Consumer advisory: GIA Cut Grade Rounding Problems

[He Scores]

Date: 2/8/2006 9:43:50 AM
Author: Serg

Date: 2/8/2006 8:58:58 AM
Author: He Scores



Ah ha! Now I see it. The 34.5 degree crown stone only has more mass since the bottom angles on the 33.0 crown stone were raised. THIS is where the mass is calculated differently...not because of the crown angles.

If everything were the same, the mass calculation would be reversed. Now I see it.

Also, IMHO, the 33 stone with the bottom angle over 41 would be a darker looking stone, and most in the trade would accept the notion that it was the ''poorer'' of the two cuts.

Did I finally get on the same page Sergey?


Bill

re:Also, IMHO, the 33 stone with the bottom angle over 41 would be a darker looking stone, and most in the trade would accept the notion that it was the ''poorer'' of the two cuts.

Diameter 5,16
Girdle valley 3 %
Girdle bonel 6%
Total Height 59,8%
Table 60%
Pavilion 41,23
Crown 32,53
Mass 0,5087
AGS grade 5

Such diamond is more bright than modern Tolkowsky cut.

Sergey...by what measure are you saying that this stone is more bright than a modern Tolkowsky cut?

Bill

 

[Garry H (Cut Nut)]

Dear Bill,

You asked by what measure would this shallow crown diamond be brighter?

I could list Q, brilliance and Fire charts from Moscow State Uni.
http://www.gemology.ru/cut/english/conferens-article/_5.htm

I could introduce you to HCA (Tools on the top bar > Cut Adviser)

And Bruce Harding's work that you can find by searching for his name here on PS - Beryl.

But it is probably more your taste to have US research from "institutions" (the chart came from someone Marty knows) for 60% table diamonds with GIA Excellent or AGS 0 proportions.

It is considered commmon knowledge that shallow crown slightly deeper pavilion diamonds are brighter with a little less fire. Tolkowsky wrote about it too.

 

[WinkHPD]

Date: 2/8/2006 1:03:15 AM
Author: Garry H (Cut Nut)
The shallower pavilion would not have made GIA Excellent or Very Good - so the cutter was possibly led astray by the Sarin GIA software that cost him $3,500 per machine plus about 20% per year licence fee.

I would call that value subtracting.

Gary, can you explain that to me. Are you saying GIA is charging $3,500 per Sarin machine for the software and another $700 (20%) per year just to use the software? Or is the 20% of something else?

Man, if I was a cutter and followed the software''s advice then got bad grades on my stones for doing so, I would be REALLY CRANKED!

Wink

 

[Garry H (Cut Nut)]

Date: 2/9/2006 1:28:44 PM
Author: Wink

Date: 2/8/2006 1:03:15 AM
Author: Garry H (Cut Nut)
The shallower pavilion would not have made GIA Excellent or Very Good - so the cutter was possibly led astray by the Sarin GIA software that cost him $3,500 per machine plus about 20% per year licence fee.

I would call that value subtracting.

Gary, can you explain that to me. Are you saying GIA is charging $3,500 per Sarin machine for the software and another $700 (20%) per year just to use the software? Or is the 20% of something else?

Man, if I was a cutter and followed the software''s advice then got bad grades on my stones for doing so, I would be REALLY CRANKED!

Wink

That is exactly right Wink.

I guess you would be feeling like a lot of cutters.
Remeber they can cut a stone for a small % of what GIA charge to grade the stone.

 

[valeria101]

Date: 2/9/2006 1:46:58 PM
Author: Garry H (Cut Nut)


Remeber they can cut a stone for a small % of what GIA charge to grade the stone.

Has this being going on for long? I sounds extremely ''natural'' that grading (let alone cut grading) and cutting would tag along gingerly as opposed to slicing competing shares of the same cake.

 

[He Scores]

Gary,
I appreciate your information. I''ve seen charts like that before and I tend to glance over them, but I won''t digress as to why.

Has anyone ever photographed a sampling of these stones?

I find it hard to believe that two stones would face similar, one being a Tolkowsky and the other with a 41.8 pav angle and a 31.5 crown, even though the chart says they''re the same.

