Anti CGC

[VintageComics]

And with THREE people assigning an aggregate grade,

 

There is no aggregrate grade. There is a Finalizer, and his grade is law.

 

 

The finalizer considers the grades of the first two and then makes a decision.

[selegue]

 

There was a guy on the EBay comics boards who is a math professor who talked there for several years about writing a grading guide and possibly some software to support it. I tried to get him to come post here to make the grading debates more interesting, but he's somewhat like the guy who started this post and dislikes certification, although I don't recall him taking the melodramatic "it's killing the hobby" stance this guy is taking. Think his name there was "gifflefunk."

 

I just wandered in to see what this long thread is about.

 

gifflefunk is on these boards -- the http://www.bipcomics.com/ guy, right?

 

He was very helpful a while back trying to straighten out some obscure publishers with inscrutable numbering systems for GCD. I'm not sure we ever finished that project.

 

Now to read more about this grading algorithm you're proposing. The oddest thing about the CGC to me is that it's coarse in the middle but fine at the edges.

 

Jack

 

 

 

[plitch]

And with THREE people assigning an aggregate grade,

 

There is no aggregrate grade. There is a Finalizer, and his grade is law.

 

A quick drive by post (I like this thread, thanks for all of the input):

 

Speaking as a Finalizer I have had other graders in the room tell me I would be flat out wrong if I gave a book grade X. If I am strongly out-voted by people I trust, I will put my opinion aside and trust in everyone else in the room.

 

But ultimatly, even when a situation like that occurs, the finalizer's grade is what goes out the door. So yes, it is law but there is no hubris.

 

That's why we call ourselves a "Grading Team".

[skypinkblu]

 

There was a guy on the EBay comics boards who is a math professor who talked there for several years about writing a grading guide and possibly some software to support it. I tried to get him to come post here to make the grading debates more interesting, but he's somewhat like the guy who started this post and dislikes certification, although I don't recall him taking the melodramatic "it's killing the hobby" stance this guy is taking. Think his name there was "gifflefunk."

 

I just wandered in to see what this long thread is about.

 

gifflefunk is on these boards -- the http://www.bipcomics.com/ guy, right?

 

He was very helpful a while back trying to straighten out some obscure publishers with inscrutable numbering systems for GCD. I'm not sure we ever finished that project.

 

Now to read more about this grading algorithm you're proposing. The oddest thing about the CGC to me is that it's coarse in the middle but fine at the edges.

 

Jack

 

 

 

 

Nick (Gifflefunk) is not a Math Professor, but I'll send him a link to this thread, the Math professor was someone else.

[tth2]

Just to separate the art from the science--how much you should deduct for a defect is an art, but once you've decided this, it's a science. A computer can't tell you how much to deduct for a defect because that's aesthetics, but once programmed with how much to deduct, they absolutely can tell you how much to deduct for a defined set of defects.

 

The gray area is separate from computers--it's whether or not it's possible at all to document how much to deduct for the infinite types of defects that are out there. Entropy and chaos theory tells us there are an infinite number of things that could happen to a comic...so what tth and Borock doubt is that it's possible to document infinite complexity. But you can by categorizing defects into taxonomical types--that's the scientific approach, and it's what graders do anyway in their heads lacking a written standard.

 

Heck, computerizing the calculation is the easy part. the calculation could even be (gasp!) hand-calculated if needed. The challenging part would be to:

1) determine all the components of a comic book

2) determine and codify all individual defects that each component may have

3) determine the value of each component's defect

4) apply an optimal methodology for a grader to capture each defect

 

It sounds like the biggest obsticle for this project would be to convince the users (GCG graders) that this is not impossible and that human observations can be standardized and quantified.

 

At any rate, it sounds like it would be a fun project!

Bill

Why don't you and FF create your computer program, and then we'll send you some raw books for you guys to grade. You can see how long it takes you to input the data and report what grade gets spit out. Then we'll send it to CGC and see what kind of grade they give. Then after a year, we'll send the same book to you and see if you input the data in exactly the way you did the first time, and see whether the same grade gets generated.

