Big difference in HCA reading for similar stones

Garry Holloway FGAA DipDT

HCA and Ideal-scope developer

http://www.ideal-scope.com and
http://www.HollowayDiamonds.com.au

Bill Leiberum
1948-2007
Thanks for everything and Rest In Peace,Bill.

Ana "The greatest experts are only as good as the sum total of what they have seen." [Souren Melikian]

Ira (Ruffles have Ridges) Z.
____________

If you're looking for diamond shopping tips...this (along with this update) might do the trick!

however...caution: known to have been regarded as armed and dangerous…

It's sort of all of the same thing.

My point was that with current incarnation of HCA you can't split hairs in making a comparison.  Unless you are on a "hard" measured point within the graph, there is enough opportunity for variance in the method that calculates scores that you can't necessarily claim one diamond's performance is actually higher or lower than another's when the difference is small. 

"Small" is a fluid term, in relation to the shape of the graph surrounding your actual point in question.

If carefully considerthe measured explicit data points that surround the nominal measured crown or pavilion angle of the diamond you are evaluating, and also consider the factors that are determining the performance that goes into HCA, you may be able to give more or less weight to a particular score, and/or to a difference between two scores.

HCA's stated original purpose was to allow a buyer (a commercial buyer at that) to rule in additional stones that might be of very high performance, yet not within AGS 0, or other traditional ranges.  Maybe it's at heart a pointer towards high performance RB territory, not a stone picker.   Some want to call it an exclusionary tool, one which isn't going to tell you which stone to pick, but rather which ones not to pick.  Which seems true within reason.  I wouldn't say averages are the issue.  It returns results based on assumptions, not on actual stones, in order to rule in (or to not rule out) more good stones than previous systems.  It's a starting point, and not a conclusion.   Some stones are still going to fall through the cracks in the assumptions, and end up outside consideration, that's hard to avoid.

It might seem very safe to say, "yes, under 2.0, that's all I'll look at".  But there are still stones that the methodology of the current HCA is probably shortchanging.  These are the stones that are right at certain thresholds, the "edge of the cliff", so to speak.

In my example (from the other thread with C34.8 for the following P angles mentioned), by the time you get to P41.5, performance has no doubt dropped off.  Look at enough scope images (real ones) and it's very clear.

The trouble is that most of the hairsplitting is attempted with stones that would be, in this case, P40.5 to P41.1 or so.

Some say under 40.6, performance is actually sliding already, though that shows up as the lowest HCA within that set of proportions.  Clearly as P increases to 41.5 performance is slipping, too.  The ultimate could really be 40.6, or 40.7, or 40.8, or 40.9, (and still nearly optimal slightly past that).  It's easier to find 40.9 selling at a premium, or certainly not for less than the other angles I just mentioned.

So 1 degree difference is a potentially very significant difference if it puts you over the threshold.  0.1 degree really isn't significant by itself, if you are trying to make a stone to stone comparison.

Consider that the real stones all do not have an actual single angle for the C and P facets, but a range, and that the extent of the range varies.  Some stones with 40.9 nominal and facets that vary .4 degree +/- aren't uncommon.  .2 is more common.  The point made about knowing the relationship between the actual pairs seems to be one of the biggest keys, probably the true isse.

Still, I find better idealscopes correlate as well to stones where the range of angles is small, and the nominal angle is in the middle of the range rather than shading toward the bottom with one or two high outliers.  Not all Vendors post the facet by facet measurements, more seem to post just the basic table of numbers, typically, so this may help distinguish stones you might want to learn more about. 

You can't take an HCA score too literally.  To use an analogy for applying HCA, let's say you want to find a certain house.  AGS cut puts you in the right time zone, while current HCA gets you into either the right state or possibly into the right area code, though what it is giving you numerically suggests the precision of a zip code.  

