2006 GIA grading report - Post info here please

[belle]

and for the presbyopic

 

[michaelgem]

Belle,

The diagram from the new GIA grading report that you posted clearly shows on the left side that the mains are not on edge, as they would be if the girdle measurement was at the peak rather than the valley. They are making it clear that the measurement is at the valley between adjacent mains, as it has always been.

I asked GIA and they responded along the lines of:

What part of:

"It is important to note that there will be no change in the GIA Laboratory’s methods for assessing culet size and girdle thickness due to the introduction of the new cut grading system."

do you not understand?

So my current belief appears accurate that the Facetware input of thin, medium, thick etc. is still based upon evaluation at the valleys in the usual manner.

Michael Cowing

www.acagemlab.com

 

[Garry H (Cut Nut)]

Date: 1/7/2006 7:55:58 PM
Author: michaelgem

What part of:

''It is important to note that there will be no change in the GIA Laboratory’s methods for assessing culet size and girdle thickness due to the introduction of the new cut grading system.''

do you not understand?

Michael Cowing

Michael what I do not understand is that you seem to be having a fight over this.
I am not.
I just find it interesting that GIA has chosen to accept girdle thickness entry of mesaurements at the mains and not the valleys on its Facetware online grading software. (This stone gets 4%).

Hmmm, interesting - 0.37mm - would this actually rate as Very Thick according to Marty''s old theory?

 

[Garry H (Cut Nut)]

Here is the chart from http://www.gis.net/~adamas/cut.html#girdle
You can see that this stone according to the GIA Facetware should have a girdle thickness between Very Thick and Extremely Thick.
(I know that GIA is using very large steps in rounding - but ....???)

 

[adamasgem]

I have been prepareing a litlle commentary on GIA FARCEWARE(TM).. but I want to warn those out there that G-D, aka GIA saw fit to include a little notice, that if you use FARCEWARE(TM), you are legally "bound" not to discredit it or say anything disparaging, otherwise the legal team out at GIA ill have fits...

“PUBLICITY. Licensee will not issue any press release or make any statement or announcement to the press, the public or any third party (including, without limitation, Licensee’s customers) which (i) reflects unfavorably on the Software or the performance of the Software, (ii) is false or misleading about GIA or the Software or (iii) is damaging to the reputation of GIA or the Software. This paragraph shall survive the termination of this Agreement.”


So I won''t be plugging any numbers into their program, but will comment freely on the loosening of any standards GIA once had...

More on next screen, everone might fing interesting, and is sure to tick someone off..

 

[adamasgem]

Regarding FARCEWARE(TM)..

One of the thing I first noticed in what they publicly published regarding thier new cut grade was that an extremely thin girdle (i.e. knife edge) could get a VERY GOOD cut grade, so I did a little searching and came across this page from their internal lab manual of 1998 (which I just happen to have had a copy of). People should find it interesting reading.

Now based on what I had previously documented about girdle thickness issues on my web site I came up with (based on GIA published information) the boundaries (Measured at the minimums of the scallops) in the mid 90''s of

Thin: Less than 0.15 millimeters
Medium: between 0.15 and 0.20 millimeters
Slightly Thick: between 0.20 and 0.23 millimeters
Thick: between 0.23 and 0.33 millimeters
Very Thick: between 0.33 and 0.40 millimeters
Extremely Thick: greater than 0.40 millimeters

By 1998 they had internally stretched the grading of girdle thickness, based on the attached such that, for a 1 carat stone, it appears that>>>

Very Thin: (New
Thin or less: Less than 0.15 millimeters (New 0.05mm-0.1mm)
To be called thin or less the stone’s girdle is allowed to be much thinner than before
Medium: between 0.15 and 0.20 millimeters (New 0.1mm-0.15mm)
To be called medium, the stone’s girdle is allowed to be much thinner than before
Slightly Thick: between 0.20 and 0.23 millimeters (New 0.15mm-0.2mm)
To be called slightly thick, the stone’s girdle is allowed to be much thinner than before
Thick: between 0.23 and 0.33 millimeters (New 0.2mm-0.3mm)
The thick range is the same width, but can be thinner than before
Very Thick: between 0.33 and 0.40 millimeters (New 0.3mm-0.5mm)
The very thick range has widened and can be much thicker than before
Extremely Thick: greater than 0.40 millimeters (New >0.5mm)
To be called Extremely Thick, the girdle can be 25% thicker than before