[LordRahl]

OK, I'm understanding now.

 

So you want a human to mark off a checklist whereby a program calculates a grade.

 

So you're still using subjective human input as raw data.

 

Right back where we've started.

 

That still doesn't take into account the feel, smell or over all appearance of a book and those are 3 very important factors when choosing a grade.

 

Comic grading is not just about a string of defects much like human health is not just about weight and height.

 

Because comics are 3 dimensional AND some subjectivity is a must to grading a book I believe it's just too much info (and too many variables) to be pigeonholed into a 2D program.

 

Of course that's just my opinion and I'm happy to be proven wrong.

 

 

You're not.

 

Well, at least this time....

 

The simple fact of the matter is this: people who have experience grading, and who use a *relatively* standard set of guidelines (ie, the 2nd edition of the OGG), will come to relatively close conclusions 99.9999934% of the time.

 

That means you're not going to get one person say a book is a "4.5", another a "1.8", and yet another a "9.2."

 

It's not going to happen.

 

And with THREE people assigning an aggregate grade, as opposed to just one, the margin for "subjectivity" is drastically reduced.

 

The real issue has always been, and remains, why then does the market insist on putting radical price differences on what amounts to tiny, fairly subjective differences in actual physical condition that could be "either/or" on any other day.

 

THAT is the real crux of the issue.

 

Because we live in a competitive society

[fantastic_four]

Because we live in a competitive society

 

Our society inhibits our natural competitiveness quite a lot by forcing us to get it under control to actually get along with other people...competitiveness is far more raw and frequent in every other uncivilized species. The bull with the biggest horns gets to lead the herd, and the dork with the biggest comic grades gets to be the envy of every other dork at cons and on Internet forums!

[gifflefunk]

I'm still around. I've been busy populating my indicia and ownership statement pages with more data (second pass; I'm in the Dell section).

 

The grading program I wrote was based on the first edition Overstreet Grading Guide and the 100-point system it contained. It works great when fed a list of quantitative (measurable) defects, but does lack any adjustments for qualitative (opinion based) defects. I still use it for objectively calculating the base grade for books in my personal collection.

 

When selling I usually go with a 3-point system (low/mid/high) for quickly grading large quantities of books. If I spend any more time grading a book I usually use the 2nd edition Overstreet Grading Guide and the 25-point system.

 

I'm not Anti-CGC, but I do get annoyed with some of the practices and the cases totally suck. When I buy a book I intend to read it and opening these freak'n slabs is very annoying.

 

I think someone could design a tamper-proof case that is easy to open. Perhaps a hinged clam-shell that has a see-through latch mechanism that unlocks the case by using a screw-driver. The latch is molded in such a way that it has visible plastic tabs that attach the axle to the walls of the latch mechanism. By turning the latch/axle with a screw driver the tabs break away (but are still sealed inside of the tamper-proof latch mechanism), the latch unlocks the case and you can now open the case to extract the book.

 

Basically a slab you could re-use for storage, even if the grade has become invalidated by breaking the "seal" (tabs) in the latch.

 

 

[AtlasFan]

Just to separate the art from the science--how much you should deduct for a defect is an art, but once you've decided this, it's a science. A computer can't tell you how much to deduct for a defect because that's aesthetics, but once programmed with how much to deduct, they absolutely can tell you how much to deduct for a defined set of defects.

 

The gray area is separate from computers--it's whether or not it's possible at all to document how much to deduct for the infinite types of defects that are out there. Entropy and chaos theory tells us there are an infinite number of things that could happen to a comic...so what tth and Borock doubt is that it's possible to document infinite complexity. But you can by categorizing defects into taxonomical types--that's the scientific approach, and it's what graders do anyway in their heads lacking a written standard.