..and like I''ve said before. Even if they do analyze identically, the trade will not pay a premium for non-tolkowsky stone with a shallow crown and deep pav.

I might be one of those stubborn dinosaurs but I''m a firm believer in market share.

Also, in the GIA first "Return of something study" they had a stone with a "high crown and virtual fish eye" that had a high rating, so I''m not saying that these stones don''t give a certain "look" via some sort of machine, however, there are other aspects of the cut that are more detracting than the stones overall appearance, and this distraction will lower the percieved value of the stone, IMHO. This applies to some of the premium stones cut today also.

This is off the original thread as someone pointed out, so I''ll just bow out and chat with you guys in Vegas.
Bill

 

[Garry H (Cut Nut)]

Its not my role to teach you about Cut, nor to help you get more nice looking diamond out of various shaped rough Bill.
No - they will not all look the same, and just why should it be that experianced trade people know what look everyione else likes?

But re the dinasuar issue and trade reluctance to change - there was a time when all good diamond had to have a small culet. Now no one cares.

 

[strmrdr]

Date: 2/9/2006 1:46:58 PM
Author: Garry H (Cut Nut)

Date: 2/9/2006 1:28:44 PM

Remeber they can cut a stone for a small % of what GIA charge to grade the stone.

So?
consumers are paying for it and they want them.
every time I hear someone in the biz complain bout cert costs what goes thru my head is...
wwwwwwwwwwwwwwaaaaaaaaaaahhhhhhhhh the labs are making money and im not keeping it all for myself.
wwwwwwwwwwwwwwwwaaaaaaaaaaaahhhhhhhhhhh
:}

 

[RockDoc]

Here''s where we were in 2002. Interesting how this has affected the discussions on this thread, and the mention about a future cut grade system from GIA.

Also commented on were the affects of minor facets, from years ago. Interesting to see how the labs have addressed what probably started here, and how we''ve split the hair a few more times.

https://www.pricescope.com/community/threads/every-facet-matters-tests-show.3720/=

Rockdoc

 

[WinkHPD]

Date: 2/10/2006 12:38:10 PM
Author: strmrdr

Date: 2/9/2006 1:46:58 PM

Author: Garry H (Cut Nut)

Date: 2/9/2006 1:28:44 PM


Remeber they can cut a stone for a small % of what GIA charge to grade the stone.

So?

consumers are paying for it and they want them.

every time I hear someone in the biz complain bout cert costs what goes thru my head is...

wwwwwwwwwwwwwwaaaaaaaaaaahhhhhhhhh the labs are making money and im not keeping it all for myself.

wwwwwwwwwwwwwwwwaaaaaaaaaaaahhhhhhhhhhh

:}

Storm, I think you missed the point of Gary''s post. I have no idea what it costs to cut a stone in India, but it costs more to cut it than to cert it for a stone like Paul cuts. I do know that for the cost of the Sarin software and the anual licensing fee, the software should not be telling the cutters to cut the stone in a manner that will yield a low grade when the stone is presented, this is patently unfair and a disservice to the cutters that the GIA is selling this software to.

GIA needs to go back and do their homework it looks to me...

Wink

 

[adamasgem]

Date: 2/11/2006 1:00:03 PM
Author: Wink


Storm, I think you missed the point of Gary''s post. I have no idea what it costs to cut a stone in India, but it costs more to cut it than to cert it for a stone like Paul cuts. I do know that for the cost of the Sarin software and the anual licensing fee, the software should not be telling the cutters to cut the stone in a manner that will yield a low grade when the stone is presented, this is patently unfair and a disservice to the cutters that the GIA is selling this software to.

GIA needs to go back and do their homework it looks to me...

Wink

Wink..
You can be sure that GIA has done their "homework", in that their legal department (which must be growing by leaps and bounds lately) has probably dotted every " i" and crossed every "T", in an effort to distance themselves from any liability for what they do in their contractual dealings with Sarin. You probably can''t even complain about the GIA/Sarin software.