Now what they are doing internally in 2006, is anyones'' guess, but my intuition tells mes that a Jeep Wheel will get a medium girdle, just so that their paper reads right.. BUT THAT IS ONLY MY OPINION

More on FARCEWARE(TM) later

 

[JohnQuixote]

Date: 1/8/2006 12:19:55 AM
Author: adamasgem

“PUBLICITY. Licensee will not issue any press release or make any statement or announcement to the press, the public or any third party (including, without limitation, Licensee’s customers) which (i) reflects unfavorably on the Software or the performance of the Software, (ii) is false or misleading about GIA or the Software or (iii) is damaging to the reputation of GIA or the Software. This paragraph shall survive the termination of this Agreement.”

Hey Marty - that paragraph has been removed from the Facetware terms of use, so I believe people are at liberty to use it AND discuss it critically now.

Now what they are doing internally in 2006, is anyones'' guess, but my intuition tells mes that a Jeep Wheel will get a medium girdle, just so that their paper reads right.. BUT THAT IS ONLY MY OPINION

Ah but remember, FW will be judging how significantly the Jeep was painted. A nice two-tone might be fine, but big racing stripes or flames could result in a downgrade.

(a little humor)

 

[Serg]

Michael ,

re:What part of:

"It is important to note that there will be no change in the GIA Laboratory’s methods for assessing culet size and girdle thickness due to the introduction of the new cut grading system."

do you not understand?


Why are sure what methods "methods for assessing culet size and girdle thickness" are same for GTL, GIA education, GIA articles? Some times GIA rules are different for different applications. GIA is very big organization.

 

[adamasgem]

Date: 1/8/2006 1:07:31 AM
Author: JohnQuixote

Date: 1/8/2006 12:19:55 AM
Author: adamasgem

“PUBLICITY. Licensee will not issue any press release or make any statement or announcement to the press, the public or any third party (including, without limitation, Licensee’s customers) which (i) reflects unfavorably on the Software or the performance of the Software, (ii) is false or misleading about GIA or the Software or (iii) is damaging to the reputation of GIA or the Software. This paragraph shall survive the termination of this Agreement.”

Hey Marty - that paragraph has been removed from the Facetware terms of use, so I believe people are at liberty to use it AND discuss it critically now.


Well, I guess my prior commentary a while ago regarding that on another trade commentary forum did some good then

I wonder if they took my suggestion and fired the idiot who thought it up, as well as the the idiots who approved it..

Now what they are doing internally in 2006, is anyones' guess, but my intuition tells mes that a Jeep Wheel will get a medium girdle, just so that their paper reads right.. BUT THAT IS ONLY MY OPINION

Ah but remember, FW will be judging how significantly the Jeep was painted. A nice two-tone might be fine, but big racing stripes or flames could result in a downgrade.

(a little humor)

Not really humor, don't you realize that FARCEWARE was probably intended to make the broadest range of stones appear good on paper, otherwise you can sell your paper.. You develop the answer you want to give, and then tailor the test conditions to give you that answer proving the point you "scientifically" wanted to "prove", that's not to say expressly that that is what happened, but all we see everywhere we look is an opening up of every seemingly "standard" we were ever taught.....

 

[Garry H (Cut Nut)]

Marty I just toned down your comments a little

Date: 1/8/2006 3:48:01 AM
Author: adamasgem

Not really humor, don''t you realize that FARCEWARE was probably intended to make the broadest range of stones appear good on paper, otherwise you can sell your paper.. You develop the answer you want to give, and then tailor the test conditions to give you that answer proving the point you ''scientifically'' wanted to ''prove'', that''s not to say expressly that that is what happened, but all we see everywhere we look is an opening up of every seemingly ''standard'' we were ever taught.....

