Grading Comparisons Between Dealers 3 3

[cheetah]

One of the advantages/disadvantages of collecting full runs of CGC-graded books is that you get to buy raw books from vendors and get them graded in order to fill in the holes. Sometimes you get good surprises, sometimes you get bad ones. I’ve been doing it long enough now that I’ve built up a pretty sizable data base of raw-to-graded books and I thought it would be a good time to divulge some of my data in regards to which vendors grade well and which ones grade poorly. There is always a lot of talk on the boards as to who sucks and who doesn’t. I think my information will be a surprise in some ways, and reinforce opinions in some others.

 

First off, here is how I set up may data. The 10.0 grading scale is a qualitative scale that doesn’t reflect quantitative difference between books. As such, I have prepared my data based on ‘Grade Increments.’ A Grade Increment is the increase in quality from one grade to another. So the difference between 5.0 and 5.5 is 1 GI. Similarly, the difference between a 9.6 and a 9.8 is also 1 GI. For each of my books, I recorded the grade advertized during the sale and the CGC assigned grade once it went through the grading process. If a book was sold as at a grade higher than it received from CGC, I recorded the Grade Increment as a negative. A book I bought as an 8.0 but which received a 6.5 from CGC received -3 GI. A book bought as a 6.5 that graded an 8.0 received 3 GI.

 

Virtually all of the books I had graded were Golden Age books. I sent them all through Matt Nelson or Joey Post for prescreens and pressing, if appropriate. With those I sent through Matt, there were books where he identified problems that I had not seen. These books were returned to me then to the vendor and were not included in the data. These returned books may skew the data to give the vendors a falsely elevated score as books that were graded much too high have been excluded from the analysis.

 

As you might expect, pressing of the books also may give the vendors a falsely high score. A book that warrants a 7.5 prior to pressing may grade 8.5 after pressing. While this throws a degree of uncertainty into the results, the books I buy are usually fairly expensive and I’d like to preserve their value so we will just have to live with a little uncertainty.

 

As far as the data goes, I have collected data on seven vendors. I have graded anywhere from 7 to 78 books with the seven vendors. All the books were graded between December 2012 and January 2014 so the results reflect recent grading trends.

 

I have not, at this time, provided data on exactly who the seven vendors are. I thought it might be interesting to discuss the data and see what preconceptions we have concerning who is ‘good’ and who is ‘bad.’ You will have to play along if you want to see the data related to the names.

 

The data and its discussion will be presented in the next post.

[cheetah]

Figure 1 presents the data on the number of books graded/vendor along with the Cumulative Grading Increments each vendor was off. This value was obtained by adding all the Grade Increments from all graded books. A note here, it is possible for a vendor to be wildly inaccurate but still receive a Cumulative Grade Increment of zero. Calling a 9.8 a 0.5 when the next book is a 0.5 but called a 9.8 gives a Cumulative Grade Increment of zero.

 

[cheetah]

Figure 2 shows the average Grading Increment each vendor was off. A negative Average GI indicates a tendency for a vendor to over-grade their books while a positive Average GI shows a vendor has a tendency to under-grade their books. It is worthwhile to reinforce that all of these books went through pressing companies. All were not pressed, but you can assume that most were.

 

The data show a wide variance in grading practices by vendor. Dealer 3 routinely under-graded their books in relation to CGC while Dealer 2 routinely over-graded their books. Both vendors averaged more than one Grade Increment difference from the CGC grade: one positive and one negative.

 

In reviewing the data, it became clear than nearly every vendor had the potential to provide a catastrophically over-graded book. The vast majority of times, this was the result of Page Quality or the presence of stains, two areas where CGC places more emphasis than does the rest of the industry. Catastrophically over-graded books ranged from -8 GI to 0 GI depending on the vendor. In order to limit the influences of these books on a vendor's score, I also prepared the Average GI without the most poorly graded book.

 

[cheetah]

Realizing that the Cumulative and Average GI scores can be misleading since wildly inaccurate grading can mimic consistently accurate grading, Figure 3 provides data showing the percent of books over-graded, under-graded and accurately graded. Based on these data, Dealers 6 and 1 were the most accurate while Dealers 3 and 4 under-graded the majority of their books and Dealer 2 over-graded the majority of their books.

 

Dealers 6 and 3 were the vendors most likely to sell a book at a grade below or equivalent to the grade assigned by CGC.

 

[cheetah]

A few things I should point out with the data. First, it is worrisome that the three vendors who scored highest on Average GI were also the three with the fewest books graded. It is possible that these are affected by the low number of data points.

 

Second, I am currently collecting data on the average increase in pressing/regrading CGC-graded books. I hope to have this summarized in the next few days. This may give a clearer idea as to what kind of average increase can be expected from pressing. This information may shed further light on how accurate/inaccurate vendors are.

 

Finally, the data show that there is considerable variability in the grading practices between the vendors that most of us use on a regular basis.

[ft88]

Cool

 

[cheetah]

Okay, that is everything.

[Scrooge]

Does Dealer 3 undergrade but overprices?

[cheetah]

Does Dealer 3 undergrade but overprices?

 

No, that is not the reputation I have heard.

[jimjum12]

Realizing that the Cumulative and Average GI scores can be misleading since wildly inaccurate grading can mimic consistently accurate grading, Figure 3 provides data showing the percent of books over-graded, under-graded and accurately graded. Based on these data, Dealers 6 and 1 were the most accurate while Dealers 3 and 4 under-graded the majority of their books and Dealer 2 over-graded the majority of their books.

 

Dealers 6 and 3 were the vendors most likely to sell a book at a grade below or equivalent to the grade assigned by CGC.

 

 

..... to make sure I've got this right.... this graph would indicate the dealer 5 did not overgrade ANY books ? GOD BLESS....