 

[strmrdr]

Wink,
To understand my comment read this:
http://journal.pricescope.com/Articles/35/1/Why-do-we-need-diamond-grading-reports.aspx

I got his post 100% :}

It all goes back to the pricing pie.
The price of diamonds at all levels all the way back to the mine is set by dividing a chunk of the estimated selling price then dividing it up allowing each stage as much profit as the rough seller wants them to have.
The cutters would like to keep the chunck of the pie the labs are getting because the rough suppliers keep taking chunks out of their slice.

Thats the way I see it,
Diamonds are not sold on a cost plus basis they are sold on what the market will take divided between the players between the end customer and the ground the rough came from with the rough sellers setting the prices.
Consumers are willing to pay $25 more, way kewl says the store owner more profit, woops next order costs $24.50 more, cutter says way kewl I get $24.50 more, woops next order of rough costs $25 more.
Cutter says hey where is my $.50
Store owner says charge me the $.50 and ill take my business elsewhere,
rough supplier says your lucky I allow you any profit at all.
Answer: hey lets take the labs share of the pie and soft grade on our own certs.

net result consumers lose

Tell me where in there im wrong and it dont work that way?

 

[RockDoc]

Should lab grading be done by the seller?

I agree with Wink about the cutting prices. I've been quoted as high as $ 600.00 per carat for AGS 0 type cutting and that was a cost to the dealer, not a cost that a retail customer would pay.

For Eightstar recutting the price is significantly higher. Although the cost of things in India is far less than the US, it is understandable that diamonds can be polished for less cost. However, I wonder if these are for the high quality type cutting that is done to attain the AGS 0 type grading by AGS.

Paul Antwerp has gone to great pains to produce AGS 0 type stones. Others have as well. Perhaps in India they are investing in high tech equipment for cutting, and I guess it is possible to attain the high cutting standard there but not probable.

Sellers grading their own stones, is no better than the foxes guarding the chickens. Sure it is bad business to grade a stone higher than it is, but the incentive to jump in certain grades to the next higher ones, have finanancial incentives, that are just too tempting for the average person in the diamond business.

But back to the topic - GIA's cut grade system. It is obvious the approach GIA has taken is far less superior to that of AGS's approach. So this is sort of a no brainer.... just insist on AGS reports, since they do have some meaning, and GIA's are questionable at best.

Rockdoc

 

[Garry H (Cut Nut)]

Wink was correct - we lost the topic, which was that GIA is selling software that cutters should use to add value by having producing diamonds with the confidence they will get a given GIA cut grade.
They then send the stone to GIA and it comes back as Poor, where their software told them the stone was Excellent.
They knew the stone had been painted - perhaps they would have been happy with Very Good.

It now has GIA stuff engraved on the girdle so can not be sold without the cert. It is .50ct and will be under 1/2ct if that is polished off.

Had they sold it as most stones of that size are - without a cert - then they could have expected to get a higher price than they would now. But they sent it to GIA because they thought it would add value.

storm the other story is another story. You can have your opinion - but please do not mix the topics.

 

[WinkHPD]

Date: 2/11/2006 1:40:26 PM
Author: strmrdr

rough supplier says your lucky I allow you any profit at all.

Answer: hey lets take the labs share of the pie and soft grade on our own certs.


net result consumers lose


Tell me where in there im wrong and it dont work that way?

I agree that DeBeers is pretty efficient at keeping the lion''s share of the profits, but I do not see where Gary was advocating for not getting certs so that we could be the foxes guarding the henhouses. I have been using certs for more than 20 years because I can not depend on cutters and manufacturers to accurately grade their stones consistently. I am always amazed when my local clients do not seem to care about the certs, but no way could you sell stones on the internet without them.

Wink

 

[Garry H (Cut Nut)]

As a matter of policy i choose to have diamonds over .90ct H+ SI2+ sent to a lab.
About 1 in 6 or so 1/2ct stones I buy already has a cert and very often we forget to give them to the client who rarely asks anyway.
The report definetly adds to the stone cost and price. And it adds to the consumers confidence in only some cases in my business.

 

[valeria101]

If I remember well, the first time lab reports were deemed ''overrated'' around here was not long ago - in a recent addition to the Diamond Journal. It''s been a while!

