Hello all
 
I recently asked a selected bunch of geostatistical 'heavyweights' about the fundamental gripe Jan Merks has about the field of geostatistics in an attempt to kill off a small seed of doubt that was growing in my primitive geostatistical brain.  For those who have pondered the same thing, I forward Isobel Clark's response (with her permission).
 
Cheers
Perry
 

Perry Collier

Senior Mine Geologist
Ernest Henry Mine  
Xstrata Copper Australia
( (07) 4769 4527
4  (07) 4769 4555
* [EMAIL PROTECTED]
" http://www.xstrata.com
 
PO Box 527
Cloncurry QLD 4824
Australia
 
"Light travels faster than sound. That is why some people appear bright until you hear them speak"

 
-----Original Message-----
From: Isobel Clark [mailto:[EMAIL PROTECTED]
Sent: Saturday, 11 June 2005 8:49 AM
To: Collier, Perry (Ernest Henry - Copper)
Subject: Re: JW Merks - help requested to kill off seed of doubt!

Hi Perry
 
I remember vividly walking into the ice breaker at the 1992 Geostats Congress in Portugal and having several people turn to me and say "what did you DO to this guy?". I have never met Jan Merks, although I have invited him (politely) to talk.
 
Merks is hung up on the seeming disparity between standard (frequentist) statistics and geostatistics. The two main examples he gives in his seminal article are:
 
(1) geostatisticians divide the variance by N when everyone knows that the variance should be divided by N-1
 
(2) the kriging system is basically just a standard regression system and hence should have effectively zero degrees of freedom.
 
The point he seems to have missed in (2) is quite simple: geostatistics is a two stage process. The rules of statistical regression do not apply because the 'covariances' (semi-variogram) calculated are based on the complete data set. Noel Cressie says the total degrees of freedom are N(N-1) and started to froth when I asked him this question - and he isn't a GEOstatistician but is a top flight statistician.
 
In (1) his point is even more fundamental. Classical variances are divided by N-1 because we estimate the mean from the data as well as the variance. Semi-variograms are divided by N because we don't -- we assume the mean difference is zero.
 
This is all explained at great length in Practical Geostatistics 2000, which we sent to him 5 years ago. Of course, you do have to read it to find that out.
 
Jan Merks is a brilliant man and has written some excellent stuff. However, he is hoist by his own petard in that he also can ignore facts which conflict with his own view of the world.
 
I am sure Peter will set you straight on this and wish you all good luck in your studies.
 
Isobel


**********************************************************************
The information contained in this e-mail is confidential and is
intended only for the use of the addressee(s).
If you receive this e-mail in error, any use, distribution or
copying of this e-mail is not permitted. You are requested to
forward unwanted e-mail and address any problems to the
Xstrata Queensland Support Centre.
Support Centre e-mail: [EMAIL PROTECTED]
Support Centre phone: Australia 1800 500 646
International +61 2 9034 3710
**********************************************************************
* 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

Reply via email to