(And separately on percentages: Percentages rounded to whole numbers and then input into HCA may be dropping too much data in the process to allow you to give the score comparison much weight unless you are dealing with extremes.   Two angles off by a tenth could go through the averaging and conversion process and be compared as if they were off by as much as one whole degree. (Some numbers that are in range for one cut grade when using tenths of %, will round up to a whole number %that is in the next cut grade).  )

So in the example that starts this thread, HCA 4.5 and HCA 1.8 would lead most people to say don't touch the 4.5.  In reality, HCA's of these stones might be much much closer to each other. 

The poles of BIC and TIC are worth considering in a comparison, but don't worry about 1.7 versus .5 or anything else within the excellent range unless both stones are on the exact hard angle combinations that were used to build the HCA scoring methodology.  The other scores are approximations which may come out higher or lower than the true performance HCA would assign if it resolved performance details that fine.  So also don't pay a premium solely for a slightly better HCA in the absence of other differences between stones.  It's a great tool, a very valuable one, and a tool that seems to have taught many people an important lesson about visual performance, but you should know how to use it.  An Excellent HCA is an Excellent HCA unless you want to get under the hood and get dirty your hands with the details.

If you are going to get into the details, it seems likely that some stones are being short-changed on HCA scores.  For one, the "threshold stones" mentioned above, whose scores are potentially exaggerated to be higher numerically because of who their "hard" point neighbors are.  But for another, the "trough stones", which in reality might deserve the lowest HCA's of all, but which are located at minima that fall between the "hard" points are held back, scoring numerically higher, because of this, too. Neighbor on left is a 1.7, neighbor on right is a 1.7, so everybody in between also gets assigned 1.7 even though inbetween the stones probably dip to something lower.

(also, a stone at a maximum in the curve might get a benefit, too, but hopefully you aren't looking for the true highest HCA numbers).

On the flip side, a low number HCA might be coming out slightly lower than "true" in certain cases, but not by all that much from the curves I have looked at.  So if you didn't have detailed info on the facets, and you didn't have a scope picture, staying in the under 2 HCA range would keep things safer, the possible variance between calculated and "true" HCA smaller, which is why that piece of advice has become gospel for many.  This means, in the example I keep drawing angles from, 40.9 P and under.

If you are going to play near the edge of the cliff, a much tighter range of angle measurements may perform very well.  Idealscopes of nominal 41.0 and 41.1's can be all over the map, but if the variance of the actual facets is low, they usually look rather good.  Actually, I think the scope of a tight P41.0 in this range looks significantly better than the virtual version that shows up in GemAdviser and elsewhere.  The table on 41.0's has much a lighter pink for the virtual stones than the Idealscopes I see from real ones with "matching" numbers.  (it's over my head technically to speculate on why the virtual one is "pinker" than the real ones, but since the true experts on this topic are around somewhere, maybe they have some input?).  Nice 41.1s work, too.  At 41.2, it's becoming likely some facets stray past 41.3, and the idealscopes usually start paying for it, too.

I would be very interested to see, though, something in more detail about the "threshold area".  Does anybody with the real software toys want to plug in a 55% table, 34.8 or 35 crown, and run the pavilions from 40.5 to 41.5 by tenths?  What are the performance numbers showing?  How do the ray tracing images change?  With a better idea of what this might imply for the shape of the perfromance curve, it would be a better map for the "edge of the cliff".

Paul Slegers
Infinity Diamonds
www.CraftedByInfinity.com

Garry Holloway FGAA DipDT

HCA and Ideal-scope developer

http://www.ideal-scope.com and
http://www.HollowayDiamonds.com.au

Bill Leiberum
1948-2007
Thanks for everything and Rest In Peace,Bill.

Ira (Ruffles have Ridges) Z.
____________

If you're looking for diamond shopping tips...this (along with this update) might do the trick!

however...caution: known to have been regarded as armed and dangerous…

Garry Holloway FGAA DipDT

HCA and Ideal-scope developer

http://www.ideal-scope.com and
http://www.HollowayDiamonds.com.au

Ana "The greatest experts are only as good as the sum total of what they have seen." [Souren Melikian]