 3) the finding of the ICEFCP, which as close as we can get (so far) to settling the matter. I believe that is reasonably balanced. (For the missing "pro-VAN" references I have in mind one or two of the early VAN papers, and the editor's introduction to the TP issue.) If there is unbalance in this it is in giving VAN equal weight whereas NPOV (specifically, WP:WEIGHT) requires proportionality.

I am a bit dismayed by your repeated reference to "anti-VAN" sources; it seems to me that you do have a bit of inclination here. That the "VAN method" is controverted is a fact; to state that Varotsos et al. claim successes without mentioning that those claims are controverted would amount to promotion of VAN. My proposed text is very simple: it states 1) the claim that the VAN method successfully predicts earthquakes, 2) that this has been criticized, and 3) the finding of the ICEFCP, which as close as we can get (so far) to settling the matter. I believe that is reasonably balanced. (For the missing "pro-VAN" references I have in mind one or two of the early VAN papers, and the editor's introduction to the TP issue.) If there is unbalance in this it is in giving VAN equal weight whereas NPOV (specifically, WP:WEIGHT) requires proportionality. ~ J. Johnson (JJ) (talk) 00:03, 16 May 2013 (UTC)

I just read the part of the ICEF paper cited -- Thank you for the link and resource! From the text (p. 335), where it says "can provide short-term diagnostic precursors for ruptures in specified source regions" it is clear that the ICEF is still commenting on the earlier incarnation of the VAN method. I wanted to add this comment in here, for emphasis. Their early method tried to identify SES and make predictions based on them. Their updated method with NT takes SES as an indicator of criticality in a region, and then looks at the seismic events in that region. Their updated method doesn't claim that SES provide a basis for short-term prediction. There's been no criticism of their updated method, as far as I know.Daniel Helman (talk) 21:48, 30 May 2013 (UTC)

PS The critical sources that the ICEF report cites are: Geller [1996], Lighthill [1996], Mulargia and Gasperini [1992] (from p. 336, footnote 151). These are all criticisms of the earlier VAN method.Daniel Helman (talk) 23:44, 30 May 2013 (UTC)

Wait -- I just found the most recent report critical of the VAN method. It's from 2010. http://onlinelibrary.wiley.com/doi/10.1029/2010EO180003/abstract It's not critical of the method itself, but rather that Varotsos claimed to have predicted a quake -- but the prediction wasn't made through official channels in Greece. Instead, it was published in arXiv. I think his predictions aren't welcome anymore in the official channels, based on his earlier method.Daniel Helman (talk) 00:10, 31 May 2013 (UTC)

Electromagnetic compatibility issues

I believe that the VAN team have never published circuit diagrams of the SES front end, which critics claim that may respond to industrial frequency and RF signals. I have only seen a set of block diagrams of the front end, without specific details (K. Nomikos: "The telemetric of VAN group", A Critical Review of VAN, 1996). Even the ICs used in the front end are not listed. I don't know of any evaluation of EMC performance of the front end by an independent laboratory. This lack of information reduces the credibility of the method and allows opponents to publish claims about TV signals, military UHF repeaters, etc. Has anyone seen a detailed description of the analog part of the VAN system ? SV1XV (talk) 07:13, 10 May 2013 (UTC)

I haven't trawled through all of the multitudinous Varotsos papers, but I have neither seen nor heard of any technical details other than what you see in Lighthill. Nor any description of how they evaluate their SES, though Nagao et al. (in the GJI issue) give some overview. Note that it is not a matter that the lack of such info allows the critics to make claims (those claims standing on their own merit). Rather, it is the lack of info necessary to reproduce the results that precludes acceptance of the claims of successful. ~ J. Johnson (JJ) (talk) 21:01, 11 May 2013 (UTC)

It is both. A key characteristic of pseudoscience is not allowing any interested third parties to independently reproduce and evaluate the method. A detailed description of the collecting system and the front end is essential in order to reproduce the experiments under equivalent conditions, and to compare VAN recordings with possible similar signals recorded by researches using different devices. Details of the digitizer and the data transmission and recording systems are less critical.

At the same time a well managed research program should have made all efforts to ensure that there are no EMC/EMI issues. I understand that in the 1980s the current EMC standardization, testing and certification infrastrucure was not available yet in Greece, however they should have investigated possible EMC issues using the limited resources available then. They could also have done it later in the mid-1990s, when such services were becoming available at a reasonabler cost.

On the other hand, I have seen how the lack of information allowed critics to make wild claims. In Greece and in Japan supporters and opponents of VAN were trying to discredit each other and gain public support. Both sides used all kinds of publicity tricks in the mass media, there was even a pro-VAN manga comic. One repeated claim by VAN opponents was that the recordings were actually subtle EMI at the VAN central facility. This was clearly unfounded, as the signals were digitized near the source and transmitted over leased lines using modems. SV1XV (talk) 05:40, 12 May 2013 (UTC)

  Ouch! Your comment that "[b]oth sides used all kinds of publicity tricks in the mass media" suggests a disparagement of the controversy, even that the criticism is just petty bitching. Having read a fair portion of the criticism, and Varotsos' responses, I would say that the criticism is valid. I haven't gotten to the EMI claim, but as to your personal basis for disbelieving it (that the signals were digitized): well, according to one report I just saw some were, but some were not. So I would be very cautious about rejecting that criticism without looking closer at its basis. Some of the other criticism (such as the ambiguity and imprecision of the telegrams, the adjustment of parameters, and various statistical analyses) seem pretty solid, and the attempted refutations not persuasive.

  BTW, that report was Nomicos & Chatzidiakos 1993 (Tectonophysics 224:39-46), which has a description of the technical aspects that I believe is more extensive than the report in Lighthill. It is kind of interesting for being state-of-the-art PC-AT based computing circa 1990. (Ouch!) ~ J. Johnson (JJ) (talk) 21:45, 12 May 2013 (UTC)

Well, the publicity tricks dnn't refer so much to the scientific publications but mainly to the use of mass media and manipulation of the public opinion by both sides. Criticism in scientific journals is mostly valid. The story about local interference at VAN HQ is just one of these stories for public consumption, any EMI to the system takes place at the measurement location. This is the view published by Stavrakis et al. as well.

Yes, you are righr, I am told that the VAN team still use electronics of the mid 1980s. Their front end claims best CMRR of 88 dB when >100 dB is specified for cheap ICs like Burr-Brown INA-110. Also there was a public discussion recently about funding, and it appeares that in the age of internet and ADSL they still spend much of their budget in leased telephone lines.

In any case, there is an old axiom which says that a valid but controvential theory becomes mainstream when it manages to survive after the death/retirement of its opponents. An ivalid theory usualy dies out and fades into obscurity after the death/retirement of its supporters. We have now reached this critical point, supporters and opponents retire one by one and heated discussions have died out. My guess is that the VAN method has no potential for further development and shall fade into obscurity. SV1XV (talk) 02:28, 13 May 2013 (UTC)

  I think the heated discussions started around the time of the Athens conference, and peaked in 1996/1997 with the GJI issue and Geller's article. Likely some editors cooled a bit. Quite aside from any PR antics I think the criticism has been pretty much accepted by most of the seismo community. But V's wife is still going strong (have you seen her new book?), and I suspect that as long as WP editors are required only to show sources in support, and not a complete review of the field, this will remain be a pop favorite for many years.