 

Heck, computerizing the calculation is the easy part. the calculation could even be (gasp!) hand-calculated if needed. The challenging part would be to:

1) determine all the components of a comic book

2) determine and codify all individual defects that each component may have

3) determine the value of each component's defect

4) apply an optimal methodology for a grader to capture each defect

 

It sounds like the biggest obsticle for this project would be to convince the users (GCG graders) that this is not impossible and that human observations can be standardized and quantified.

 

At any rate, it sounds like it would be a fun project!

Bill

Why don't you and FF create your computer program, and then we'll send you some raw books for you guys to grade. You can see how long it takes you to input the data and report what grade gets spit out. Then we'll send it to CGC and see what kind of grade they give. Then after a year, we'll send the same book to you and see if you input the data in exactly the way you did the first time, and see whether the same grade gets generated.

 

Thats sounds like a nice idea, but in order to create a usable model, an analyst would need detailed input from "subject matter experts" regarding the nitty-gritty details on the grading of comics. The developer would also need to be aware of the entire process, the environment that the process would be performed in, who the users are, etc.

 

Its really about nailing down all the components of a comicbook and the all the possible defects that each component could have. Writting a computer program is just an option to record the information in the process. The score could be calculated by hand if needed. Its the process that matters, not the program.

 

I have seen too many developers think that they understand enough of a process to fully automate something -- and have the end product fail miserably becuase they did not have input from the people who would be using the system (or process).

 

...but feel free to send me your raw books anyway!

[fantastic_four]

The grading program I wrote was based on the first edition Overstreet Grading Guide and the 100-point system it contained. It works great when fed a list of quantitative (measurable) defects, but does lack any adjustments for qualitative (opinion based) defects.

 

Great to see you around here! (thumbs u I eventually quit the EBay boards because they were too newbie-oriented, which is fine, but eh, gets boring...your insightful and well-written posts about grading were what kept me reading that board for any length of time.

 

What are some examples of the qualitative defects you've been unable to model? I'm guessing you mean attributes that lacking any actual tool to measure them people tend to measure by assigning their own value judgments, such as staple rustiness, reflectivity, brightness of colors, page whiteness, and gloss. Did you for some reason rule out just factoring those in the same way people end up doing it anyway--by assigning an artificial score to those qualities and factoring anything below perfect in with the rest of the defects? The easiest example of this to relate to since CGC started putting it on their labels is page quality--there is no scientific way to measure it that's easily accessible to collectors, so CGC makes their best guess based upon visual analysis, assigns one of their qualifiers, and if it's too low for the overall grade, they factor it in. CGC measures this on the w, ow/w, ow, cr/ow, cr, tan/cr, tan, slightly brittle, and brittle scale, which unless I'm missing some tweener colors (which I probably am) is 9 notches, so mathematically, anything below 9 (or higher than 1, whichever way you want to start) is a defect and gets factored in.

 

Is assigning judgment-based grades to qualitative defects imperfect? Absolutely...but it's the only thing we can practically do. And it's what we do anyway in our heads as we grade.

[showcase22gr1959]

Because we live in a competitive society

 

Our society inhibits our natural competitiveness quite a lot by forcing us to get it under control to actually get along with other people...competitiveness is far more raw and frequent in every other uncivilized species. The bull with the biggest horns gets to lead the herd, and the dork with the biggest comic grades gets to be the envy of every other dork at cons and on Internet forums!

 

Well said and quite true most of the times. This what makes a great market for collecting. (thumbs u

[fantastic_four]

Why don't you and FF create your computer program, and then we'll send you some raw books for you guys to grade.