If that was their goal, then having the light come from one area above the stone, with relative darkness avbove and behind the observer would certainly make most shallower, tending to be lower yeilding, stones look pretty bad.
Also the use of a light colored stone tray makes leakage look neuatral rather than dark as it would appear most of the time to most diamond observers.

 

[JohnQuixote]

Date: 1/8/2006 3:48:01 AM
Author: adamasgem

Date: 1/8/2006 1:07:31 AM
Author: JohnQuixote

Date: 1/8/2006 12:19:55 AM
Author: adamasgem

“PUBLICITY. Licensee will not issue any press release or make any statement or announcement to the press, the public or any third party (including, without limitation, Licensee’s customers) which (i) reflects unfavorably on the Software or the performance of the Software, (ii) is false or misleading about GIA or the Software or (iii) is damaging to the reputation of GIA or the Software. This paragraph shall survive the termination of this Agreement.”

Hey Marty - that paragraph has been removed from the Facetware terms of use, so I believe people are at liberty to use it AND discuss it critically now.


Well, I guess my prior commentary a while ago regarding that on another trade commentary forum did some good then

I wonder if they took my suggestion and fired the idiot who thought it up, as well as the the idiots who approved it..

Now what they are doing internally in 2006, is anyones' guess, but my intuition tells mes that a Jeep Wheel will get a medium girdle, just so that their paper reads right.. BUT THAT IS ONLY MY OPINION

Ah but remember, FW will be judging how significantly the Jeep was painted. A nice two-tone might be fine, but big racing stripes or flames could result in a downgrade.

(a little humor)

Not really humor, don't you realize that FARCEWARE was probably intended to make the broadest range of stones appear good on paper, otherwise you can sell your paper.. You develop the answer you want to give, and then tailor the test conditions to give you that answer proving the point you 'scientifically' wanted to 'prove', that's not to say expressly that that is what happened, but all we see everywhere we look is an opening up of every seemingly 'standard' we were ever taught.....

It think it was prudent of them to remove the 'Publicity' statement.

Their new system appears to allow a much more liberal range of diamonds into the top grade than the AGS system does, but I don't believe that's a surprise.

I am surprised at the rounding in FW. 5% increments would seem to make the inclusion of stars/lgf a token gesture. Moreover, PA in increments of .2 & CA in increments of .5 can result in wholly different judgments, depending on what direction you take the rounding if an angle is in the middle. Perhaps they will be making revisions (?) In any event - whether you are a fan of FW or not - there is a necessary adjustment for depth when you 'round up' (or down) PA/CA numbers falling in the middle which FW does not mention.

BTW, my painting gaffe was in reference to this paragraph from FWs Additional Assumptions:

"The actual overall cut grade received from GIA for a diamond may differ from that predicted by the Software for the following reasons: ...(2) The diamond has been manufactured using non-standard brillianteering such that the diamond is significantly “painted” or “dug-out” (i.e., the diamond has its upper- and/or lower-half facets fashioned at tilt angles different than those traditionally used for standard round brilliant diamonds)... "

It will be interesting to see what 'significant' means, as we all know that these brillianteering methods can be responsibly used to improve light return and face-up appearance.

 

[michaelgem]

I don''t buy in to all this unproductive vituperation, but I am in concert with and have made related observations to these remarks of Garry''s:

" having the light come from one area above the stone, with relative darkness above and behind the observer would certainly make most shallower, tending to be lower yielding, stones look pretty bad.
Also the use of a light colored stone tray makes leakage look neutral rather than dark as it would appear most of the time to most diamond observers."

I believe that the lighting and grey stone tray were attempts to obtain standard lighting conditions close to typical observation conditions, so that results would better correlate to perceived reality. I don''t believe it was designed to produce some desired result.