  Saw a curious statement in Yoshii (1993) last night. He said that the opponents may have "vested interests. On the other hand, the VAN group is only concerned with the discussion of the scientific feasibility of earthquake prediction as a result of SES detection." Right. Nothing to do with professional credibility and control of a sizeable budget. ~ J. Johnson (JJ) (talk) 22:23, 13 May 2013 (UTC)

Yes, I ordered the book from Amazon a few days ago and my copy shall arrive in a few days. SV1XV (talk) 02:39, 14 May 2013 (UTC)

I have seen only the toc, and a couple of pages from several chapters. I am quite curious as to whether she addresses the criticism, aside from just dismissing it. But not curious enough to buy it! ~ J. Johnson (JJ) (talk) 00:00, 16 May 2013 (UTC)

A look at google books ngrams for "VAN method" is interesting [1] - it appears to be dying away (shame there's no data after 2008). Mikenorton (talk) 07:15, 16 May 2013 (UTC)

Nice. But I suspect that the sudden fall-off after 2007 is simply the edge effect of incomplete data. I wonder if something Georef could be mined for mentions per year. ~ J. Johnson (JJ) (talk) 19:20, 16 May 2013 (UTC)

The concept of "natural time"

I believe this section is too technical and esoteric. I would delete most of the material and replace it with the following, much simpler version:

Since 2001 the VAN team has tried to improve the accuracy of the estimation of the time of the forthcoming earthquake. Their methodology involves the through introduction of the concept of natural time. This is a parameter which puts weight on a process based on the ordering of events. (two references) The VAN team claims successful predictions, that from 2001 to 2010, of twenty five of the 28 major earthquakes in the region from N36 to N41 and E19 to E27. were successfully predicted Predictions are part of papers uploaded in arXiV. [minor detail] (reference) Their work was documented in a book published by Springer in 2011, titled "Natural Time Analysis: The New View of Time." (book ref)

If this deletion is accepted, we could even delete the separate section and move the remaining text into "Description of VAN". SV1XV (talk) 15:52, 14 May 2013 (UTC)

Yes! I fully agree. I haven't seen any plain statement of what "natural time" is supposed to be about, neither by them nor anyone else, and suspect it's not so much esoteric as esoteric gibberish. But the text needs some wordsmithing (I'll take a whack at it), including what these regions are. If we mention their predictions we really should look around for some comments on them. If there is no comment, then there these might qualify as WP:FRINGE (see also WP:PROFRINGE). ~ J. Johnson (JJ) (talk) 02:09, 15 May 2013 (UTC)

If NT is used as a parameter to describe the evolution of an earthquake, then it is not fringe, it is just a technique. It becomes a fringe theory when the VAN team suggest that it is a general characteristic of almost all dynamic systems, especially when they include biology, cardiology etc, so it is better not to include jere such claims. SV1XV (talk) 03:01, 15 May 2013 (UTC)

Whether "natural time" is theory, technique, or something else, if they are the only people using it — e.g., "not broadly supported by scholarship in its field" — then it is likely fringe. ~ J. Johnson (JJ) (talk) 03:36, 15 May 2013 (UTC)

I read their 2011 book by Springer. Natural time is the name they give to their time series analysis technique. Their use of language is somewhat misleading -- For example, in the book they introduce a term called "entropy in natural time" which is just a variance term working with data and has nothing to do with thermodynamics. Natural time is the time series analysis they are using currently for SES prediction. BTW I'm 134.139.205.253 and had made changes to this article as I was doing research on microtectonic deformation mechanisms and related electrical phenomena. I don't know whether the VAN method (as it stands now, with Natural Time analysis) is replicable. Daniel Helman (talk) 07:04, 27 May 2013 (UTC)

I figure if Springer has recently published a book by them, then they are legitimate, despite the earlier criticisms. Maybe I'm being naive. Better would be if some seismologists did the work to test their current method. Daniel Helman (talk) 07:11, 27 May 2013 (UTC)

Thanks. Sounds like you have some basis for understanding what they are saying -- if they are actually saying anything. You might be interested in downloading some of V's papers on this; see his bibliography, esp. part IV. Are you aware of any criticism of this "natural time" stuff? They have been claiming more successful predictions, but I suspect that "natural time" stuff is just a way for them to get more knobs to twist (degrees of freedom). I haven't seen any criticism (have you?); I wonder if most seismologists now ignore VAN. ~ J. Johnson (JJ) (talk) 21:29, 27 May 2013 (UTC)

Oh their language is so difficult. It took a long time to get a clear picture of what they are doing. Their major methodological flaw still seems to be in how they build the selectivity map of a region. They make the assumption that if there is a time lag between electric and magnetic signals, then that's an indication that there was an original signal which traveled from a distance. They make the further assumption that a signal traveling from a distance must be natural, and not artificial. To their credit, there are some 25 causes for telluric currents, and there's no real way to tell, other than looking at frequency data and modeling transmission. I like the fact that they are trying. They set up multiple pairs of electrodes, in order to make sure that signals are not just local noise. They also set up electrode pairs at different distances, to get a sense of whether the electric potential changes with distance. They set up electrode pairs orthogonally, in order to make some sense of the direction of an incoming signal. They also do a detrended fluctuation analysis (DFA), to see whether the data stream is coherent. It's messy, in my opinion, but the mechanism itself is not clean. The only SES that are definitely real are the ones associated with the earthquakes themselves, occurring within a few hours of the quake. What causes these SES might cause earthquake lightning too, for example, but the mechanism itself is unknown. Since some occur before the quake, they are part of a proximal mechanism which has not yet been described. To me, this is wonderfully interesting science and conjecture! The other classes of SES -- whether coherent signals or a gradual change to the base level of electric variations -- these could be due to other causes. It's unclear what the method is of connecting SES detected at a site with the seismic events in a region, to build the selectivity map. Also, earthquakes happen at depth, so that might simplify, or complicate the modeling. They don't generally take a vertical reading of voltage. The variations they do note are on the order of millivolts.

Regarding natural time analysis -- I haven't seen any criticism of it. I'm not an expert! Their method does seem legitimate -- simply signal filtering, with some terms introduced to give weight to the ordering of events. This kind of work is empirical -- if they get results, then it's acceptable. I'm concerned that some of the criticism of their earlier method was directed at their catalog of seismic events. They may have been leaving some earthquake data out deliberately -- at least that was the implication of some of the criticism. Or, they may just have taken seismic data from one source and not another. None of their work has been retracted, as far as I know. The criticism might be unfounded. In the Springer book they did include a set of seismic events they predicted between 2001 and 2010, and it wouldn't be that hard to see whether they have left out any. I think mostly their work has been ignored. Again, I do like the fact that they are trying to do something. Natural time -- that's just their way of saying time series analysis. There's nothing particularly controversial about it, other than their choice of language, as far as I can tell. It's just data analysis, looking for criticality in the data. Mostly they like to have their filtered data with values between zero and one, so that the numbers get smaller when they are multiplied together, and you can pick out changes or set different critical thresholds easily. Thanks for the link to their work. BTW, some of their articles have the term "entropy" -- and this term has nothing to do with thermodynamic entropy. They've just constructed an equation to look at variance in the data (similar to their kappa term) and picked "entropy." A confusing choice of nomenclature.Daniel Helman (talk) 04:55, 28 May 2013 (UTC)

"[S]imply signal filtering"? That has always seemed to me to involve heavy statistics, not notably a VAN strength. And V's article "Is time continuous?" (which I haven't read) makes me wonder if he thinks he has some profound fundamental insight.

