RE: [ai-geostats] Geostats Scam?
Here we go again list! Welcome to the world of Jan Merks, "Mach" Perry Collier Senior Geologist Rio Tinto Technical Services Phone: +61 7 3327 7676 Mobile: 0408 015 837 -Original Message- From: Mach Nife [mailto:[EMAIL PROTECTED] Sent: Wednesday, 8 February 2006 4:59 AM To: ai-geostats@unil.ch Subject: [ai-geostats] Geostats Scam? Hi, This guy (Jan W Merks) seems to devote his life trying to prove that Geostatistics is a scam... If it's true I'm gonna start looking for alternatives. http://www.geostatscam.com Anyone has a point of vue on this? machnife __ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
[ai-geostats] Re: Geostats Scam?
Hi Mach (?) Jan Merks is a consultant with a formidable and well-earned reputation in sampling theory and applications. I understand he sits on many committees in Canada which define standards for sampling and evaluation. The first of his "geostatistics scam" articles appeared under the title "Geostatistics or voodoo statistics" in 1992, in just about every mineral industry publication from the Engineering & Mining Journal to the Northern Miner newspaper. In this article he references Michel David's book (1978) and my 1979 book Practical Geostatistics. In December 1993, after considerable pressure from my colleagues and many more publications of this article, I wrote a personal letter to Dr Merks inviting him to come down to Reno, Nevada and put his thoughts at a short course I was about to teach. I have now put both of these faxes up on the Web at http://www.kriging.com/correspondence so that you can judge for yourselves what the position was then. Please forgive the quality of the reproduction as thermal fax paper tends to fade with time! Since then, Dr Merk's has made it his life's work to visit every site possible (e.g. Amazon) and post negative reviews and comments about geostatistics. His comments are coherent and persuasive and have influenced many people, like yourself, against this whole field. His premise is that statistical theory does not apply to auto-correlated or spatially related variables. This will come as an unpleasant surprise to all statisticans involved in the study of stochastic variables including such authors as Sir David Cox (Emeritus Professor of Statistics, Imperial College London and author of the basic textbook in stochastic processes), Noel Cressie (Director of Spatial Statistics at Ohio State University) and even Brian Ripley (Professor of Statistics at Oxford University) -- not to mention Roger Mead (now retired, formerly Head of Department of Applied Statistics at Reading University, England) who taught me spatial statistics in 1969 before I'd ever heard of geostatistics. Make no mistake, there are flaws in geostatistics both as a theory and in application. There is plenty of room for improvement across the spectrum and (hopefully) people around the world are working on this as we read. There are also advances in other approaches to spatial estimation which I (for one) watch with interest in the anticipation of new tools for the real world. Criticism should be seen as a good thing and an aid to development. However, relentless negativity serves no-one on either side of this non-discussion. An exchange of ideas is more profitable than an endless stream of insults on either side. In the meantime? Learn what you can and judge for yourself whether the ideas of geostatistics make sense in practice and could be applicable to your own problems. Isobel Clark http://www.kriging.com/courses * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
RE: [ai-geostats] Geostats Scam?
machnife, This has been discussed. You can search the archives to read the discussion. My best guess is that people posting regularly on AI-Geostats do not agree that Geostatistics is a scam :-). You can also read what A. Journel had to say about the argument in material Merks has posted on his website (at least you used to be able to, I haven't checked recently). Paul Walline NOAA Fisheries -Original Message- From: Mach Nife [mailto:[EMAIL PROTECTED] Sent: Tuesday, February 07, 2006 10:59 AM To: ai-geostats@unil.ch Subject: [ai-geostats] Geostats Scam? Hi, This guy (Jan W Merks) seems to devote his life trying to prove that Geostatistics is a scam... If it's true I'm gonna start looking for alternatives. http://www.geostatscam.com Anyone has a point of vue on this? machnife __ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
[ai-geostats] Geostats Scam?
