Grading between ‘Good’ and ‘Very Good’ — debating a specific example

eugeneandresson

It gets a score of 13.5



😗



I am not disagreeing, but I still don't see how this is relevant when trying to put an honest fair-and-square price to a watch I was looking at in this scenario (unless I was trying to #buycheapsellexpensive).

Lets open up the rabbit hole by switching to examples, and bear with me, and (anybody) feel free to improve/extend the example with more data points and different gradings etc. I am sucking these watch values from thin air based on e.g. the unloved 145.006/16's).

I want to buy a watch similar to @kov's 'excellent' exotic I see for sale. I grade @kovs watch at 7/10. Because I am a motivated collector, and I follow these things, I grade the watch I want to buy a relative 7/10 compared to his...its not as nice...I also give it a 5/10 total grading (case is similar, but the dial is not as attractive (a few spots/marks and some drag on the minute counter, and dirty-ish lume on some plots, but all plots present) and there is some lume loss on one of the hands (I would rate it less if there was some lume loss on the dial)). Lets assume his watch has an established value of 3.5k (lets further assume we saw multiple sell for a similar price in that condition, thus a realistic market value) and thus the 'highest' point in our value grading table. There was a shagged-dog or 3 that went for round 1k with a single or two issues (horrible lume / scraped lume / shagged dial / overpolished case etc) rated 3/10. I need to use the table to help me figure out the price. Here is the 'market' for the sales we have seen, for what everyone is calling 'unbiased grading' (showing spaces for grades that do not exist) data that is all over the internet.

'*' is where I grade the watch I want to buy, which is exactly in the middle of our know datapoints. So unbiased value calculation based on this would be (following the line, or linear interpolation) :

(1 + (3.5 - 1)/2) = 2.25k

for a 5/10 watch based on this data.

Fair enough, lets make the table simpler.

I have rated the watch I want to buy at roughly 7/10 compared to the best we have (@kov's at 3.5k ... scarlet is excellent, but not necessarily the best 😉 ). So a 'only data we have' based value calculation on this would be :

0.7 * 3.5k = 2.45k

However, my gut feeling (based on the watch pictured in my mind) is telling me that 2.1k would be the correct value for this watch...I really do not find dirty dial lume attractive, even though 'all dial lume present' is the only rating criteria, and whilst minor, hand lume loss now bothers me (I am a pedant).

Some conclusions:

There is a slight difference ($200 in this example, against the 'bias') ... so 'team-no-bias' you win a point (im on no team, by the way 😉 )...
- which would get 'larger' when using this method on a very inaccurate data set e.g. single datapoint 400k auction outlying result on a very rare reference
- which would make no significant difference on 90% of the watches people are hunting (down to 105.003s), based on the data points already there, where data is more accurate (not 1 off, many sales at similar prices) and ranges from 'shagged dog' to 'Mint'

In the second case one would use this method to get to the correct 'starting point' on the curve and then based on intuition (the brain is a wonderful machine, and does very complex things quite seamlessly), move along from there...

Using an absolute grade would require no room for intuition (other than coming up with the correct grading and following all the guides).
- this becomes way more complicated to follow (will average-Joe be able to use it, or even do a simple 'follow the line' calculation?) but would love to see such a system for watches!

A grade of 1 to 10 is most likely not enough 'resolution' for an absolute grading system ... might have to be out of 100 ... or more.

How to deal with subjectivity / attraction (or ugliness)? I grade the 'top example watch' at 7/10 ... but the owner may think its an 8/10.

Using any system like this implies a certain level of experience in the thing being graded, so not for 'noobs' (how can a noob grade the polished state of a case, for example, which will in a stiff rigid 100+ point system need to be done)
- the person using it will in anycase have to know what the rating watch would look like.

This is not trivial. Say what you like, but until someone can come up with a better grading/value system, that encompasses everything, SM101 is actually ok (for speedmasters). And apart from giving something an accurate rating for correctness sake, I still don't see how I benefit by doing so when I am trying to work out a value if I don't have the full dataset for all possible ratings for the particular watch.

Anybody want to try and come up with a 'coin' like approach for watches?

PS: While not 'conclusive', these are random watch-values that first came to mind based on more pedestrian watches. I did not choose them to make things look close ... perhaps using different values, or a bigger spread, the picture looks radically different.

watchlovr

Also, using the comic book/coin grading systems suggested so strongly above, how can you grade a brown dial on a Speedmaster?

pdxleaf

@eugeneandresson

A lot of food for thought.

First, i think we can all agree on using Scarlett Johansson as the Excellent example.

I think i understand your difficulties with the grading system concept. Correct me if i get this wrong. I think it is based on different expectations for what a grading system should or would do.

From your example you are expecting that a grading system should be able to determine the current market value of a watch relative to all other graded watches. An analogy would be like financial analysis of a company to determine market value by comparing assets to debt, cash to current liabilities, etc...

My expectation for what a grading system should do is different. Using the financial analysis example, my thought is that a grading system would be more like an audit of a company (stick with me.). The audit uses a standard to test the acounts of different companies. The audit determines if the bits are accurate, is the cash amount accurate, has the company correctly reported the amount of debt?

A person uses audited results to do their analysis. The grading system that I think is being proposed would'nt do the analysis. Rather, it would set the standard for how to judge different entities or watches. I think @JwRosenthal is suggesting the same (don't want to speak for others.)

In some respect speedmaster101 does this now. He has a standard for grading and then he looks at the market to determine what the values are for each group. His chart is not indicating whether the market is accurate or even realistic between grades. But his chart is a grading system that can in turn be used to compare similar objects.

Hopefully this makes some sense.

Also, I am not trying to convince you or anyone that a grading system will work or is even necessary. I am only trying to understand the objections. It seems that most people use a system. Perhaps it's only that people think it isn't necessary or workable.

airansun

Nope. Not.

Sorry.

Oh, and my wife tells me Scarlett isn’t an object... so, there’s that too.

STANDY

...I am quiet happy to have a top of the Very Good and bottom of the Very Good. Which is what the OP has alluded to wanting to express.

pdxleaf

@watchlovr



Maybe another scale should be:
Blah.
Yay, it's nice.
Ohhh, can I see that?
The hairs on my arms are standing up.
Holy Shit!