Nice to see the followup.

May I play the ''conspiracy theorist'' just once and pick up Wink''s post, twist it around, blend in a bit from Garry''s article about how underpaid (great value-added) cutting is relative to (dodgy, unreliable) grading, and ask... if new grading technology would be anywhere thrown in.

It sounds like if Mr. Nobody (i.e. not a human operator) does the grading, whoever handles that could make the application worth as much as the human error on current lab reports.

Not to mention those could be ''certificates'' as opposed to the ''reports'' or... something like that. Pipe dream - of a non-smoker to boot.

 

[oldminer]

The automated grading of color, fluorescence and light behavior in the ImaGem system may well become an accepted world-wide standard in months or years to come. No device is presently more reliable, repeatable, accurate or scientific in reporting these results. What is lacking now is knowledge, in the main community of gemologists, diamond dealers and appraisers, of the repaeatability and accuracy of this system.

It takes time to get recognized and checked out. It also takes time to perfect technical systems to get them to the real "ready point". If people who want to know more would consider creating a blind test of color grading non-fluorescent and blue fluorescent diamonds, Imagem would like to be a particpant. Someone unbiased would be welcomed to run the test diamonds at Imagem.

What needs to be recognized is that the GIA itself and other labs that make a valid effort at GIA conformance, are only about 65% consistent in their own color grading using the GIA system. We know this by having a large selection of already graded GIA diamonds in the database and having compared them to Imagem''s digital measures of color which are substantially more repeatable than any human, visual grading.

It is our belief that once several users of Imagem are processing diamonds in better conformance to GIA standards than the GTL itself can do, a new era of better conformance to the existing color grading standard will arise. In regard to grading light behavior, we have sufficient data now to process round, princess, marquise and three similarly shaped, yet differently faceted branded cuts which are proprietary to a sight-holder and are being marketed in the USA for the past couple of years by one of the largest retail jewelry chains. In the coming weeks, a lab I am a partner in, DGLA, will be processing over 26,000 diamonds for light behavior analysis with Imagem. Several other sight-holders are lined up to test and deploy grading euqipment in-house or in conjunction with labs being set up now in Mumbai and Surat.

I''m glad to show anyone interested the latest output and product line. It has evolved substantially and rapidly since we introduced this lab system in January 2005 in Philadelphia. If you have not looked in the past couple of months, it has changed and improved a great deal thanks to some extensive and intensive testing and data building.

The building of performance models is very useful to cutters, but the measurement of performance directly on each diamond is of more importance to end-users. No small thing is skipped when one measures directly. There are much fewer arguments about accuracy, rounding, and subjectivity when numerical results of the measures are provided. One simply makes a choice based on performance numbers and their own taste in the look of the diamond. I don''t think all the discussion in this thread will lead to a better result. It seems that there are going to some elements that cannot be agreed upon when it comes to modelling.

 

[valeria101]

Date: 2/12/2006 2:07:41 PM
Author: oldminer


The automated grading of color, fluorescence and light behavior in the ImaGem system may well become an accepted world-wide standard in months or years to come. No device is presently more reliable, repeatable, accurate or scientific in reporting these results.

Out of curiosity... will these reports guarantee robustness of results upon re-grading, for example. Is it enough to call the resulting paper a 'certificate'?

And is there any way for these grades to be tested if someone wanted to (via an independent method) ?

In fact, I am trying to hint at some of the disclaimers that makes current 'paper' a matter of expert opinion rather than measurable certitude.

Hope it is OK to ask.

 

[adamasgem]

Date: 2/12/2006 2:07:41 PM
Author: oldminer
What needs to be recognized is that the GIA itself and other labs that make a valid effort at GIA conformance, are only about 65% consistent in their own color grading using the GIA system. We know this by having a large selection of already graded GIA diamonds in the database and having compared them to Imagem''s digital measures of color which are substantially more repeatable than any human, visual grading.

Dave:
Repeatability and accuracy are two different things. You can be 100% wrong all the time, or 100% right all the time and you are 100% repeatable. GIA, for example, developed their "Horizon database", in my opinion, so that they could in part, give the illusion of repeatability.

