Here are the specs of a very interesting diamond:
Measurements 6.70 – 6.73 x 4.17 mm
Carat Weight 1.16 carat
Color Grade J
Clarity Grade SI1
Cut Grade Very Good
Depth 62.1 %
Table 57 %
Crown Angle 34.5°
Crown Height 14.5%
Pavilion Angle 41.0°
Pavilion Depth 43.0%
Star Length 55%
Lower Half 80%
Girdle Medium to Slightly Thick, Faceted, 4.5%
Polish Very Good
Clarity Characteristics Crystal, Feather
Clouds are not shown.
The question is: how can it be that a diamond is graded very good by the GIA yet seemingly has GIA excellent proportions?
First a word of caution, diamonds like this are more likely to be a cause for concern rather than a good deal. If it truly had excellent optics, a good vendor will price it accordingly. More often than not, the cut grade is reflecting something else that is otherwise not presented on the grading certificate.
At first, this diamond sounded like one that had possible painting or digging. But this is unlikely because such a diamond is usually accompanied with a comment that says, “Cut grade is based on the brillianteering of the half-facets”.
The girdle variation also does not vary so much to a point where I would be immediately concerned. So what could be the cause? Well we know that the GIA assesses cut using proportions but it places an arguably greater weight on the visual appearance of a diamond.
Light Performance Images
(Images are courtesy of Good Old Gold, used with permission.)
Looking at the H&A images, it is obvious that there is some twist and some variances in the lower girdles although it is important to note the camera-tilt in the hearts image. You can draw a line of symmetry at a diagonal, from the top-right corner to the bottom-left corner. Apart from relatively larger hotspots, I can’t really see what would justify the very good grade. I have certainly seem many GIA excellent cuts that had worse looking images.
On GIA’s Facetware Cut Estimator, these specs do give a very good result and if you change the girdle thickness to 4%, then it gives an excellent result. However, GIA allows slightly thick girdles to attain an excellent cut grade and this is why I ruled out the girdle.
Because of this, I did more research on the girdle thickness. It turns out that the initial grade of the girdle thickness is determined visually under 10x and the guidelines for a slightly thick girdle is that it appears “obvious at 10x”. It appears that it is possible for a larger diamond with a 4.5% girdle thickness to have a 4.5% girdle that is considered ‘thick’. The guidelines for a thick girdle is that it is “very obvious at 10x”. This means it is possible that this diamond’s cut grade came down to a value judgment of whether a 4.5% girdle in this case was obvious or very obvious. Now Facetware allows you to specify the girdle min and max so it is still not evident that the 4.5% girdle was the cause of the downgrade. It is useful to know that a thick girdle is an automatic downgrade to very good. The only possibility is that Facetware considers this diamond’s girdle thickness as borderline, and because girdle thickness is measured at the girdle valleys that there is a chance the girdle is ‘very obvious’ at the girdle bezels where the GIA tells us the girdle thickness is typically 1.7% thicker than at the valley. This means there is a possibility that the final cut grade was simply based on the Facetware result, which may or may not be the standard practice in borderline cases.
Also, you can note that on the certificate of this diamond, the total depth is reported as 62.1% so it is interesting to note that GIA’s Facetware estimates the total depth as 62.7%. The discrepancy is explained by the GIA because Facetware calculates the total depth using rounded numbers whereas the number reported on the certificate is based on measurement. However, because GIA allows up to 63% total depth to attain their excellent cut grade, the total depth here also shouldn’t be the issue.
So far, it does look like this is one of the best looking very good cuts I have seen.
The GIA talks about weight ratio here.
The equation to calculate the weight ratio is as follows:
Weight Ratio = Actual Weight / (Average Diameter^3 x 0.003592)
Please note that the GIA says that the weight ratio is not used as a separate parameter for cut grading although weight ratio is a very important part of cut grading. It is actually taken into account of using the other proportions in GIA’s Facetware Cut Estimator.
I calculated this for the diamond in question and it actually received a weight ratio of 1.067.
Here is a quote from the GIA article:
“Excellent proportion grades most commonly have Weight Ratio values between 1.02 and 1.05.”
This would indicate that the diamond in question is 1.6% shy of an excellent proportion grade.
However, GIA also states this:
“When the estimated Weight Ratio is between 1.08 and 1.09, the diamond’s rounded proportions are quite close to the border between Excellent and Very Good cut grades. For such cases, it would be prudent to use the Facetware program to check both the specific proportions of the diamond and the surrounding proportions. If a change of one unit to table percentage, crown angle, pavilion angle or average girdle thickness changes the grade, the diamond may be too close to that grade border for comfort. For most cases in which the Weight Ratio is 1.09, the diamond is likely to receive a Very Good cut grade.”
According to the above quote, it would seem like the upper-limit of the weight ratio is 1.08 rather than 1.05. Also, a single unit of girdle thickness would likely move this diamond into the excellent cut grade and the diamond is too close for comfort, wow!
An interesting point about this diamond is that if the cutter had reduced the girdle thickness by just 0.1%, then this diamond could have received a GIA excellent cut grade yet still main its carat weight!
Whether the weight ratio limit is 1.08 or 1.05 is unclear but it does seem like GIA’s Facetware uses 1.05. A 1.05 means that the diamond can weigh 5% more than the reference weight and this is the same standard that AGS adopts in their AGS PGS software.