I doubt if V et al. will ever retract any of their papers, despite some pretty solid criticism. The problem isn't so much leaving out any data as leaving out "the rules of the game" by which their results should be judged. Basically, they keep fudging the acceptance criteria. They keep saying they have predicted earthquakes, but close examination keeps showing problems. That there is no current criticism is probably because most seismologists think they have nothing new. (But, "citation needed"!) ~ J. Johnson (JJ) (talk) 01:07, 29 May 2013 (UTC)

Data can be listed as information, and signal filtering is the things a person does with the data to make patterns stand out. It's similar to statistics, in that both are trying to arrange or modify data to be able to say something novel. In statistics, the main point is to show that a small group of data (the statistic) can be generalized for all the data (the population). In signal processing, the data modifications can be for any number of reasons. For music, people change the data for aesthetic reasons. For astronomy, people change the data to get a maximum contrast on the feature you're looking for. The "natural time" analysis that V et al. are doing is close to music and to astronomy work -- trying to heighten a feature.

Regarding "Is time continuous?" the title is catchy. But it's just a piece about signal processing and how to weight events within an ordered series. The "natural time" weighting isn't complex. The first event gets a value of 1 over the total number of events. That's multiplied by the energy from the first event. The second event gets a value of 2 over the total number of events. The third event gets a value of 3 over the total number of events. This method just puts extra weight on the newest event. As far as I know, there's been no criticism of their "natural time" analysis. The criticism of their statistics was damning enough that they gave up on predicting earthquakes based on specific SES -- and started with their new method of time series analysis, looking at a set of SES data for each region. Their time series methods have changed, as well. They have two methods where they look at seismic data once the SES have reached a critical state. The first was to look at variance terms for the seismic data. Their second was to split up the seismic events into overlapping regions, and look at the variance terms as they appear in these regions. They will probably come up with a third method, or a fourth, to optimize their time series analysis. This also seems reasonable to me, since the method is empirical. It would be good if some outside scientists take the time to confirm or find flaws in what they are doing now -- to set up some electrodes and actually try to replicate their results. Springer published their work, so, hopefully, the scientific community will step up and explore it. I don't see much of a risk here. This is how science works. If their new method with time series analysis is confirmed by others, then it really will be a good day for seismology. If their new method is found wanting, then they will go back and try to resuscitate it somehow. If they are shown to have committed fraud, then the science publishers will retract their work, and the government in Greece might put them in jail. It is breathtaking, the scope of what they are trying to accomplish. For me, the validity of their work hinges on whether SES are valid for the smaller quakes which are used to build the selectivity map of a region. Those SES might not be valid! I'm coming from a naive perspective: I don't assume that they are committing fraud outright, despite: (1) very large claims, (2) confusing use of language, (3) questions over which seismic data they include in their analysis, and (4) unclear presentation of their analysis. Point number three means that, at the least, their scientific methods could be a bit wanting. But Springer has published their work, and the journal "Tectonophysics" continues to publish their work. I'd like to see someone try and replicate their results. Someone probably will try, since Springer is a major science publisher.Daniel Helman (talk) 01:19, 30 May 2013 (UTC)

All very interesting. But NT is about how to process (measure) something, which is different from the thing measured. And the original "VAN method" did not use NT, so we have to ask: is NT an augmentation of "the VAN method", or something new? Not entirely new as long as it is based on SES, which seems to still lack a correlation with earthquakes. The claimed successes seem to be more to do with how they select their data.

BTW, in the last paragraph of the NT section "successfully predicted" and "according to the VAN team" are at the opposite ends of the sentence, weakening this key caveat; I may jump in and tighten that. Also, nice work you're doing. Especially with the references. When ever I do a mass change like this I get nervous that people won't understand it and keep jamming references into the notes. I rest easier seeing that you have good grasp of that. (Attaboy!) ~ J. Johnson (JJ) (talk) 18:00, 30 May 2013 (UTC)

Thanks for the praise and comments! I think the NT section would be more readable if the math in the middle paragraph were laid out in larger type. It's not all that complex, but the small type makes it hard for the reader. I may try and see how other pages with math are formatted, and then reformat that paragraph. The article is coming along!Daniel Helman (talk) 21:19, 30 May 2013 (UTC)

PS I don't want to jump to any conclusions about their new method. I haven't seen any criticism of it yet.Daniel Helman (talk) 21:19, 30 May 2013 (UTC)

  Thank you for doing a good job. I'm so glad it doesn't all depend on me. :-)

  A large part of the criticism is that their predictions are broad and imprecise, and having made a prediction they subsequently stretch it to fit what ever earthquakes do occur. Having been called on that, and not showing any interest in tightening "the rules", I think the criticism is still applicable to their predictions regardless of how they derive them.

  You know about the wiki-markup for doing the math? Though this does raise a question of what level of math is appropriate. I certainly have no objection to writing for a higher level, but we do have to present it in a way that won't entirely baffle the general audience.

  One other thing: I think we should change the header to «"Natural Time" analysis».

  No, two other things! Are you familiar with "named refs"? I hate them (ask for details), and unfortunately there are bots running around that think it a Good Idea (bah!) to merge identical <ref>s into named refs. I point this out because I see you have cited Varotsos, Sarlis & Skordas 2011 twice. The thing to do is differentiate them in some way. Specifying a page number is one way. I see we have several instances of duplicate refs, which I'll look at, but I don't have VSS 2011 so I'll leave that to you. ~ J. Johnson (JJ) (talk) 20:12, 31 May 2013 (UTC)

Wonderful to be sharing the work. The new VAN method is to use SES to make medium-range predictions, and then use seismic data for short-range prediction, from what I know. Typically, short-range prediction with seismic data is done by filtering out certain events (like aftershocks) so their updated VAN method seems in line with current seismological practice. I haven't seen any criticism of their using SES to make medium-range predictions. I also haven't seen any criticism of their using the SES medium-range prediction to initiate a short-range prediction based on seismicity. They have two versions of that short-range algorithm, but I'm not aware of anyone in the literature criticizing that. Having a medium-range prediction seems to be a moral good, despite the moral failing of the earlier VAN method. They had essentially been yelling "fire" in a crowded theater. Their earlier (1980s) telegrams generated a lot of ill will. That is why they only post to arXiv now, and not to the official earthquake agency in Greece. When you write "the criticism is still applicable" I'm a bit confused. There has been no criticism of their updated method, as far as I know. On the other hand, the folks in Japan are using the updated VAN method successfully. These two groups are in touch with each other, so that doesn't qualify as "independent" verification -- and I do think that independent verification is critical. Regarding Wikimarkup for math -- I've never done it, but I'm sure I'll have a good time at it (though feel free to step in, if you feel inclined!) Regarding the level of math -- The highest level in the NT section is an expression with a summation. I also think it's a bit hard to understand right now. I was hoping that with in a math format, it would be readable. If you can read the expressions, they're really not very demanding at all! There's some exponents, parentheses and new terms -- but no knowledge of calculus, specialized functions or anything like that is needed. Regarding the NT header: I think it has at least one thing going for it -- it emphasizes why the NT section is here, namely for earthquake prediction. Regarding duplicate refs, I'll go through and try and put in chapter numbers. I already returned the book to the University of Houston library (inter-library loan) and don't have the page numbers anymore. I'll have to look and see if Amazon has a TOC available, and can find the chapters from there. I was pleased that you'd put page numbers in a lot of refs, and now I see one reason why. it may take me a few days to get to any of this -- alas. I'm finishing up a 500 page paper, and then have a journal article to revise, and a chapter to a book to write (unrelated). I'm so glad to be here, collaborating! Regards.Daniel Helman (talk) 22:27, 31 May 2013 (UTC)

