AGS vs Sarin numbers (for HCA)

[diasurfer]

Here''s one for the cut nerds out there ... Garry? Rockdoc?

I used the "Search by Cut Quality" tool to locate a diamond I am interested in. It got a 1.5 Ex-Ex-Ex-VG Overall:Excellent on the Holloway Cut Advisor. I noticed that the angles listed on the Search tool to produce the 1.5 match those on the AGS cert for this diamond. (I confirmed by entering them from the certification into the HCA myself).

I then took a look at the Sarin report for this diamond. I entered the numbers listed on the Sarin report. For crown and pavilion angles, I entered the angles in the left hand column (the ones on the right are the max-min ranges) and for depth and table I used the angles in the right hand column (numbers in left are in millimeters). I used the same value of 0.4 for the culet.

The HCA produced 1.9 Ex-VG-VG-VG Overall:Excellent.

Both times the HCA indicated an Excellent cut and I think to conclude anything beyond that is pushing the limits of the HCA (instructions are to use it as a weed out tool, not as the final word). Still, I have read recommendations on this forum (probably posted by amateurs) something along the lines of "choose Stone A over Stone B because A has 3 Ex and 1 VG and Stone B has 2 Ex and 2 VG".

As a professional scientist, I am used to reporting numbers with confidence intervals based on various assumptions or empirical knowledge of intrument noise, sample size and independence, underlying probability distribution, etc etc. So my questions are:

1. does the AGS also use the Sarin to produce the measurements on their certification?

2. If not, which is more accurate, the Sarin report or the AGS numbers? why?

3. If AGS does use a Sarin, should I interpret the differences in my example above as due to random variability, most likely due to instrument noise?

4. And if so, is the difference between and 1.5 and a 1.9 on the HCA itself "in the noise" and therefore not statistically significant?

5. And if that is true, since PriceScope is dedicated to consumer knowledge, might it be a good idea to note for the consumer what constitutes a significant difference in the HCA and what does not?

thanks and aloha

 

[WorkingHardforSmallRewards]

If you do a search for this you will find alot of information. When I had a Sarin analysis run it came back only .1 different on the HCA but the crown height was like .7% off and a few other measurement numbers were off, thank god not the angles very much...somehow. Check it out,

my understanding from reading those threads (cut nut and others of well reknown participated in most of them) is that yes AGS does use a Sarin and that in the end the AGS numbers should be considered closer, however, measuring to within .0X places or .X degrees on an object so small already will inherently cause for inaccuracies. Somewhere along the way Gary and the others discuss how it needs to be improved yada yada and the final conclusion i took away was that probably neither of them were perfect. But of course the AGS is going to be the numbers to put the most faith in, especially without knowing any real information about the machine used. There are also discussions on different machines and their different abilities and software updates, etc.

Check it out--but its going to be a late night for both of us I think


Also remember that there is a lot of discussion--as I believe you alluded to--on how the HCA is a tool used to weed out bad performers, but not necessarily to be used to evaluate better diamonds within a certain range, especially the top range. And even with the Sarin analysis it meets the under 2 limit I have heard alot of the big names on PS put out.

 

[Regular Guy]

Date: 6/20/2007 12:29:24 AM
Author:diasurfer

Still, I have read recommendations on this forum (probably posted by amateurs

yo...that's me...

something along the lines of 'choose Stone A over Stone B because A has 3 Ex and 1 VG and Stone B has 2 Ex and 2 VG'.

I've said something like that in the past and have been called a weenie. I, like you, want to drive the HCA to perform, because of other constraints making this an attractive tool, when diamonds at a distance represent an excellent value. Unfortunately, it's for weeding out. Nevertheless...I join you in wanting it to do as much as it can...

As a professional scientist, I am used to reporting numbers with confidence intervals based on various assumptions or empirical knowledge of intrument noise, sample size and independence, underlying probability distribution, etc etc. So my questions are:

1. does the AGS also use the Sarin to produce the measurements on their certification?