Hi, This guy (Jan W Merks) seems to devote his life trying to prove that Geostatistics is a scam... If it's true I'm gonna start looking for alternatives. http://www.geostatscam.com Anyone has a point of vue on this? machnife __ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
RE: [ai-geostats] Sill not reaching 1
Colin, I suggest you don't have to complicate yourself. If using alternative a), do not touch the nugget; and there;s no need for "scaling" anything. The zonal anis. takes the form of an additional structure, valid only in the direction you're interested in, as you point out with the same range as your last structure, but with 0 variance contribution in the other main directions of anisotropy. It's a simple additional structure that gets added to the end for that particular structure. If this is confusing, think of how you would model a hole-effect in one direction (pseudo-sedimentary or strata-bound deposits, with sequences of mineralized/unmineralized material, or perhaps coal seams). If option b), should not use the zonal-turned-geometric anisotropy to determine your search ranges! At most, use the other directions as guide, and don't worry about never reaching the range (or in fact the variance) when you search for data in that particular direction. In fact, I'm a subscriber of limted search radii, even if less than the variogram ranges; a bit more conditional bias (what I call a healthy dose), just enough to get better local accuracy and better prediction of the ex-mine material, something not always or easily accepted. This of course depends a lot on the characteristics and amount of data you may have. Hope this helps, Cheers, Mario Colin Badenhorst <[EMAIL PROTECTED]> wrote:Hi All, Thanks for the many replies thus far. Just to make it clear: the data has not been normalised. The samples (from drill core) have simply been composited to equal support lengths. Sorry about the confusion between gamma and rho, but I think you know what I mean the left hand axis of the variogram graph. Theoretically, nugget + sill = sample variance, except in the absence of trend or drift. Mario seems to have seen exactly what the problem is. There is indeed zonal anisotrophy in one particular direction, for the variogram under question. I have considered both Marios suggestions: 1. Adding the last structures variance component, keeping the range the same. In fact, I wanted to add the component to the nugget effect (a sort of conservative approach by allocating more of a random component to grade variability, if that makes sense), but I was advised against this. It was suggested I try scaling, which is where I am coming unstuck. 2. Manipulating the 2nd structure, but then ending up with a massive range, which as Mario has correctly predicted, raises eyebrows. I also run the risk of negative weights because of the very large search area. The intermediate direction is rather poor too, so this is perhaps not an option. I hope that explains the problem better. Regards, Colin From: Mario Rossi [mailto:[EMAIL PROTECTED] Sent: 07 February 2006 10:51To: Colin BadenhorstSubject: Re: [ai-geostats] Sill not reaching 1 Colin,what may be happening is that you have a strong zonal anisotropy, I'm assuming you're talking about a specific direction.You probably know that the zonal can be either modeled as:a) an additional nested structure in the specific direction where it occurs, just add the variance component, same range; orb) convert the zonal into a geometric anisotropy, taking the last nested structure asymptotically until it reaches 1.0 at a very long distance in that direction. Easier done than explained, it may raise eyebrows in people that don't understand or have a geostats background when you try to explain 2-3 5km ranges in a 500m long deposit, for example... Also, need to be careful with intermediate directions, and make sure they fit well. Hope this helps, Cheers, MarioColin Badenhorst <[EMAIL PROTECTED]> wrote: Hi All, I am performing estimation for Ni as a penalty element to our main ore, using nested spherical correlogram models. One of these shows abundant and robust experimental data reaching a plateau at gamma = 0.90, and is thus modelled in such a way that the sill of this spherical model is at 0.9, instead of at 1.00. Im not sure of this, but is there a strict requirement that the sill differentials and nugget need to sum to 1.00 for estimation purposes. I use Vulcan (effectively GSLIB) for estimation. If it is a requirement that they add to 1.00, it has been suggested that I could scale the nugget and sill to a value of 1, but Im not sure exactly how to do this. Does anyone have any comments and/or suggestions? Regards,Colin This e-mail and its attachments, is confidential and is intended for the addressee(s) only. If you are not the intended recipient, disclosure, distribution or any action taken in reliance on it is prohibited and may be unlawful. Please note that any information expressed in this message or its attachments is no
RE: [ai-geostats] Sill not reaching 1