For one thing, GIA has explained that a black tray overemphasized the negative appearance due to light leakage, while a white tray disguised it. The grey tray was thus an attempt to obtain better correlation with perception.

The standard lighting chosen, as Garry points out, would show the poor performance of shallow cut diamonds due to retro-reflection from the vicinity of the observer’s head. It must have made clear the fallacy in their erroneous belief stated in their original 1998 research report that I discussed in the Journal of Gemmology e.g.:

http://www.acagemlab.com/Article1/Article1.htm#brill024

"In general, WLR increases as crown angle decreases. ... These results suggest that, at the reference proportions, a diamond with a 23° crown angle is brighter than a stone with any other crown angle greater than 10°. ... Ironically, the highest WLR values are obtained for a round brilliant with no crown at all" GIA

Michael Cowing

www.acagemlab.com

 

[JohnQuixote]

Michael,

To be sure, I certainly hope you don't think I was vituperating without merit... You know I try to be productive when I get my 'tuper on.

To that end, I think you would agree that the rounding in FW creates dichotomy when there is no accompanying instruction to resolve numbers falling inside the increments, wouldn't you?

 

[adamasgem]

I am wondering if the NEW GIA/GTL reports giving the ''measurements" and angles of of the stone, mimic the FARCEWARE(TM) rounding "standards" (or lack thereof), ie

1) Measure the stone using a Helium/Sarin/Ogi or whatever
2) Average the measurements
3) Re Round the averaged measurements to numbers of convenience..


If so, I ask the question, would these GIAGTL "reports" constitute a FRAUD on the consumer ?, being intentionally misleading, so much so that the numbers would then be useless if the consumer wanted to use GIA "measurements", to get another "opinion".

The example of GIA/GTL paper posted prior on this thread seems to suggest that this is what is happening.. But, again, GIA legal has probably tried to cover this up with the "fine print" on the back of the report.. I''d appreciate if someone with a new report would scan the "fine print" on the back...

 

[JohnQuixote]

Date: 1/8/2006 5:12:37 PM
Author: adamasgem
I am wondering if the NEW GIA/GTL reports giving the 'measurements' and angles of of the stone, mimic the FARCEWARE(TM) rounding 'standards' (or lack thereof), ie

1) Measure the stone using a Helium/Sarin/Ogi or whatever
2) Average the measurements
3) Re Round the averaged measurements to numbers of convenience..


If so, I ask the question, would these GIAGTL 'reports' constitute a FRAUD on the consumer ?, being intentionally misleading, so much so that the numbers would then be useless if the consumer wanted to use GIA 'measurements', to get another 'opinion'.

The example of GIA/GTL paper posted prior on this thread seems to suggest that this is what is happening.. But, again, GIA legal has probably tried to cover this up with the 'fine print' on the back of the report.. I'd appreciate if someone with a new report would scan the 'fine print' on the back...

Marty, I wonder this too.

We, and other vendors, provide a Sarin scan for every diamond independently of its lab report. We have always maintained that the lab report is the document of authority (in the case where there may be a slight disparity between our scan and the lab’s).

If the new GIA report includes rounded proportions it is - at the least - a weaker technical document than a Sarin scan and - at the most - misleading to any who don’t realize the implications. Perhaps that’s not the case, and the numbers on the report are actual Sarin numbers (?)

I think fraud is a strong word but I agree, if this is what is being done, there should be obvious notation to that effect. Maybe the ‘fine print’ you're hoping for will be in a large font, bold and highlighted.

 

[strmrdr]

Garry the official lighting envirement box is here.

http://www.gia.edu/geminstrument/

 

[adamasgem]

Date: 1/8/2006 5:53:57 PM
Author: JohnQuixote

Date: 1/8/2006 5:12:37 PM
Author: adamasgem
I am wondering if the NEW GIA/GTL reports giving the ''measurements'' and angles of of the stone, mimic the FARCEWARE(TM) rounding ''standards'' (or lack thereof), ie

1) Measure the stone using a Helium/Sarin/Ogi or whatever
2) Average the measurements
3) Re Round the averaged measurements to numbers of convenience..