PS I don't think NT warrants its own wikipedia page. But maybe the math could be cut from here, and added to a new wikipedia page on natural time analysis. On the other hand -- the only application of NT has been for SES and earthquake prediction. Heart health (cardiogram traces) and other phenomena have more successful time series protocols, so it is unlikely that NT will be picked up outside of seismology. I do think it's important that a reader should be able to find some treatment of NT in wikipedia that includes the math.Daniel Helman (talk) 22:56, 31 May 2013 (UTC)

  I agree. NT's notability seems to be solely and strictly as an augmentation to "the VAN method", so here it belongs. But hardly more than a mention until they show it's actually useful. (If they have any particular insight more profound than just tweaking the math, perhaps that can described without the math?) They claim successes, but they seem to lack credible showing that their claims are any better than before. And I don't think there is anything better in their "medium-range" predictions; the criticism is still applicable.

  Regarding chapter pagination: note that that goes into the full reference (showing what portion of the containing work. The specific references, using the {{harv}} templates, takes the specific page (sometimes pages), section, or such being cited. It's pretty obvious once you "see" it, but sometimes people miss it at first glance.

  Good luck with all that real work. Take a few days, this stuff isn't urgent. ~ J. Johnson (JJ) (talk) 21:40, 3 June 2013 (UTC)

---

I looked at a little of the "natural time" material, and it looks like junk. (Just window dressing to burnish the discredited SES selection process.) Varotsos also claims that it can identify cardiac patients at risk of sudden death, but Google search results suggest he is being ignored by the medical community. I think there is a strong case for this being a WP:fringe theory, which should be clearly identified as such. ~ J. Johnson (JJ) (talk) 23:17, 18 June 2013 (UTC)

While Varotsos in Greece and Uyeda in Japan are on faculty at major universities, and considering that Springer has published a recent collection of the updated VAN method that hasn't been discredited -- I think it's premature to call their work a fringe theory, and such a designation wouldn't help the reader as much as being very clear in this article about what is claimed (a successful prediction protocol involving SES and time series analysis), what has been demonstrated (a successful prediction protocol for a limited number of events based on purported SES -- such that the SES activity, when it reaches criticality, creates a medium range prediction, which is then used as a seed for seismic data analysis) , and what has yet to be demonstrated (all of it needs to be repeated by outside researchers). Springer doesn't generally publish fringe science. But I do agree with you -- the presentation of Varotsos et al.'s work is a little sloppy. I wouldn't characterize Uyeda's work as sloppy. I wouldn't call their time series analysis window dressing -- they have abandoned individual short-term predictions based on SES, and now favor using SES for medium-term prediction.134.139.205.142 (talk) 18:21, 26 June 2013 (UTC)

References

Currently all of the "references" (full citations) seem to be setup in cite templates, but in each individual footnote. I am thinking of pulling them all out and putting them in a real "References" section (not just a spot for {{reflist}}) because this is easier. But there are some other complications, like then I would have to put short cites in the text/notes. Comments? ~ J. Johnson (JJ) (talk) 22:28, 25 May 2013 (UTC)

Did it. Replaced full references in the text with Harv short cites. (Some of these could use page specifications.) Corrected some errors and augmented the references, put them in alpahabetical order in "References". ~ J. Johnson (JJ) (talk) 23:38, 29 May 2013 (UTC)

"Primary sources" tag.

I propose removing the "primary sources" tag that asks for secondary or tertiary sources to be added. The statement is true, but the reality is that essentially there are no secondary or tertiary sources on this topic (Wikipedia leads the pack!) except for the Mary Lazaridou-Varotsos book, which is seriously COI. ~ J. Johnson (JJ) (talk) 22:29, 31 May 2013 (UTC)

I concur. I stuck many of these primary sources in here. I don't know of any secondary sources in English. There may be some in Greek or Japanese.Daniel Helman (talk) 22:34, 31 May 2013 (UTC)

And the original tagger didn't stick around to justify the tag. Okay, I'm taking it out, and anyone that objects should suggest what other sources to use. ~ J. Johnson (JJ) (talk) 18:13, 1 June 2013 (UTC)

Criticisms of VAN

Anyone else think that some of the text for the bullet points rambles and is unclear? I think each point should be rewritten for style. I'll do it if I can get around to it -- very busy right now. :) Daniel Helman (talk) 22:41, 31 May 2013 (UTC)

  Before we delve into style I wonder if we ought review just what the criticism is. That would be a bit of a project in itself.

  By the way, regarding your last edit ("Inaccurate predictions ... lead to public criticism."): we need citations to sources. As that paragraph is about the public criticism in Greece (not the scientific criticism), I expect the best sources are in Greece, and in Greek, which does pose a bit of a challenge. There may be other editors that can help with this. There is a corresponding article in the Greek Wikipedia, but all the references there are scientific. (And 24 out of the 30 references by Varotsos or supporters, suggesting that they have "captured" that article.) ~ J. Johnson (JJ) (talk) 18:41, 1 June 2013 (UTC)

Agree. I'll look for a source to add.Daniel Helman (talk) 21:32, 1 June 2013 (UTC)

I have a bunch of articles (not all) from the TP vol. 224 and GRL vol. 23; I'll take a look through them. If you (or anyone else?) have the Lighthill book that would be good to check. Perhaps we should look for Greek editors. And how do you feel about a review of the criticism? ~ J. Johnson (JJ) (talk) 22:03, 2 June 2013 (UTC)

I'm looking forward to posting a draft of VAN criticism here, but it will take me another couple of days to get some other work done first. Best wishes.Daniel Helman (talk) 18:51, 8 June 2013 (UTC)

How about we run up an outline first, to get things organized? I'll start hacking on a list. BTW, I came across Stiros 1997. Pretty interesting. ~ J. Johnson (JJ) (talk) 20:32, 8 June 2013 (UTC)

Sounds good to me. I'll post something in the next few days.Daniel Helman (talk) 08:20, 10 June 2013 (UTC)

I'm about to start another section we can use as a work area. ~ J. Johnson (JJ) (talk) 21:55, 10 June 2013 (UTC)

Summary of criticism.

Scientific articles critical of VAN in chronological order, with summaries.