 

Think about how long it takes to be a tight grader yourself. Multiply that by a factor of 3 to 5 to get it into code--not to mention the much more difficult and likely impossible challenge of making the user interface so fast to use that people would actually use it so that it wouldn't slow down their grading, which is the driving reason I haven't started already. Next, think about how long it would take to not only be a tight grader, but to resolve the amount that you personally weigh your defects with the way most other educated, objective tight graders weigh their defects, as nobody wants to use a grading system--software or otherwise--that isn't reflecting the way tight graders are actually grading. Realize that this is a never-ending task that to be useful needs to evolve as people refine the weights they assign to defects as time goes on. I would only even start down this path is with collaborative Internet-based software that allows the comics community itself to develop that standard as a whole. I haven't seen that type of collaborative software yet but I've had the specs in my head for almost a decade now...wikis are close to it but not visual or visually structured enough to emphasize the hierarchical nature of the information. Or maybe I'm being overly-pessimistic about the collaborative web-based software commonly available. All of this is the giant PITA I envision with computer-based grading and much tighter grading scales in general.

 

These kinds of systems only get created to be much good when they're applied to areas of knowledge that are of such wide interest that it's worth the time and money it costs to develop them...or when someone's so passionate about it that they just do it anyway, which is rare. They're not impossible, they're just impractical for comics--the market isn't big enough. Yea, Action 1 8.5 sold for a million and 8.0, but it sold for that much even given the common knowledge that CGC can vary by a notch, meaning the 1.5 mil book could've been the 1 mil book or vice-versa on any given grading day. We grudgingly accept the imprecision in grading, we live with it, and we keep buying...if the market was bigger, if these were shares or gold or some other mass-market asset, there'd be enough money riding on this imprecision that someone would find a profit in eliminating it. There may be such a profit to be made in comics...I just can't see where it's at.

[VintageComics]

I can't believe you took the time to type that all out.

 

I think Tim was being facetious.

 

 

 

[fantastic_four]

I can't believe you took the time to type that all out.

 

He was being snide. I, on the other hand, am a huge grading geek...I did it for me, not for him.

[VintageComics]

I can't believe you took the time to type that all out.

 

He was being snide. I, on the other hand, am a huge grading geek...I did it for me, not for him.

 

I love detail and attention to detail...but not for the sake of detail.

 

[fantastic_four]

I love detail and attention to detail...but not for the sake of detail.

 

It's for the sake of precise grading...I'll debate grading all day, even with the most argumentative of posters, whereas I avoid almost all other arguments...unless I'm in a combatative mood. Not for the sake of arguing or being right no matter what, for the sake of figuring out what we or I haven't figured out yet.

[tth2]

Why don't you and FF create your computer program, and then we'll send you some raw books for you guys to grade.

 

Think about how long it takes to be a tight grader yourself. Multiply that by a factor of 3 to 5 to get it into code--not to mention the much more difficult and likely impossible challenge of making the user interface so fast to use that people would actually use it so that it wouldn't slow down their grading, which is the driving reason I haven't started already. Next, think about how long it would take to not only be a tight grader, but to resolve the amount that you personally weigh your defects with the way most other educated, objective tight graders weigh their defects, as nobody wants to use a grading system--software or otherwise--that isn't reflecting the way tight graders are actually grading. Realize that this is a never-ending task that to be useful needs to evolve as people refine the weights they assign to defects as time goes on. I would only even start down this path is with collaborative Internet-based software that allows the comics community itself to develop that standard as a whole. I haven't seen that type of collaborative software yet but I've had the specs in my head for almost a decade now...wikis are close to it but not visual or visually structured enough to emphasize the hierarchical nature of the information. Or maybe I'm being overly-pessimistic about the collaborative web-based software commonly available. All of this is the giant PITA I envision with computer-based grading and much tighter grading scales in general.

So basically after all that, you`re agreeing with me that it`s too hard, which is what I said in the first place.

 

These kinds of systems only get created to be much good when they're applied to areas of knowledge that are of such wide interest that it's worth the time and money it costs to develop them...or when someone's so passionate about it that they just do it anyway, which is rare. They're not impossible, they're just impractical for comics--the market isn't big enough. Yea, Action 1 8.5 sold for a million and 8.0, but it sold for that much even given the common knowledge that CGC can vary by a notch, meaning the 1.5 mil book could've been the 1 mil book or vice-versa on any given grading day. We grudgingly accept the imprecision in grading, we live with it, and we keep buying...if the market was bigger, if these were shares or gold or some other mass-market asset, there'd be enough money riding on this imprecision that someone would find a profit in eliminating it. There may be such a profit to be made in comics...I just can't see where it's at.