If so, I ask the question, would these GIAGTL ''reports'' constitute a FRAUD on the consumer ?, being intentionally misleading, so much so that the numbers would then be useless if the consumer wanted to use GIA ''measurements'', to get another ''opinion''.

The example of GIA/GTL paper posted prior on this thread seems to suggest that this is what is happening.. But, again, GIA legal has probably tried to cover this up with the ''fine print'' on the back of the report.. I''d appreciate if someone with a new report would scan the ''fine print'' on the back...

Marty, I wonder this too.

We, and other vendors, provide a Sarin scan for every diamonds independently of lits ab report. We have always maintained that the lab report is the document of authority (in the case where there may be a slight disparity between our scan and the lab’s).

If the new GIA report includes rounded proportions it is - at the least - a weaker technical document than a Sarin scan and - at the most - misleading to any who don’t realize the implications. Perhaps that’s not the case, and the numbers on the report are actual Sarin numbers (?)

I think fraud is a strong word but I agree, if this is what is being done, there should be obvious notation to that effect. Maybe the ‘fine print’ you''re hoping for will be in a large font, bold and highlighted.

RE: Maybe the ‘fine print’ you''re hoping for will be in a large font, bold and highlighted.


IS THIS SIZE TYPE APPROPRIATE ?
I THINK IT MIGHT WELL BE!

 

[belle]

the ''fine'' print

 

[adamasgem]

Date: 1/8/2006 6:07:10 PM
Author: belle
the ''fine'' print

Thanks Belle..

I guess the fine print says it all, by saying nothing other than: don''t trust our reports or anything on them

It is a shame, that with advances in measurement technology, like the Helium, that GIA and the internal management powers to be, apparently chose to resort to the protractor and yardstick to measure what the diamond is about.

All they would have to do, as an educational institution, is state the computed measurement variance (or sigma) about the measured mean values, but no, they know better.. let''s just fib a little about the numbers ....

 

[michaelgem]

John,

RE your comments:

>>> I am surprised at the rounding in FW. 5% increments would seem to make the inclusion of stars/lgf a token gesture. Moreover, PA in increments of .2 & CA in increments of .5 can result in wholly different judgments, depending on what direction you take the rounding if an angle is in the middle.

I think you would agree that the rounding in FW creates dichotomy when there is no accompanying instruction to resolve numbers falling inside the increments, wouldn''t you? <<<

I don''t see your comments as vituperative. Your point deserves addressing. IMO, considering the importance of lower half length to the diamond''s optical performance and thus its beauty, this parameter should be measured and specified more accurately than rounding to 5% increments. GIA, on the other hand, seems to find lower half length less critical than you and I, perhaps more a matter of taste. This is reflected in this seemingly coarse quantization.

We may find our concerns already addressed, but likely not to your satisfaction, on GIA''s website. I seem to remember reading or being told that the choice of increments for each parameter related to what I think of as just noticeable differences (JND''s) obtained by statistical analysis of their 70+ thousand paired comparison database. It was also said to relate to an analysis of the accuracy obtainable in practice for each parameter during laboratory measurement using the Sarin machines.

I have made inquiries, and I''ll let you know if I hear or find anything more definitive.

Michael Cowing

www.acagemlab.com

 

[Garry H (Cut Nut)]

Date: 1/8/2006 11:52:01 PM
Author: michaelgem

...considering the importance of lower half length to the diamond''s optical performance and thus its beauty, this parameter should be measured and specified more accurately than rounding to 5% increments. GIA, on the other hand, seems to find lower half length less critical than you and I, perhaps more a matter of taste. This is reflected in this seemingly coarse quantization.

We may find our concerns already addressed, but likely not to your satisfaction, on GIA''s website. I seem to remember reading or being told that the choice of increments for each parameter related to what I think of as just noticeable differences (JND''s) obtained by statistical analysis of their 70+ thousand paired comparison database. It was also said to relate to an analysis of the accuracy obtainable in practice for each parameter during laboratory measurement using the Sarin machines.