My short-form naming scheme (optional!) is straight forward: author(s), year (yy), a code for the journal, volume. In some cases page number (opt.) follows an underscore.

---

• MulargiaGasperini92-gji_032

Apparent success of VAN predictions is just the result of altering the "rules of the game", can be confidently ascribed to chance. Stronger correlation with prior events.

• DrakopoulosEt93-tp224_223

Of 13 VAN predictions from Sept. 1986 to April 1988 none are deemed successful, because there was no quake, or none within the prediction parameters. In seven EQs with comparable magnitudes occured, but no SES reported, "contrary to VAN’s assertion of one to one correspondence between SES and earthquakes."

• GruszowEt96-grl23

Independent observation of signals interpreted by the VAN group as SES for the May 13 Kozani earthquake have magnetic and other characteristics that suggest they are generated by artificial (industrial) sources.

= Special issue of Geophysical Research Letters:

• Wyss96-grl23_1299

"Inaccuracies in seismicity and magnitude data" used by VAN, including unspecified or doubtful magnitudes, and addition of undocumented EQs. The '"one to one in correlation" between ESs and EQs, as claimed by VAN, does not exist' as the record of actually observed EQs is misrepresented, and the record of claimed observations of SESs is unreliable. VAN also makes major changes in the hypothesis, abandoning claims that were previously vigorously defended.

• WyssAllmann96-grl23_1307

Agree with Jackson [1996] and Rhoades and Evison [1996] that the VAN works should not be evaluated because parameters are not defined. "VAN continue to promote untestable statements instead of a well defined hypothesis." Evaluation of 22 claimed successes shows (depending on the parameters used) 6, 11, or 14 hits, with the probability of occuring by chance ranging from 0.50 to 0.98. Also, 79% to 98% of the eligible earthquakes are missed by VAN.

• Kagan96-grl23_1315

The high rate of sucess claimed by VAN is "due either to the retroactive adjustment of prediction rules or the the non-randomness of seismicity." The VAN prediction's rate of success can be attributed to chance

• MulargiaGasperini96-grl23_1323

"No published work has so far provided a sound support to the candidacy of the VAN signals as earthquake precursors. The problem is complicated by a chronic fuzziness in the text of VAN's predictions, which often allow different interpretations, and by the absence of any clear statement by VAN of the laws of the game." VAN predictions in 1987-1989 "are not associated with earthquakes beyond chance", but do "appear to be signficantly associated in reverse time".

• MulargiaGasperini96-grl23_1327

Previous assessment confirmed: "VAN predictions have alarm and success rates too low to accept them as candidate precursors". Tendency "to follow rather than precede earthquakes."

• StavrakakisDrako96-grl23_1347

By analyzing all of the VAN-telegrams available to us, we conclude that in no ease was a mainshock predicted. Whenever the reported VAN-correlation seems to be satisfactory, the correlated event was either an aftershock or the correlation was not unique, meaning that any of several earthquakes in the same focal region could be arbitrarily correlated with the telegram."

• DrakopoulosStavra96-grl23_1355

Analysis of a supposed prediction in January 1991. "This ... case demonstrates the lack of reliability of SESs for use in earthquake prediction." Public announcement also in direct contradiction of code of ethics concerning earthquake predicition.

• Jackson96-grl23_1363

"The VAN technique as not yet been stated as a testable hypothesis." The telegrams "do not fully specify the predicted events, and all of the published statistical evaluations involve many subjective ex post facto decisions." "Presently the VAN method cannot be described as successful predicting earthquakes in beyond chance, in spite of several nominally positive results. In every one of the tests, the rules of the game were specified after the occurrence of the earthquakes used in the test. ..." Geophysically implausible. No reported SES anomalies at the time of an earthquake, when the geophysical effects might be expected to be greatest

• RhoadesEvison96-grl23_1371

"Objective tests on the performance of the method, using independent data, cannot begin until the VAN hypothesis and the null hypothesis have been fully formulated." Lacking a thorough formulation of the VAN hypothesis the position we are in is quite similar to what it would be with a proposal which had no scientific basis."

• Burton96-grl23_1379

"... VAN is plainly not successful in many aspects of their predictions. There still appears to be a lack of a rigorously controlled data gathering and prediction exercise uncontaminated by a posterori consideration of methodolgy."

• Stark96-grl23_1399

The null hypothesis "might be rejected not because VAN does well, but because the probabilistic model for seismicity is wrong." 'The statement in the Appendix of Varotsos et al. [1996] that earthquakes are statistically independent if they are "isolated in time and space" is false.'

• KaganJackson96-grl23_1433

"... there is no convincing proof that the VAN method predicts any earthquake beyond chance." Varotsos and Lazaridou [1996] argue that the prediction rules for 1987-1989 were established before the test period. This is not so. The prediction rules ... have been adjusted retroactively." Tolerances have grown with time, 'suggestng a consistent pattern of expanding the windows to include more events as "sucesses." This window enlargement is still continuing...." "... there is not yet a specific hypothesis to test."

= In the Lighthill volume:

• ?? [Someone else needs to list these.]

= Subsequent:

• Geller97-gji131_425 ‡

§4.5 "There appears to be no convincing evidence that any of the electrical signals observed b VAN are earthquake precursors." "VAN's 'predictions' are vague and ambiguoius .... Thus VAN are not making earthquake predictions in the first place." "VAN's 'successful predictions' are not statistically significant ... [and] correlate much better with preceding, rather than subsequent, earthquakes, as they were issued preferentially during periods of heightened seismic activity. Simple prediction algorithms using only seismicity dta perform at least as well as VAN." A critical reexamination of ten claimed successes shows only one hit, and that was part of a swarm. Additional comments on evaluation in §4.7.

• BernardEt97-gji131_467

"... VAN’s method cannot be properly evaluated as long as these predictions are not clearly and unambiguously presented, referring to well-defined rules; hence, it is not, sensu stricto, a testable prediction method. The only public alerts which seem to have worked up to now are those produced by nature itself, in the form of foreshock sequences strong enough to generate anxiety and to lead people to leave buildings."

• Stiros97-gji131_478

A "successful earthquake prediction in Greece would be of, at best, limited benfit to society. ... alternative use of funding could be expected to save more lives with much greater certainty. Over the past 15 years, the VAN group’s research on earthquake prediction has absorbed a substantial fraction of the resources devoted to earthquake research and protection in Greece. However, the VAN method has not advanced the nation’s policy on earthquake protection planning, its results continue to be widely questioned by both the Greek and international scientific communities, and the underlying model is not commensurate with currently accepted thinking on earthquake generation and tectonophysics."

"... this group’s work has been presented and/or interpreted as a ready- to-use, operational and successful method for short-term earthquake prediction." "With one exception, VAN’S claims of successful predictions have not been officially recognized as accurate or reliable by Greek authorities." Scientifically unsound. No reason to believe that the empirically compiled 'SES selectivity maps', based on only a few years of observatons, are reliable.