 

[adamasgem]

Date: 2/12/2006 2:07:41 PM
Author: oldminer

The building of performance models is very useful to cutters, but the measurement of performance directly on each diamond is of more importance to end-users. No small thing is skipped when one measures directly. There are much fewer arguments about accuracy, rounding, and subjectivity when numerical results of the measures are provided. One simply makes a choice based on performance numbers and their own taste in the look of the diamond. I don't think all the discussion in this thread will lead to a better result. It seems that there are going to some elements that cannot be agreed upon when it comes to modelling.

Dave: It depends on EXACTLY what you are measuring. A black box, with an undefined lighting environment, which may or may reflect "realistic" situations CANNOT be accepted as any sort of "standard", even if it is "repeatable". DO NOT discount the use of modeling.

As an example, the graph below represents the weighted light return metrics (cosine squared weighting) for some of the originally modeled (non absorbing) GIA stones, where the total angle of diffused lighting varies from 0 to 90 degrees from overhead. As you increase the total lighting angle you get less "discrimination" between stones, a smearing of differences in any "metric", but the question is is the "metric" valid for consumer or scientic study. What does it mean.. Some stones respond better to higher angle lighting than others can be inferred from these data, but the diffuse "envirionment" modeled is not realistic, is it necessarily.

You can design a lighting envirionment and metric to "prove" whatever you want to, make one class of stone "better" than another or not.

When you directly "measure" a stone's "performance", you throw into the mix, the "color" of the stone, as differential absorption rears it head, so that the "measure" on a "J" color stone cut exactly the same is going to be different than on a "D" color stone. You have to make the qualitative adjustment for color versus cut, as well as human visual perception through modeling and translate that to a realistic envirionment to be meaningful, otherwise it is just a number.

I, for one, do not blindly accept the mantra that so and so machine is best, especially when the "machine" is a "black" box, and the minions refuse to define the rationale behind the metric. Sorry, It is just more HYPE to me, until then.

 

[oldminer]

Marty, I know the difference, as you do, between repeatability, and accuracy. I am making no atempt to claim one without the other. BOTH are required. BOTH are accomplished. It is a matter of continued fine tuning that will be an ongoing project. Making correct color measures of the odd UV fluorescent colors is something to address. Another is the problem of grading diamonds with certain gemstone inclusions or feathers which interfere with light transmittance.

Just like in the early 1900''s when folks said cars will never catch on because tehy can''t jump a fence with one of them. Well, cars still don''t jump fences with cars, but somehow the horse has been replaced. Technology may not solve every outlying situation that happens with diamond grading, but I would suspect 98 percent or more of near colorless diamonds can be graded with automation. This represents a huge number of the total diamonds which get graded by labs. The odd stones, the other 1 or 2 percent may well need human grading for many years to come.

Valeria: The robustness of the system allows re-identification although it is not based on color recognition, although a part of a re-recognition might be color grade. I won''t get onto the argument about what constitutes certification as the term is tossed about to the extent now that what it actually means has been blurred. I will say that color grading and light behavior grading will soon be done accurately and repeatably to the extent you will be able to bank on the results. This will be an important change in how people trust grading results.

 

[Garry H (Cut Nut)]

Dave we have been waiting for answers and independent testing.
We have asked for STL files so that we can compare scanner accuracies.
It would be better to have this data than read claims.
So far there is a lot of rhetoric and little info / evidence.
The little info that has been shared and from my own personal experiance suggest to me Imagem has a ways to go. I hope you can get there because the goal is a worthy one.
But show first and tell later.

 

[strmrdr]

Date: 2/12/2006 4:20:52 PM
Author: adamasgem

Date: 2/12/2006 2:07:41 PM


I, for one, do not blindly accept the mantra that so and so machine is best, especially when the ''machine'' is a ''black'' box, and the minions refuse to define the rationale behind the metric. Sorry, It is just more HYPE to me, until then.

I agree with Marty.
No one machine so far and no one lighting environment can tell you how a diamond performes because they perform differently under different lighting.

