happytraveler
Rough_Rock
- Joined
- Jul 12, 2004
- Messages
- 5
Size ranges
- Thread starter happytraveler
- Start date
- Status
There are a few reasons I've heard mentioned for the phenomenon. Firstly, people do snatch them up really quickly. Secondly, cutters often have a choice between a .9x ct ideal cut diamond, and a 1.0x non-ideal (or close to ideal) cut stone... since there's such a huge price jump at 1ct, it just doesn't make financial sense to cut those diamonds like that -- those .9x ct ideal cuts often take more time to cut, and they don't make as much money as they could on them.
happytraveler
Rough_Rock
- Joined
- Jul 12, 2004
- Messages
- 5
Thanks, that makes a lot of sense.
I just wanted to say thanks to Leonid for putting together this site and the philosophy behind it.
On a sidenote, I'm an engineer and enjoy math - I've been able to reverse engineer within 3% tolerances some vendors' pricing models. If anybody cares, you can set up a General Linear Model similar to designing experiments with Six Sigma and Taguchi methods. It will give you correlation coeffecients that show how a vendor values certain factors.
I just wanted to say thanks to Leonid for putting together this site and the philosophy behind it.
On a sidenote, I'm an engineer and enjoy math - I've been able to reverse engineer within 3% tolerances some vendors' pricing models. If anybody cares, you can set up a General Linear Model similar to designing experiments with Six Sigma and Taguchi methods. It will give you correlation coeffecients that show how a vendor values certain factors.
- Joined
- May 3, 2001
- Messages
- 7,516
Regular Guy
Ideal_Rock
- Joined
- Jul 6, 2004
- Messages
- 5,963
MelanieC
Shiny_Rock
- Joined
- Jun 13, 2004
- Messages
- 127
I agree with Wink. That might have well been posted in another language. It went way over my head!
I definately don't see too many in the right under 1ct range. When I was buying my diamond though, he had a .98 F SI1 that he had just got in and was waiting on the Cert so I couldn't see what the measurements were. I ended up going with a 1.08ct diamond instead.
I definately don't see too many in the right under 1ct range. When I was buying my diamond though, he had a .98 F SI1 that he had just got in and was waiting on the Cert so I couldn't see what the measurements were. I ended up going with a 1.08ct diamond instead.
happytraveler
Rough_Rock
- Joined
- Jul 12, 2004
- Messages
- 5
Sure -
I apologize in advance to the people that this is likely to bore!
If you look at a bunch of diamonds from one vendor, you guess at their formula. They might price by adding a % of what they paid for the diamond, or they might try to use a formula to price their diamonds "smarter" than the formula that their wholesaler used. This would take into account some or all the determinents of value.
Maybe every time you drop a color grade, you guess that the price per carat will drop by X dollars. That's linear. And everybody already knows that all things equal, doubling the weight is more than doubling the price. So maybe that's a square factor, or maybe exponential.
When the analysis is done, you can say in statistical terms how much value, in price per carat, a dealer (or the market) places on a factor (color, clarity, even flourescence). That's the correlation coeffecient.
The next test is the variance of the correlation coeffecient. That sounds complex, but it's just telling you how often the diamond specs you fed into it fit your model well. All other things equal, if a color upgrade from G to F sometimes is worth $100, sometimes $400, sometimes $600 - then that vendor isn't using a highly predictable system, or a system at all. You'd have a very high variance on your correlation coeffecient.
When you feed a General Linear Model a bunch of data, it's really good at taking a bunch of "orthoginal factors" and pretending like you had only changed one thing at a time, even though maybe each of the 500 diamonds you analyzed changed 2, 3, or 4 factors at a time. Color and clarity are probably orthoginal factors in their middle regions.
These Taguchi methods are why in the 80s, import cars got a reputation for being higher quality than domestics. Mr Taguchi was helping Japaneese makers find out that when they changed one thing (maybe size of a brake pad), it failed a lot less. But they only had to make a few cars that changed a bunch of carefully chosen things at once - and they were able to infer a lot of useful data that made them more reliable.
I apologize in advance to the people that this is likely to bore!