 

-jimbo(a friend of jesus) (thumbs u

[cheetah]

Yes, Dealer 5 did not over-grade anything but I have only graded 7 of their books.

[jimjum12]

Yes, Dealer 5 did not over-grade anything but I have only graded 7 of their books.

 

..... still, that's not a bad sample group. Unfortunately, it appears that several are not, shall we say, on the same page as CGC. This is an impressive effort on your part, Jeff...... shows a high degree of discipline on your part. I'm excited to see more details. For me, a dealer who overgrades is not always a deal breaker for me... so long as they're consistent. With pressing involved, the data may skew the differentiation between accuracy and consistency, but still, this is a study the likes of which I don't recall ever seeing. Your attraction to so many different types of books (you 'ho. you....) makes this data all the more compelling. GOD BLESS...

 

-jimbo(a friend of jesus) (thumbs u

[thehumantorch]

Very interesting thread. I'm sure it took you a lot of work to put it together I'm sure we would all like to know the dealers are.

[cheetah]

Very interesting thread. I'm sure it took you a lot of work to put it together I'm sure we would all like to know the dealers are.

 

I assume the thread will have a short shelf-life once I put the dealer's names down. I expect it will be locked or deleted once the riff-raff from CG get wind of it.

[cheetah]

Yes, Dealer 5 did not over-grade anything but I have only graded 7 of their books.

 

..... still, that's not a bad sample group. Unfortunately, it appears that several are not, shall we say, on the same page as CGC. This is an impressive effort on your part, Jeff...... shows a high degree of discipline on your part. I'm excited to see more details. For me, a dealer who overgrades is not always a deal breaker for me... so long as they're consistent. With pressing involved, the data may skew the differentiation between accuracy and consistency, but still, this is a study the likes of which I don't recall ever seeing. Your attraction to so many different types of books (you 'ho. you....) makes this data all the more compelling. GOD BLESS...

 

-jimbo(a friend of jesus) (thumbs u

 

Thanks, Jimbo. It's been a matter of finding a decent way of presenting it. The process has given me a lot of insight into how the same dealer can grade different types of books differently. There are definite trends. For example, Pre-code Horror books are routinely overgraded more often than superhero books. I assume dealers grade on a curve for books that are less common in high grade.

[rjpb]

Yes, Dealer 5 did not over-grade anything but I have only graded 7 of their books.

 

..... still, that's not a bad sample group. Unfortunately, it appears that several are not, shall we say, on the same page as CGC. This is an impressive effort on your part, Jeff...... shows a high degree of discipline on your part. I'm excited to see more details. For me, a dealer who overgrades is not always a deal breaker for me... so long as they're consistent. With pressing involved, the data may skew the differentiation between accuracy and consistency, but still, this is a study the likes of which I don't recall ever seeing. Your attraction to so many different types of books (you 'ho. you....) makes this data all the more compelling. GOD BLESS...

 

-jimbo(a friend of jesus) (thumbs u

 

The only dealer who seems routinely weak is #2. Both #4 & #7 seem to have a tendency to overgrade more than undergrade, but in both cases the majority of books had grades that agreed with CGC or were undergraded. Though without knowing what percentage were pressed, the initial grades may well have been softer. With a slightly larger sample size, and an 80% chance of getting an undergraded book ( or at least improvable book), dealer #3's stats are actually more impressive than even #5's.

[jimjum12]

Very interesting thread. I'm sure it took you a lot of work to put it together I'm sure we would all like to know the dealers are.

 

I assume the thread will have a short shelf-life once I put the dealer's names down. I expect it will be locked or deleted once the riff-raff from CG get wind of it.

 

.... this is exactly the type of information that SHOULD be in a forum like this.... I hope it doesn't happen that way. A retailer in the game for the long haul should welcome this oppurtunity for self assessment. GOD BLESS....

 

-jimbo(a friend of jesus) (thumbs u

[jimjum12]

Yes, Dealer 5 did not over-grade anything but I have only graded 7 of their books.

 

..... still, that's not a bad sample group. Unfortunately, it appears that several are not, shall we say, on the same page as CGC. This is an impressive effort on your part, Jeff...... shows a high degree of discipline on your part. I'm excited to see more details. For me, a dealer who overgrades is not always a deal breaker for me... so long as they're consistent. With pressing involved, the data may skew the differentiation between accuracy and consistency, but still, this is a study the likes of which I don't recall ever seeing. Your attraction to so many different types of books (you 'ho. you....) makes this data all the more compelling. GOD BLESS...

 

-jimbo(a friend of jesus) (thumbs u

 

Thanks, Jimbo. It's been a matter of finding a decent way of presenting it. The process has given me a lot of insight into how the same dealer can grade different types of books differently. There are definite trends. For example, Pre-code Horror books are routinely overgraded more often than superhero books. I assume dealers grade on a curve for books that are less common in high grade.

 

..... another slight inaccuracy that could occur here is with the nature of your sample group in it's ratio to the total field of possibilities..... to clarify.... your focus is high grade where the point spreads from 8.0 to 9.0 are 10 points, though only 2 increments..... which is unavoidable. As most of us probably know.... most overgraded NM's will actually fall into this range, with those retailers being more harshly penalized for their lack of consensus with CGC.....from an increment standpoint. I apologize if I'm overthinking this.... but this type of stuff fascinates me. GOD BLESS...

 

-jimbo(a friend of jesus) (thumbs u

[cheetah]

That's alright. It's not definitive but the trends with the higher sample sizes seem pretty clear.

 

I'll post the vendors tomorrow night. Maybe between now and then, folks can take a few minutes to list the vendors from most to least favorable.

[Johnny545]

I'll bet a dollar I know who dealer 3 is...