• PhamEt98-grl25

Long transient electric signals without measureable correlated magnetic signals, such as claimed for SESs, are attributed as artificial radio signals. Observations VAN interpreted by VAN as SESs prior to the 1995 Kozani earthquake are attributed to a radio signal. "... careful studies of electromagnetic noise ... should preced lengthy discussions of the statistical signficance of recorded signals." Absence of a magnetic signals has been considered by VAN as a firm criterion to select genuine SES. This paper shows examples of electric signals from artificial sources which are not accompanied by observable magnetic varitions. "The criteria adopted by the VAN group are not sufficient to distinguish SES either from the noise due to remote sources or from natural MT disturbances." The characteristics of the electromagnetic noise need to be thoroughly studied prior to discussions of the statistical significance of the issued predictions.

• JacksonKagan98-eos79

Despite GRL 23 and the Lighthill book, "the myth persists. Here we summarize why the VAN group's claims lack validity." "The SES are supposedly observed only at special sensitive spots depending on the epicenters. ... are reported only as precursors, not at the times of the quakes themselves when the greatest stress changes occur. No other seismic, geodectic, or geophysical changes regularly accompany the SES." Laboratory piezoelectric effects don't scale up. No model explains the sensitive spots; no evidence of spatial variations in conductivity such as VAN claim. "Published works by the VAN group lack documentation, adequate formulation of hypotheses, and testing. ... VAN's claims of successful predictions for larger quakes in Greece collapse under scrutiny. VAN's hypothesis is ill-posed, and the many published descriptions of it are incomplete and inconsistent [Kagan and Jackson, 1996]. Not only are VAN's "predictions" absurdly vague, but VAN also some times claim "successful predictions" that violate their own stated criteria and common sense." "The vagueness of the predictions has grown steadily with time." "The VAN hypothesis badly violates physical intuition, it is too vague to test, and its proponents' claims of successful prediction are greatly exaggerated."

• PhamEt99-pepi114

Origin of so-called SES signals unknown. "... transient electric signals looking like SES can be generate3d by digital transmitters of the radio-telecommunication network. ... It is concluded that some of the signals at [the VAN IOA] site are probably of artificial origin, and the the criteria used by the VAN group are no sufficien to guarantee that the so-called SES are not man-made. Without an extended and thorough study ... [the VAN findings] are of dubious significance."

• ChouliarasStavrakakis99-eos80

"Support for VAN's earthquake predictions is based on false statements". Dispute claim of Uyeda [1998] that the strong earthquakes at Pirgos city on March 5 and 26, 1993, despite heavy damage, resulted in "no loss of life the citizens were prepared due to VAN predicytions." "... this statement is completely false." Following a panic on the 25th of Feb. Varotsos withdrew a prediction, saying "VAN has no indications of a forthcoming earthquake." Note that the IOA station is next to an antenna park, and the Pirgos station (where most of 1980s predictions were derived) was discovered to lie on a buried grounding grid used by a powerful radio transmitter. VAN "flatly refuse" to make raw data available.

• Papadopoulos10-eos91

A claim to have predicted the 14 Feb 2008 earthquake is "unjustified". "None of the claims for successful VAN predictions is justified."

• Hough 2010 (book) ‡

"Sensitive" sites makes hypothesis unfalsifiable.

"Proponents are quick to point to successes .... Closer scrutiny reveals a different story."

"... outright falsifications on the part of VAN proponents."

"As a prediction method, most seismologists consider VAN to have been resoundingly debunked."

‡ Indicates a non-primary source.

---

Tentative listing of main criticisms (extracted from above).

• Not a testable hypothesis. § WyssAllman96; Jackson96; RE96; KJ96; BernardEt97; JK98
• Major changes in hypothesis. § Wyss96
• Geophysically implausible, scientifically unsound. § Jackson96; Stiros97; JK98
• Likely man-made sources. § GruszowEt96; PhamEt98; PhamEt99; CS99
• Unspecified or indefinite parameters. § MG92; Wyss96; WyssAllmann96; MG96; Jackson96; JK98; SD96
• Retroactive adjustment of prediction rules ("rules of the game"). § MG92; Kagan96; KJ96
• Tendency to follow rather than precede EQs. § MG92; MG96; SD96;
• Misrepresentation of EQ record. § Wyss96
• Assertion of one-to-one relationship of SES & EQ demonstrably false. § Drak0Et93; Wyss96
• Lack of documentation, failure to share data. § JK98; CS99
• Failure to follow code of ethics for predictions. § DS96
• False claims and exaggeration. § CS99; Papa10
• Waste of money. § Stiros97

Daniel's comments

(What follows is comments from Daniel Helman) -- You have put in a huge amount of effort into this. It is wonderful! Here are some comments -- I'd like to assert that * Papadopoulos10-eos91 was written in bad faith. The text of the note is here: Uyeda and Kamogawa [2008] reported on the VAN experimental method for short-term earthquake prediction (VAN was named for three Greek physicists, Panayiotis Varotsos, Kessar Alexopoulos, and Konstantine Nomicos), which reportedly recorded seismic electric signals (SESs) before the Mw 6.8 earthquake on 14 February 2008. They claimed that a prediction is documented by P. A. Varotsos et al. (Seismic electric signals and 1/f “noise” in natural time, version 3, 2008; available at http://arxiv.org/abs/0711.3766v3) and in the newspaper Ethnos (10 February 2008; http://www.ethnos.gr/article.asp?catid=11424&subid2&tag=8777&pubid=444473). This claim is unjustified because the prediction was not submitted to Greece's Permanent Special Scientific Committee for the Assessment of Seismic Hazard and the Evaluation of Seismic Risk. According to Greek legislation, the committee is officially charged with analyzing and vetting earthquake hazard assessments (including predictions and ongoing seismic crises) and ultimately advising the government. Through civil protection authorities, the government handles the social, economic, and other negative consequences of impending earthquakes. But perhaps more scientifically grievous, this prediction was not documented elsewhere beforehand.

And a rebuttal is here: http://onlinelibrary.wiley.com/doi/10.1029/2010EO180004/abstract "We welcome the critical comments of Papadopoulos [2010]. We must point out, however, that most of them are incorrect.

First, on 1 February 2008, P. A. Varotsos et al. (Seismic electric signals and 1/f “noise” in natural time, version 3, 2008; available at http://arxiv.org/abs/0711.3766v3) did document the seismic electric signal (SES) activity for the M 6—class earthquake, recorded at station PIR (one of the SES measuring stations located close to Pirgos city in western Greece) on 14 January 2008, and assigned the epicentral area (approximately 250 × 250 kilometers). Two weeks after the documentation, the Mw 6.9 Greek earthquake in question occurred, on 14 February 2008. As to the occurrence time of this earthquake, the newspaper Ethnos on 10 February 2008 (http://www.ethnos.gr/article.asp?catid=11424&subid2&tag=8777&pubid=444473) reported it as “imminent.” Second, we did not mention two relatively small Patras shocks on 4 February 2008 simply because VAN also documented them (P. A. Varotsos et al., 2008, version 3) based on separate SES activity recorded on 10 January 2008 at Patras."

To continue: Papadopoulos ought to have known that Varotsos et al. are barred from reporting their results directly to Greece's Committee for the Assessment of Earthquake Hazard, and therefore his note is disingenuous. That's not to say that the updated VAN method has been confirmed -- just that Papadopoulos' note is particularly odd.