If you look at a bunch of diamonds from one vendor, you guess at their formula. They might price by adding a % of what they paid for the diamond, or they might try to use a formula to price their diamonds "smarter" than the formula that their wholesaler used. This would take into account some or all the determinents of value.
Maybe every time you drop a color grade, you guess that the price per carat will drop by X dollars. That's linear. And everybody already knows that all things equal, doubling the weight is more than doubling the price. So maybe that's a square factor, or maybe exponential.
When the analysis is done, you can say in statistical terms how much value, in price per carat, a dealer (or the market) places on a factor (color, clarity, even flourescence). That's the correlation coeffecient.
The next test is the variance of the correlation coeffecient. That sounds complex, but it's just telling you how often the diamond specs you fed into it fit your model well. All other things equal, if a color upgrade from G to F sometimes is worth $100, sometimes $400, sometimes $600 - then that vendor isn't using a highly predictable system, or a system at all. You'd have a very high variance on your correlation coeffecient.
When you feed a General Linear Model a bunch of data, it's really good at taking a bunch of "orthoginal factors" and pretending like you had only changed one thing at a time, even though maybe each of the 500 diamonds you analyzed changed 2, 3, or 4 factors at a time. Color and clarity are probably orthoginal factors in their middle regions.
These Taguchi methods are why in the 80s, import cars got a reputation for being higher quality than domestics. Mr Taguchi was helping Japaneese makers find out that when they changed one thing (maybe size of a brake pad), it failed a lot less. But they only had to make a few cars that changed a bunch of carefully chosen things at once - and they were able to infer a lot of useful data that made them more reliable.
Lynn B
Ideal_Rock
- Joined
- May 9, 2004
- Messages
- 5,609
pqcollectibles
Ideal_Rock
- Joined
- Feb 22, 2003
- Messages
- 3,441
Those are 2 highly desireable carat weight ranges, 0.9-0.99 and 1.25, and they go quickly. If you're interested in diamonds in either size range, you might try sourcing thru a Vendor to look for you, and be prepared to wait until the right diamond comes along.
As far as pricing, I like Paul Slegers explanation better.
"If you take the basic price of I-SI2 at a ratio of 100, then an I-SI1 is about 115, an H-SI2 is about 111 and an H-SI1 is about 130. Therefore, you cannot directly compare an I-SI2 with an H-SI1. The second one has inherently a 30% higher price, so if you get the I-SI2 at the same price, you are paying a premium of 30%."
https://www.pricescope.com/community/threads/would-you-buy-a-hof-diamond-if-it-was-only-5-10-more-should-i.16958/
Much simpler to understand.
As far as pricing, I like Paul Slegers explanation better.
"If you take the basic price of I-SI2 at a ratio of 100, then an I-SI1 is about 115, an H-SI2 is about 111 and an H-SI1 is about 130. Therefore, you cannot directly compare an I-SI2 with an H-SI1. The second one has inherently a 30% higher price, so if you get the I-SI2 at the same price, you are paying a premium of 30%."
https://www.pricescope.com/community/threads/would-you-buy-a-hof-diamond-if-it-was-only-5-10-more-should-i.16958/
Much simpler to understand.
icelady
Brilliant_Rock
- Joined
- Nov 25, 2003
- Messages
- 1,030
noobie
Brilliant_Rock
- Joined
- Mar 3, 2004
- Messages
- 1,318
pqcollectibles
Ideal_Rock
- Joined
- Feb 22, 2003
- Messages
- 3,441
----------------
On 7/13/2004 9:12:22 PM noobie wrote:
Yes PQ, but I'm still pretty sure that you understood his statistical analysis description perfectly----------------
On 7/13/2004 8:54:51 PM pqcollectibles wrote:
Much simpler to understand.
----------------
----------------
Ssshhhhhhhhhhhhh, Noobie!! The "Dumb Blonde" routine has worked for me so far. I don't wanna blow it!!
- Status
How to Buy the Best $25,000 Engagement Ring: A Smart Shopper’s Guide
How to Buy the Best $25,000 Engagement Ring: A Smart Shopper’s Guide - 02/24
300x240.png)