Michael Cowing

www.acagemlab.com

Has anyone bookmarked the link to the rapaort article that gIA wrote? - They discussed their reasons (excuses) for such wide tolerances.
As i remeber they have said that they want appraisers to be able to guestimate and use the system. Just as weak an arguement me thinks.

 

[Garry H (Cut Nut)]

Date: 1/8/2006 12:43:48 AM
Author: adamasgem

Regarding FARCEWARE(TM)..

One of the thing I first noticed in what they publicly published regarding thier new cut grade was that an extremely thin girdle (i.e. knife edge) could get a VERY GOOD cut grade, so I did a little searching and came across this page from their internal lab manual of 1998 (which I just happen to have had a copy of). People should find it interesting reading.

Now based on what I had previously documented about girdle thickness issues on my web site I came up with (based on GIA published information) the boundaries (Measured at the minimums of the scallops) in the mid 90''s of

Thin: Less than 0.15 millimeters
Medium: between 0.15 and 0.20 millimeters
Slightly Thick: between 0.20 and 0.23 millimeters
Thick: between 0.23 and 0.33 millimeters
Very Thick: between 0.33 and 0.40 millimeters
Extremely Thick: greater than 0.40 millimeters

By 1998 they had internally stretched the grading of girdle thickness, based on the attached such that, for a 1 carat stone, it appears that>>>

Very Thin: (New <0.05mm)

Thin or less: Less than 0.15 millimeters (New 0.05mm-0.1mm)
To be called thin or less the stone’s girdle is allowed to be much thinner than before

Medium: between 0.15 and 0.20 millimeters (New 0.1mm-0.15mm)
To be called medium, the stone’s girdle is allowed to be much thinner than before

Slightly Thick: between 0.20 and 0.23 millimeters (New 0.15mm-0.2mm)
To be called slightly thick, the stone’s girdle is allowed to be much thinner than before

Thick: between 0.23 and 0.33 millimeters (New 0.2mm-0.3mm)
The thick range is the same width, but can be thinner than before

Very Thick: between 0.33 and 0.40 millimeters (New 0.3mm-0.5mm)
The very thick range has widened and can be much thicker than before

Extremely Thick: greater than 0.40 millimeters (New >0.5mm)
To be called Extremely Thick, the girdle can be 25% thicker than before


Now what they are doing internally in 2006, is anyones'' guess, but my intuition tells mes that a Jeep Wheel will get a medium girdle, just so that their paper reads right.. BUT THAT IS ONLY MY OPINION

More on FARCEWARE(TM) later

I just opened the attachment on Marty''s article - interesting.
Do you think you could make some new lines on this old chart of yours Marty showing where you see the changes and their scale? Perhaps enlarge it a bit?

 

[adamasgem]

I just opened the attachment on Marty''s article - interesting.
Do you think you could make some new lines on this old chart of yours Marty showing where you see the changes and their scale? Perhaps enlarge it a bit?
The Holloway Cut Adviser & Ideal-Scope Cut Crusader!
www.ideal-scope.com
www.preciousmetals.com.au


Hi Garry..
That is on my agenda to do, and program into my SAS2000 all the various GIA definitions I''ve come across and can substantiate via writen documentation, obtained in various manners..

Getting this information out of the GIA black hole is sometimes difficult and out of date...

what they publish internal to the lab is not what they always teach and publish externally.. And it probably changes at the whim of management. Take for example the Rapaport article, where they say that girdle thickness is a visual estimate, which is inconsistent with the older internal lab manual.. what the heck they are actually doing is anyone''s guess.

It seems a shame that Millions of dollars of industry monies were spent to generate a technically corrupt presentation in a report. They give the excuse that, in effect they have to take into account that not all users of the system have accurate measurement devices, so they go to the lowest common denominator to create a "system", and seemingly report the LOWEST ACCURACY NUMBERS they can spit out.... AND THIS FROM AN EDUCATIONAL INSTITUTION that should strive and promote the HIGHEST ACCURACY represention of the stone, something that the makers of the instrumentation try to do..