You think so? Rating debt is a huge business with hundreds of billions of dollars riding on it, as we`ve just seen over the past couple of years. And yet it`s done by very imprecise rating companies with lots of subjective grading. Now you could argue that maybe that`s the problem, since they got things so badly wrong, but I`m sure if there were a better model that could be used, someone would`ve created it by now.

 

At the end of the day, for some things you just can`t replace human judgment.

[gifflefunk]

Page color is one factor, which I think could be made quantitative if paper "type" identification data was accumulated. But that starts to get into a level of forensics that is beyond a basic system.

 

Other qualitative areas are cover gloss. Again a basic classification system of dull, normal, and shiny could be created, but it isn't easily measurable for a basic system.

 

The impact a defect has on appearance is another qualitative issue. A color breaking crease on a black cover stands out more than that exact same crease on a white cover book. While the two creases could be quantitatively similar (in the same location, same area affected by length and width, etc.) they could have different qualitative values when it comes to impacting how a book presents those defects. Namely the aesthetic factors.

 

Then there are aesthetic factors fall into the realm of personal likes and dislikes. There are some collectors who have issues with the vertical amount of back cover that appears on the front cover of a book. Others have issues with staple placement (are they perfectly placed on the cover fold or are they offset onto either the front or back cover). Are the staples perfectly aligned. If one staple is 3 degrees off from being parallel with the spine is enough to make some collectors crazy.

 

Who is to say that a book with some slight back cover wrapping to the front side is a worse defect than having the staples off-spine or slightly misaligned? These are factors that could be measured, but how their values impact the grade are not clearly defined by the collecting community.

[VintageComics]

How about non color breaking indents, varying degrees of fading, tanning on the inside covers, different sized cuts....the list goes on and on.

 

It's like trying to discuss the reinvention of the wheel.

 

It ain't gonna happen.

 

You'd have more productive discourse on trying to figure out how to make human grading more consistent than trying to replace it with a machine.

 

 

[AtlasFan]

Page color is one factor, which I think could be made quantitative if paper "type" identification data was accumulated. But that starts to get into a level of forensics that is beyond a basic system.

 

Other qualitative areas are cover gloss. Again a basic classification system of dull, normal, and shiny could be created, but it isn't easily measurable for a basic system.

 

The impact a defect has on appearance is another qualitative issue. A color breaking crease on a black cover stands out more than that exact same crease on a white cover book. While the two creases could be quantitatively similar (in the same location, same area affected by length and width, etc.) they could have different qualitative values when it comes to impacting how a book presents those defects. Namely the aesthetic factors.

 

Then there are aesthetic factors fall into the realm of personal likes and dislikes. There are some collectors who have issues with the vertical amount of back cover that appears on the front cover of a book. Others have issues with staple placement (are they perfectly placed on the cover fold or are they offset onto either the front or back cover). Are the staples perfectly aligned. If one staple is 3 degrees off from being parallel with the spine is enough to make some collectors crazy.

 

Who is to say that a book with some slight back cover wrapping to the front side is a worse defect than having the staples off-spine or slightly misaligned? These are factors that could be measured, but how their values impact the grade are not clearly defined by the collecting community.

 

This is exactly what I'm talking about...forget the automation for a moment...each component part of a comic book would have to be identified and all the possible defects or levels of measurement (in this example, page quality) would have to be quantified.

 

Only when this is all done, would it be possible to come up with some kind of system to record the data and calculate the grade score.

 

This is not an impossible task and I would think that the folks who run CGC must have already considered this option long ago. The real question is: Why have they not implemented it?

 

Is it too costly?

Would it slow down the grading process?

Is there not enough demand for precision?

 

--or---

 

Do the folks at CGC like to "keep it loose" so they can choose to be "soft" sometimes and "firm" at other times.