Hi All, Thanks for the many replies thus far. Just to make it clear: the data has not been normalised. The samples (from drill core) have simply been composited to equal support lengths. Sorry about the confusion between gamma and rho, but I think you know what I mean – the left hand axis of the variogram graph. Theoretically, nugget + sill = sample variance, except in the absence of trend or drift. Mario seems to have seen exactly what the problem is. There is indeed zonal anisotrophy in one particular direction, for the variogram under question. I have considered both Mario’s suggestions: 1. Adding the last structures variance component, keeping the range the same. In fact, I wanted to add the component to the nugget effect (a sort of conservative approach by allocating more of a random component to grade variability, if that makes sense), but I was advised against this. It was suggested I try ‘scaling’, which is where I am coming unstuck. 2. Manipulating the 2nd structure, but then ending up with a massive range, which as Mario has correctly predicted, raises eyebrows. I also run the risk of negative weights because of the very large search area. The intermediate direction is rather poor too, so this is perhaps not an option. I hope that explains the problem better. Regards, Colin From: Mario Rossi [mailto:[EMAIL PROTECTED] Sent: 07 February 2006 10:51 To: Colin Badenhorst Subject: Re: [ai-geostats] Sill not reaching 1 Colin, what may be happening is that you have a strong zonal anisotropy, I'm assuming you're talking about a specific direction. You probably know that the zonal can be either modeled as: a) an additional nested structure in the specific direction where it occurs, just add the variance component, same range; or b) convert the zonal into a geometric anisotropy, taking the last nested structure asymptotically until it reaches 1.0 at a very long distance in that direction. Easier done than explained, it may raise eyebrows in people that don't understand or have a geostats background when you try to explain 2-3 5km ranges in a 500m long deposit, for example... Also, need to be careful with intermediate directions, and make sure they fit well. Hope this helps, Cheers, Mario Colin Badenhorst <[EMAIL PROTECTED]> wrote: Hi All, I am performing estimation for Ni as a penalty element to our main ore, using nested spherical correlogram models. One of these shows abundant and robust experimental data reaching a plateau at gamma = 0.90, and is thus modelled in such a way that the sill of this spherical model is at 0.9, instead of at 1.00. I’m not sure of this, but is there a strict requirement that the sill differentials and nugget need to sum to 1.00 for estimation purposes. I use Vulcan (effectively GSLIB) for estimation. If it is a requirement that they add to 1.00, it has been suggested that I could scale the nugget and sill to a value of 1, but I’m not sure exactly how to do this. Does anyone have any comments and/or suggestions? Regards, Colin This e-mail and its attachments, is confidential and is intended for the addressee(s) only. If you are not the intended recipient, disclosure, distribution or any action taken in reliance on it is prohibited and may be unlawful. Please note that any information expressed in this message or its attachments is not given or endorsed by Lisheen Mine unless otherwise indicated by an authorised representative independently of this message. Lisheen Mine does not accept responsibility for the contents of this message and although it has been scanned for viruses Lisheen Mine will not accept responsibility for any damage caused as a result of a virus being passed on. * * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats Mario E. Rossi GeoSystems International Ph: 561-495-8797 Fax: 561-498-1262 Relax. Yahoo! Mail virus scanning helps detect nasty viruses! This e-mail and its attachments, is confidential and is intended for the addressee(s) only. If you are not the intended recipient, disclosure, distribution or any action taken in reliance on it is prohibited and may be unlawful. Please note that any information expressed in this message or its attachments is not given or endorsed by Lisheen
Re: [ai-geostats] Sill not reaching 1
yes- but shouldn't use 'gamma' to mean correlation, it confuses me. 'rho' is the usual symbol. In which case, the correlogram shouldn't 'reach a plateau at gamma = 0.9' Would expect correlation (and covariance) to decrease with lag. Colin- we need a clearer description of your problem! Dan - Original Message - From: Maria N. Morales Boezio To: Dan Bebber ; Colin Badenhorst ; ai-geostats@unil.ch Sent: Tuesday, February 07, 2006 10:05 AM Subject: Re: [ai-geostats] Sill not reaching 1 When working with correlograms the maximum theoretical sill is 1 with no need of previously normalizing data. Maria - Original Message - From: Dan Bebber To: Colin Badenhorst ; ai-geostats@unil.ch Sent: Tuesday, February 07, 2006 7:34 AM Subject: Re: [ai-geostats] Sill not reaching 1 nugget + sill = sample variance unless you have normalized the data prior to analysis. So you could get any value from raw data. Dan Bebber - Original Message - From: Colin Badenhorst To: ai-geostats@unil.ch Sent: Tuesday, February 07, 2006 9:17 AM Subject: [ai-geostats] Sill not reaching 1 Hi All, I am performing estimation for Ni as a penalty element to our main ore, using nested spherical correlogram models. One of these shows abundant and robust experimental data reaching a plateau at gamma = 0.90, and is thus modelled in such a way that the sill of this spherical model is at 0.9, instead of at 1.00. Im not sure of this, but is there a strict requirement that the sill differentials and nugget need to sum to 1.00 for estimation purposes. I use Vulcan (effectively GSLIB) for estimation. If it is a requirement that they add to 1.00, it has been suggested that I could scale the nugget and sill to a value of 1, but Im not sure exactly how to do this. Does anyone have any comments and/or suggestions? Regards, ColinThis e-mail and its attachments, is confidential and is intended for the addressee(s) only. If you are not the intended recipient, disclosure, distribution or any action taken in reliance on it is prohibited and may be unlawful. Please note that any information expressed in this message or its attachments is not given or endorsed by Lisheen Mine unless otherwise indicated by an authorised representative independently of this message. Lisheen Mine does not accept responsibility for the contents of this message and although it has been scanned for viruses Lisheen Mine will not accept responsibility for any damage caused as a result of a virus being passed on.* * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm )* To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED]Signoff ai-geostats * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm )* To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED]Signoff ai-geostats * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
[ai-geostats] geostatistics and geology
Dear list, I just want to know if anybody could hint me with some references or books regarding the relationship between geostatistics and geology (of ore deposit). Furthermore, I want to know explicitly how geostatistics can explain the geology of ore deposit, for example about: coal depositional, sedimentation process in sedimentary deposit, geologic model of hydrothermal deposit, and so on. At this moment I am dealing with the geometrical and quality data of coal deposit, but I have difficulty to connect the geostatistics to the coal depositional process. I appreciate very much your helps and suggestions. Regards, M. Nur Heriawan - Graduate School of Science and Technology Kumamoto University Kurokami 2-39-1, Kumamoto 860-8555, JAPAN URL: http://www.mining.itb.ac.id/heriawan __ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
Re: [ai-geostats] Sill not reaching 1
nugget + sill = sample variance unless you have normalized the data prior to analysis. So you could get any value from raw data. Dan Bebber - Original Message - From: Colin Badenhorst To: ai-geostats@unil.ch Sent: Tuesday, February 07, 2006 9:17 AM Subject: [ai-geostats] Sill not reaching 1 Hi All, I am performing estimation for Ni as a penalty element to our main ore, using nested spherical correlogram models. One of these shows abundant and robust experimental data reaching a plateau at gamma = 0.90, and is thus modelled in such a way that the sill of this spherical model is at 0.9, instead of at 1.00. Im not sure of this, but is there a strict requirement that the sill differentials and nugget need to sum to 1.00 for estimation purposes. I use Vulcan (effectively GSLIB) for estimation. If it is a requirement that they add to 1.00, it has been suggested that I could scale the nugget and sill to a value of 1, but Im not sure exactly how to do this. Does anyone have any comments and/or suggestions? Regards, ColinThis e-mail and its attachments, is confidential and is intended for the addressee(s) only. If you are not the intended recipient, disclosure, distribution or any action taken in reliance on it is prohibited and may be unlawful. Please note that any information expressed in this message or its attachments is not given or endorsed by Lisheen Mine unless otherwise indicated by an authorised representative independently of this message. Lisheen Mine does not accept responsibility for the contents of this message and although it has been scanned for viruses Lisheen Mine will not accept responsibility for any damage caused as a result of a virus being passed on.* * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm )* To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED]Signoff ai-geostats * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats
[ai-geostats] Sill not reaching 1
Hi All, I am performing estimation for Ni as a penalty element to our main ore, using nested spherical correlogram models. One of these shows abundant and robust experimental data reaching a plateau at gamma = 0.90, and is thus modelled in such a way that the sill of this spherical model is at 0.9, instead of at 1.00. I’m not sure of this, but is there a strict requirement that the sill differentials and nugget need to sum to 1.00 for estimation purposes. I use Vulcan (effectively GSLIB) for estimation. If it is a requirement that they add to 1.00, it has been suggested that I could scale the nugget and sill to a value of 1, but I’m not sure exactly how to do this. Does anyone have any comments and/or suggestions? Regards, Colin This e-mail and its attachments, is confidential and is intended for the addressee(s) only. If you are not the intended recipient, disclosure, distribution or any action taken in reliance on it is prohibited and may be unlawful. Please note that any information expressed in this message or its attachments is not given or endorsed by Lisheen Mine unless otherwise indicated by an authorised representative independently of this message. Lisheen Mine does not accept responsibility for the contents of this message and although it has been scanned for viruses Lisheen Mine will not accept responsibility for any damage caused as a result of a virus being passed on. * * By using the ai-geostats mailing list you agree to follow its rules ( see http://www.ai-geostats.org/help_ai-geostats.htm ) * To unsubscribe to ai-geostats, send the following in the subject or in the body (plain text format) of an email message to [EMAIL PROTECTED] Signoff ai-geostats