However I think color grading is one area that is perfect for machine automation and have high hopes that it gets adopted.
Clarity to a lesser extent could be automated but im less sure of how well that will work but im willing to be convinced.

How well does the imagem machine do at at telling it cant color grade this properly and raising an alert?

 

[Rod]

Question from the consumer side. I still don''t understand much of what you experts are writing about. Since it is all pretty much directed at the GIA, I am curious to know if the GIA monitors this site and if what you experts are saying is being heard at the GIA and whether what you''re concerned about will be addressed by the GIA.

GIA.........Are you listening? Or is this entire thread merely academic?

I do appreciate having the opportunity to read your posts. I''ve learned so much from you guys and gals and it''s an honor to be able to even address such minds in the diamond world. It seems you have our best interests as consumers in mind. It also seems you have the interests of the hard working, and obviously underpaid diamond cutters in mind as well.

Thanks..........

 

[RockDoc]

Date: 2/12/2006 7:08:11 PM
Author: Rod
Question from the consumer side. I still don''t understand much of what you experts are writing about. Since it is all pretty much directed at the GIA, I am curious to know if the GIA monitors this site and if what you experts are saying is being heard at the GIA and whether what you''re concerned about will be addressed by the GIA.

GIA.........Are you listening? Or is this entire thread merely academic?

I do appreciate having the opportunity to read your posts. I''ve learned so much from you guys and gals and it''s an honor to be able to even address such minds in the diamond world. It seems you have our best interests as consumers in mind. It also seems you have the interests of the hard working, and obviously underpaid diamond cutters in mind as well.

Thanks..........

ROD

You can bet your "bippy" GIA monitors what is being written. Now whether they will react positively is the question.

Hopefully, they will before people find out how "unimportant" the cut grading is on their new reports, and they lose the standing of being the "world''s authority".

Watch Peter Yantzer - AGS''s lab directory on CNN tomorrow morning 845-10am

It will be interesting to hear consumer opinion of his appearance.

Maybe GIA will go on and explain the basis of their system...

Rockdoc

 

[pricescope]

Date: 2/12/2006 8:59:22 PM
Author: RockDoc


...Watch Peter Yantzer - AGS''s lab directory on CNN tomorrow morning 845-10am ...

Has anybody seen it?

 

[JohnQuixote]

Date: 2/13/2006 12:59:03 PM
Author: Pricescope

Date: 2/12/2006 8:59:22 PM
Author: RockDoc

...Watch Peter Yantzer - AGS''s lab directory on CNN tomorrow morning 845-10am ...

Has anybody seen it?

Peter did not appear on the show this AM. I heard from AGS that because of weather he never made it to NYC. No word on rescheduling.

 

[adamasgem]

Date: 2/12/2006 4:54:17 PM
Author: oldminer
Marty, I know the difference, as you do, between repeatability, and accuracy. I am making no atempt to claim one without the other. BOTH are required. BOTH are accomplished. It is a matter of continued fine tuning that will be an ongoing project. Making correct color measures of the odd UV fluorescent colors is something to address. Another is the problem of grading diamonds with certain gemstone inclusions or feathers which interfere with light transmittance.

Fluoresecence is a simple problem to address, YOU FILTER OUT THE UV, so that it doesn''t have an effect, the way it was originally taught...!!!!!!!! And WHO THE HELL CARES about whether the "grade" agrees with the NEW GIA color grades

see http://www.adamasgem.com/giafluor.html

Inclusions which effect color grading are well handled by the SAS2000 Spectrophotometer Analysis System, raising a red flag that something is amiss..

 

[Modified Brilliant]

Date: 2/13/2006 1:04:42 PM
Author: JohnQuixote

Date: 2/13/2006 12:59:03 PM
Author: Pricescope

Date: 2/12/2006 8:59:22 PM
Author: RockDoc

...Watch Peter Yantzer - AGS''s lab directory on CNN tomorrow morning 845-10am ...

Has anybody seen it?

Peter did not appear on the show this AM. I heard from AGS that because of weather he never made it to NYC. No word on rescheduling.

Nasty weather yesterday in NYC. Just returned from NYC about an hour ago. 27+ inches...a new record!

www.metrojewelryappraisers.com