Most of the criticisms in the above list refer to the VAN method before time series analysis. I see the three outstanding criticisms of the VAN method as: (1) The origin of purported SES is hard to confirm. Many may be spurious. (2) The method has not been tested by scientists outside of Greece and Japan. Uyeda seems to be working closely with Vartosos et al. (3) False alarms endanger the public, and more care ought to be taken in issuing claims.

Let me run down the list of major criticisms, and add two cents to each:

• Not a testable hypothesis. § WyssAllman96; Jackson96; RE96; KJ96; BernardEt97; JK98

Earthquake prediction is empirical, and is based on positive results. If predictions are not successful, then the method is not valid. Therefore, I would not include this as a major criticism.

• Major changes in hypothesis. § Wyss96

Earthquake prediction is empirical, and is based on positive results. Empirical analysis often changes, as more data and more techniques are adopted. Therefore, I would not include this as a major criticism.

• Geophysically implausible, scientifically unsound. § Jackson96; Stiros97; JK98

Scientifically unsound I think is perhaps true -- in that the origins of purported SES have not been thoroughly cataloged, though the method of distinguishing SES from other signals has been stated clearly in the 2011 book. It remains to be seen whether artificial seismic events can be distinguished from electic signals of other origin, as test of their method. Geophysically implausible I would also include -- but a lot of this has focused on whether SES can in fact travel to the monitoring station, if the origin of SES are solid state (and not groundwater-related) phenomena. Implausibility does not preclude a thing, and this criticism ought to be included.

• Likely man-made sources. § GruszowEt96; PhamEt98; PhamEt99; CS99

Related to the origin of purported SES, and worth including.

• Unspecified or indefinite parameters. § MG92; Wyss96; WyssAllmann96; MG96; Jackson96; JK98; SD96

Earthquake prediction is empirical, and is based on positive results. Empirical analysis often changes, as more data and more techniques are adopted. Therefore, I would not include this as a major criticism.

• Retroactive adjustment of prediction rules ("rules of the game"). § MG92; Kagan96; KJ96

Earthquake prediction is empirical, and is based on positive results. Empirical analysis often changes, as more data and more techniques are adopted. Therefore, I would not include this as a major criticism.

• Tendency to follow rather than precede EQs. § MG92; MG96; SD96;

I would include this.

• Misrepresentation of EQ record. § Wyss96

I would include this.

• Assertion of one-to-one relationship of SES & EQ demonstrably false. § Drak0Et93; Wyss96

I would not include this -- as the VAN method no longer makes this claim.

They did make the claim, and its demonstrated falsity goes to their credibility. That they no longer make this claim raises the question of whether the method is thereby revised, which I address below where this point is repeated. ~ J. Johnson (JJ) (talk) 21:36, 26 June 2013 (UTC)

• Lack of documentation, failure to share data. § JK98; CS99

I would not include this. Data are available in arXiv, and in regular publications.

• Failure to follow code of ethics for predictions. § DS96

I would include this.

• False claims and exaggeration. § CS99; Papa10

I would not include this as a major criticism. I think this concept is covered better in the topic immediately above. This one is inflammatory.

• Waste of money. § Stiros97

I would not include this -- since the VAN method with time series analysis has not yet been shown to be false. Earthquake prediction, if successful, will not have been a waste of money.

In general, I think what distinguishes a "major" criticism from minor criticisms is its usefulness to scientific understanding. I would class the following as minor criticisms, as they are explained by the empirical nature of prediction:

• Major changes in hypothesis. § Wyss96
• Unspecified or indefinite parameters. § MG92; Wyss96; WyssAllmann96; MG96; Jackson96; JK98; SD96
• Retroactive adjustment of prediction rules ("rules of the game"). § MG92; Kagan96; KJ96

I would class the following as mean-spirited, and not even "minor" in nature, but, instead, untrue:

• Not a testable hypothesis. § WyssAllman96; Jackson96; RE96; KJ96; BernardEt97; JK98
• Lack of documentation, failure to share data. § JK98; CS99
• False claims and exaggeration. § CS99; Papa10
• Waste of money. § Stiros97

I would class the following as no longer true of the VAN method, since they revised it in 2001 to include time series analysis:

• Assertion of one-to-one relationship of SES & EQ demonstrably false. § Drak0Et93; Wyss96

I am especially concerned about this idea that because "the VAN method" was revised ~2001, all prior criticisms do not apply. If this was indeed the case, then we would need to distinguish "VAN method 1" from "VAN method 2" (or "VAN-NT method"?). If VAN-2 was totally different and independent from VAN-1 then that would be valid, but the same would apply to the claimed successes. And it would not be "30 years of [alleged] success", but 20 years for VAN-1, and ten or more for VAN-2.

In fact these are not separate and independent methods. Unless the revision addresses some prior criticism — remedies some defect — all the prior criticism still applies. And it remains a fact that VAN did claim a one-to-one relationship between SES and EQs, and that this is demonstrably false. This undermines the credibility of not just the VAN method, but the VAN team. I suspect this is why we do not see criticism of VAN-NT, and why (above) it "has not yet been shown to be false" — no one thinks it is worth bothering with. VAN-NT is based on VAN-1 (which is discredited), and the earlier criticism still applies.

Interesting points. My desire in working through this section was two-fold: (1) that the VAN method has been updated, and the section might be confusing to the reader, since it deals mostly with a version of the VAN method that hasn't been used for more than a decade; and (2) readability. I see that the earlier method was trying to make one-to-one (short term) predictions based on purported SES. A large electric signal was supposed to predict a large earthquake. I do not see this in their new method. The only thing in common is that they are using the same equipment to gather SES data. I see that the more recent method uses time series analysis on the group of signals in a region, looking for trends in timing and strength. The new method gathers data for a long period of time. There is no one-to-one correlation with the newer method. There is no prediction based on the SES, either, with the new method. The prediction in the new method is based on seismic data. The SES initiate the study of the seismic data, not electric data. Further, Springer just recently published a book of their updated method. It's not easy for a scientist to have a book published by Springer. They are extremely well-regarded. I'm not sure what leads you to think that a one-to-one correlation method would be the same as a time series analysis method. Could you comment some more on this? Thanks for all your efforts!Daniel Helman (talk) 01:25, 6 July 2013 (UTC)

  Glad to see you back.

  You touch a number of points; I hardly know where to start. Do keep in mind what I said above, that if the revised version ("VAN-2") is entirely separate, than the claim of "30 years of success" is false, as VAN-2 has not been around that long. Likewise, to the extent that VAN-2 does depend on detection of SES then it is subject to all the criticism of SES, including the unresolved issues of industrial noise, selectivity, generation, and transmission.

  But the concept of "VAN-2" breaks down, as there seem to be several sub-versions. E.g., in their more recent papers they seem to applying their "natural time" analysis to seismicity only, but in their earlier papers they base it on SES. So which is it? This is one of the main criticisms of "VAN" generally: lack of an explicit hypothesis.