Yah, some of the instrumentation accuracy may be suspect and manufacturers of instrumentation vary (maybe substantially) in their accuracy, but shouldn''t that be handled and treated just as it is.. If you have "perfect" measurements, shouldn''t you use them?
The powers in charge of FARCEWARE(TM) and lab reports don''t seem to think so instead they fudge the report; if you think there is an inaccuracy in your measurement, put the inaccuracy where it belongs, as a stated or computed variance (or one signa) in the measurement..PERIOD That is the way it is supposed to be handled..

I''m going to program FARCEWARE(TM) rounding into my 3D raytrace and see what the fudged results give..

 

[valeria101]

Date: 1/8/2006 11:52:01 PM
Author: michaelgem

I seem to remember reading or being told that the choice of increments for each parameter related to what I think of as just noticeable differences (JND''s) obtained by statistical analysis of their 70+ thousand paired comparison database.

Is GIA''s the first cut grading system that acknowledges any sort of rounding and error level in their measurements? JND?

 

[smarty33]

From my experiences so far with mounted stones, a microscope and a table gauge - 5% is about as close as you can get.. particularly for Star Length.

I''ve been seeing a few old style reports with an insert with all the new info''. I have just spoken to GIA and got the lowdown on these:

Any reports on Round Brilliants issued between August 1st 2005 and December 31st 2005 can be sent to Carlsbad and a new report with cut grade etc will be issued free of charge. There is no need to see the diamond again.

 

[adamasgem]

Date: 1/9/2006 4:46:39 AM
Author: valeria101

Date: 1/8/2006 11:52:01 PM
Author: michaelgem


I seem to remember reading or being told that the choice of increments for each parameter related to what I think of as just noticeable differences (JND''s) obtained by statistical analysis of their 70+ thousand paired comparison database.

Is GIA''s the first cut grading system that acknowledges any sort of rounding and error level in their measurements? JND?

Valeria..

The answer is NO.
First of all, any report IMPLIES a basic level of accuarcy for the numbers presented, in the QUANTIZATION of mean values given, the average, unfortunately NONE present the standard deviation about the mean.
That implicaton of basic QUANTIZATION is given by the least significant figure given , in other words, if I say $1 it means somthing different than $1.00 . That is basic grammer school ROUNDING to the least significant bit presented in the decimal or metric system.

When we know the numbers are averages, we also used to KNOW (or thought we knew) the average has been computed and rounded to the least significant bit given, but we don''t know the uncertainty about that mean value, but we know that there is some value, but we imply that the the AVERAGE has been computed to the least significant bit given

GIA, in their infinant wisdom, has apparenly chosen to change all that.. By the way for those who don''t know, GIA uses the international (Diamond Dealers Club) standard for rounding WEIGHT without telling any one really. A stone weighing 0.998 carats would be called 0.99 and one weighing 0.999 would be called 1.00 carats.. Not the US standard where a 0.996 carat could be legitimately called 1.00. I guess that is why AGS gives the weight to three decimal places, and lets the buyer or seller make the decision.

Now we have the FARCEWARE(TM) standards, which are inconsistent with any normal thought pattern, where different rounding standards apply to each number. It is one thing to use the actual numbers and apply them in a multidimentional database which is internally quantized to various levels, but it is another thing entirely to present a number which is arbitrarily quantized or rounded and anything but factual, and present it as factual.

It is sort of like the phone company which will charge you for 2 minutes if your call goes over 1 minute by one millisecond, and another charges you to the nearest tenth of a second. You can make a decision who to by from because their policies are stated when you purchase. GIA''s paper apparently doesn''t state what their "standards", so the consumer doesn''t have a clue.

I don''t care what they say, what they have apparently done on their NEW paper is another bit of consumer misinformation.