Yes

2. If not, which is more accurate, the Sarin report or the AGS numbers? why?

3. If AGS does use a Sarin, should I interpret the differences in my example above as due to random variability, most likely due to instrument noise?

Bigger houses like WF specifically recommend to defer to the lab versus their own in house sarin (which is a high quality machine, vs lower quality ones with more predicted variability...presumably because labs are expected to either or both have even higher quality ones, or use more rigor in maintaining them...

4. And if so, is the difference between and 1.5 and a 1.9 on the HCA itself 'in the noise' and therefore not statistically significant?

Pass...though apply answer to #3 if appropriate...

5. And if that is true, since PriceScope is dedicated to consumer knowledge, might it be a good idea to note for the consumer what constitutes a significant difference in the HCA and what does not?

Well...the HCA author has done that...at least on the back of an envelope for us. For example...I think he'd say you could tell a 1 from a 3.

Problem is partly...HCA doesn't intent to map a diamond closely...and assumes a number of generalities for unmeasured attributes...which are now even more routinely scored by GIA and such...like for Star and pavilion facets. Garry H qualifies...saying these differentials are infrequently significant. Still, he'd prefer to trust a) a more close mapping, I think, to a specific diamond, as would be done by say AGS, vs b) the general outcome you'd expect from an HCA, such that c) I think blind, he'd prefer a so-called poorer performing diamond, like one getting even 3.5 recently reported here, and an AGS0, versus one not closely measured, but still getting 0 - 2 on the HCA. But...you'd be safer if he was saying that...

thanks and aloha

Ditto

 

[angeline]

This is so funny. (in a wry sort of way) I asked almost the exact same question some time ago. As a physicist by training I couldn't understand the method of reporting numbers by GIA, and by the Sarin machine. I don't think anyone really got what I was getting at so I left the thread alone after a while.


Anyway here you go, all 5 pages of it... Differences in AGS and Sarin

a

 

[strmrdr]

As far as the hca goes 1.5 and 1.9 are equivalent.
The hca was written before the lgf% was available to most consumers and the lgf% has a huge effect on the kind of light return a diamond has.
The hca is still valid for determining if the average crown angle works with the average pavilion angle but the rest of it has been passed by now that more information is readily available.
Use the numbers as a screening tool ignore the rest.

 

[diasurfer]

Wow.
thanks for posting that link angeline. I should have searched harder for it before posting myself. Took a while to get through that one. Instrument accuracy, rounding, ... and more.

I think you posed the question very well. And I think flopkins put it best in response to a JohnQuixote comment:

There is no doubt that there will be more accuracy in the future, but I wish there was some caveats in the reports, stating precisely what John outlined -
''Average CA'' = x.xx ± 0.x degrees

Perhaps this is too technical for most folks, but even a brief statment noting the fact that certain measurements are averages, and accurate to a certain degree, would be very useful in the interest of full disclosure!!! (For the techinically minded science person that I am!!)

Consider this post another vote for the AGS and GIA to do just that.

Now to figure out what lgf% is ...

 

[angeline]

Thanks diasurfer, it''s nice to know someone understands where I was coming from. I thought the sneering comments on how some people demand accuracy that doesn''t exist were the worst. After all I''m not demanding accuracy where it doesn''t exist, just accurate reporting to the limits of the machine used is all.

Cheers,

a

 

[strmrdr]

Date: 6/20/2007 4:15:16 PM
Author: diasurfer
Now to figure out what lgf% is ...

http://www.goodoldgold.com/Articles/MinorFacets/

Basically the minors fine tune the light return.
The cant make a silk purse out of a sows ear but in some cases can improve or hurt the performance depending on the c/p relationship.

 

[diasurfer]

Well storm that link is going to take awhile to get through.

Did read a long thread from January discussing the whether 75% is "too short" of an lgf, or not, or whether that''s the whole story, and what the B-scope prefers etc etc.

My stone of interest has 77% ... sounds like that is an acceptable mid-range place to be.

thanks to all who have contributed to the thread.

aloha