  I point out that the "one-to-one correlation" of SES and EQ was an early claim, but that is not the method. That they now run their analysis through an additional epicycle of "natural time" does not fix any of the deficiencies. ~ J. Johnson (JJ) (talk) 22:46, 10 July 2013 (UTC)

Yes, what you are calling VAN-2 is entirely separate. It is the same as the difference between placing bets in blackjack based on what cards are dealt, versus placing bets in blackjack based on card counting. The former looks at a one-to-one correlation, while the latter looks at a time series of events. The early VAN method was based on looking for a one-to-one correlation between seismic electric signals and impeding seismic activity. The latter method uses SES as a medium range predictor, and then uses natural time analysis of the seismic (not electric) record starting from that critical point to make short term predictions. This new method has a benefit, in that regular seismic analysis has always failed, and so people are now in the practice of throwing out certain seismic events (such as foreshocks or aftershocks) or of picking a zero-point in time, to try for predictions based on seismic (not electric) events. Thus, the current VAN method is in line with the current work in earthquake prediction using seismic (not electric) signals. Yes, what you say is correct. "30 years of success" is false. 30 years of attempts is closer to the truth, and that is not a bad thing! I would like to emphasize that, though Varotsos' communications can be odd, the work of Uyeda in Japan is not. And these two are working together. Finally, the work is published by Springer, and reputable. The main thing, I think, for the criticisms is to do precisely what we are doing here. That is, (1) to be clear about which criticisms still apply to the updated method; and (2) to emphasize why the VAN method earned ill will from the Greek public historically. I don't find the criticisms of their earlier technique to be very useful for the reader. Look -- I'm not saying that the new method should be put on a pedestal. It's just basic science. They've come up with a new method, and it deserves to be evaluated. The readers of this page also deserve a clear description of the new method, and what its potential is. Thanks for your thoughts!Daniel Helman (talk) 05:41, 23 July 2013 (UTC)

Here's some more: Yes, what you write is true. The method based on time series analysis suffers from a few weaknesses:

(1) Though they outline a few methods for distinguishing artificial from genuine signals in the Springer book, these still need to be verified extensively, especially if the method is to be used in other metropolitan areas, like Los Angeles. (2) Selective sensitivity of stations is based on correlation of electric signals with seismic events. There are some examples from the book of artificial seismic events (blasting) producing electric signals. A standard protocol should be devised so that a specified number of artificial events can be used to constrain how likely the correlations of natural seismic events are with the purported SES. (3) Generation is still an issue, but newer papers are extant, such as those by Freund, and these look at mechanisms. I don't find a criticism saying "it's physically impossible" to be good science. One must fit the analysis to the observed data, and not the other way around. (4) Transmission is still an issue, but faults are much more electrically conductive than the surrounding rock, and articles have been published showing transmission is possible. Again, I don't find the criticism saying "it's physically impossible" to be good science. If one model is wrong, then one looks for another to explain the phenomena.

But regarding a "lack of specific hypothesis" -- I find what they are doing to be good science for empirical phenomena. They try a method, find out it doesn't work, and then refine it, to get a new method. This means that they must be clear in their claims (and they aren't so clear in separating out their various successes and failures!) Regards.Daniel Helman (talk) 06:04, 23 July 2013 (UTC)

There are a couple of points where I think you are a bit confused. E.g.: the original VAN "method" did not look for a "one-to-one correlation", it was based on a supposed prior observation that there is such a correlation. So look for the SES, and (allegedly) an EQ follows.

Another point: I think you are putting too much credence in the book. Springer sells books, peer-review is not required for publication. And if you look carefully you will see that MLV is not entirely candid about various matters. E.g., in the preface she says: "All evaluations to date of the VAN method from international committees and conferences have been positive." Which is patently false. Further on she says various journals have commented on their work, but does not mention the critical nature of the comments themselves. (Having not read the book itself, I can only wonder how she spins the chapter on "VAN evaluations".) This is the kind of stuff that should make you very wary.

I would also take issue that their new method deserves to be evaluated. No method "deserves" to be evaluated, unless it can be shown to offer some insight, such as accurately predicting earthquakes. The basis for investigating VAN is claimed success in prediction. But 1) this claim is dubious, and 2) the method allegedly used is incompletely described and keeps changing, which precludes proper evaluation. That no one has critically evaluated VAN-2 likely reflects scientific consensus that it does not deserve evaluation. Given that, it is improper for us as editors to invert that to say there is nothing against VAN-2. From a scientific viewpoint they have not proved their claims, so there is nothing to say for it.

One last point, regarding iterative science. You say they "try a method, find out it doesn't work, then refine it". For sure they have been refining their method, but seems they never acknowledge that a previous version was deficient. (This was someone else's comment, I forget who.) A naive reading gives the impression that ALL of their half-dozen or so variations work. ~ J. Johnson (JJ) (talk) 23:38, 25 July 2013 (UTC)

---

This leaves the following as major criticisms of the VAN method, with an emphasis on the origin and transmission of SES:

• Geophysically implausible, scientifically unsound. § Jackson96; Stiros97; JK98
• Likely man-made sources. § GruszowEt96; PhamEt98; PhamEt99; CS99
• Tendency to follow rather than precede EQs. § MG92; MG96; SD96;

And the following as major criticisms of the VAN method, with an emphasis on their methodology:

• Misrepresentation of EQ record. § Wyss96
• Failure to follow code of ethics for predictions. § DS96

Let me know what you think of what I've written, above? And do you agree that these are the general categories of the major criticisms of the VAN method? (1) The origin of purported SES is hard to confirm. Many may be spurious. (2) The method has not been tested by scientists outside of Greece and Japan. (3) False alarms endanger the public, and more care ought to be taken in issuing claims. Daniel Helman (talk) 21:20, 24 June 2013 (UTC)

It looks like you've also been chewing hard on this. Okay, this is going take a couple of nights to mull over. Well, one preliminary comment I think, which I'll add in a moment. ~ J. Johnson (JJ) (talk) 22:45, 25 June 2013 (UTC)

Daniel: I have taken the liberty of bolding your comments to better show where you disagree with the criticism. If you object please feel free to revert this edit. ~ J. Johnson (JJ) (talk) 21:48, 26 June 2013 (UTC)

Neutrality of recent edits questioned

I have tagged the article as non-neutral. A series of edits on 6 May (starting here) from an IP address geolocated to the University of Athens (home of VAN) generally blunted or removed criticism, while adding comments that these have been "refuted by VAN" (etc.). The general trend is to soften or remove criticism, and promote the views of VAN. I rolled back those edits, but they were largely restored by 77.69.86.91 (geolocated to an Athens suburb). The same IP did a similar series of edits about VAN at Earthquake prediction, which I have discussed in detail at Talk:Earthquake prediction#Re 195.134.89.153's VAN edits. (S/he also tagged the section on the VAN predictions with neutrality, but has not provided any discussion.)

Hard to say what should be done, as this article really needs a deep review, and revision. But no where near the top of my priorities. If anyone with some familiarity with the field (and access to a good university library) wants a challenging summer project, this would be a good time to speak up. ~ J. Johnson (JJ) (talk) 22:45, 22 May 2016 (UTC)

This is where someone should have protected the article from neutrality, and did not. We would better respect scientific publications, in a scientific article, and do not bring own conclusions inside or let them exist. Scientific criticism is answered with scientific publications, not in talk pages of our encyclopedia. Please point to the neutrality issues in the article, moving the tag there to protect the reader, or add scientific criticism that is not mentioned to balance the article.--77.69.86.91 (talk) 06:59, 23 May 2016 (UTC)