 

[valeria101]

Thanks for the detailed story...

After reading this, it is quite clear that my previous 2-word post mixed up issues of statistical significance and the gap between statistical and practical significance.

In fact, I was mostly referring to the last sort.

 

[oldminer]

I have long supported measurements as a way to screen diamonds as an aid in making a smart final choice. My first AGA cut grade parameter charts were published in 1987. Today, we are faced with many more measurements that may be taken. This larger number of measures now serves to inhibit their use in a screening process unless one has a computer and a rather complex program which predicts light behavior. The measuring of tiny facet lengths makes for great discussion, but unless each facet length and angle is accurately measured, the diamond may perform differently than a simple average of them all would predict, even when plugged into the computer and complex predictive programs. Rounding off makes it worse yet. Ugh!

One would think that by now more folks would begin to understand that you can make the screening process endlessly complex and challenging without adding accuracy to the end result. You will ultimately fail to predict the outcome of how light interacts with a particular diamond by measuring facets and angles. Sure, you''ll get some good information that may assist a consumer in screening potential diamonds, but the real answer will remain as elusive as measuring each facet and angle with accuracy.

A final "grade" should be "MEASURED" not predicted. You won''t get to where you want to go by choosing the wrong fork in the road. Look back over your shoulder and notice the other direction that the fork in the road took. You have not gone that far down the wrong path that you can''t readily go back and consider direct assessment as the more useful and most honest grading technology.

Consumers want to believe that experts always employ accurate grading techniques. We all seem to be well aware of the shortcomings of the Sarin, Ogi or the GIA system in grading as well as in disclosure and accepting responsibility. For those dealers interested in this alternative direction, I''d be pleased to assist you with the proper technical contacts and/or demonstrations.

 

[valeria101]

Date: 1/9/2006 1:42:42 PM
Author: oldminer

You won''t get to where you want to go by choosing the wrong fork in the road. ... You have not gone that far down the wrong path that you can''t readily go back and consider direct assessment as the more useful and most honest grading technology.

It must be that your question has been directed to experts, not me. Would it be ok to clarify what the either/or fork is about? I have some trouble understanding why metrics and models are mutually exclusive.

Here''s why...

Unless there is a single physical metric sufficient to describe quality, isn''t a model needed to indicate what exactly should be measured?



For example:

As far as I know, only one diamond brand is still measuring performance by a single physical metric (total light return) for its sales pitch: the Context and Spirit cuts. Both suffer from the fault that GIA''s study refers to when saying that the best light return comes from a diamond without a table. Sure enough, the choice of this unique metric no particular model of diamond cut to back it up - the same machine measures reflectivity of anything else in the same way.


Sure enough, I am not trying to take or argue for either branch in the road - just wondering whether they are convergent or divergent

 

[oldminer]

Yes, I suppose I am speaking to the "experts", but not on a high technological plane. It is one of logic and a rather simple proposition.

One fork of the road is choosing to make as many physical measures of a diamond as one may wish to do and then go into a program to predict how well that stone handles light. The accuracy of the final grade predicted will depend on how accurately each parameter is measured. It will also depend on measuring all or as many parameters that really matter to the predicted outcome. If one collects inaccurate measurements, and/or insufficient measurements can you hope for a correct prediction? No. If one rounds off, will that result in accuracy? No. If the tools used for measurement are inherently inaccurate, will the prediction be accurate? No.

The other fork of the road uses high grade measuring of the light performance itself. If light is what we want to know about, then why would one attempt to "predict" it, if it can be directly "measured" repeatably, accurately and scientifically?

Diamonds which appear to be cut identically are not identical. They do not handle light identically if they are cut even a tiny bit differently. If one diamond has an inclusion in the center, it will handle light differently than the exact same cut stone with no inclusion in the center. This just makes common sense. You will never predict light performance based solely on physical measures with accuracy as you will have neglected both inclusions and transparency. Every diamond is different, no matter how identically they might be cut.

I hope that makes the points and simplifies what the message is.