Not a CBCS member yet? Join now »
CBCS Comics
Not a CBCS member yet? Join now »
CBCS GradedComics Restoration and ConservationGrading Help NeededQuestions

CGC will now allow CBCS slabs to be regraded/retagged with CGC labeling....7707

Moderator Jesse_O private msg quote post Address this user
@shrewbeer Thank you for your response. That does give me more confidence in your system. However, I still have questions.

"Categories are used rather than search terms. Far too many people spam their listings with "cgc CBCS pgx""

I totally agree that too many people spam their listing and that is my concern with this data. But people still spam by categories. They will check the certification even when it is a raw book in order to get their books seen.

"Large samples are then taken of each category. These are scanned by eye, looking at the slab picture to ensure it is properly categorized. This past month sample size was 2%, 50%, 25% (CGC PGX CBCS)"

Who takes these samples and looks at them? I understand that CGC is the largest listing, but I hardly think that a 2% sample size is statistically significant.

"From there we can then extrapolate an initial accuracy rate. For example, in the past data set, it was 88% for CGC, 95% for PGX, and 98% for CBCS (this tells me people who list CBCS books are more honest and accurate than others lol)"

This is my major concern. The accuracy of the extrapolation based on the sample size. According to this 88% of all CGC listings (with a 2.8% MOE) is accurately listed. I find that very hard to believe. I know my sample of Hulk 181 is skewed because it is a very popular book to spam, but that one listing was 41% accurate. I just spot checked "God Country 1 CGC" 58 out of 81 are labeled correctly (72% accurate).

I just don't see how a 2% sample size is significant enough to reliably extrapolate from. And how is that sample taken? Is it just a price range? Do they choose DC or Marvel?

Having said all that, I do appreciate the explanation tremendously!!! This is a huge undertaking by you and I commend you for it. I hope you do not take my critique personally. I am happy that someone is tracking this and I believe that it will prove invaluable in the future. I just want to make sure that the process is accounting for everything.
Post 26 IP   flag post
COLLECTOR shrewbeer private msg quote post Address this user
Quote:
Originally Posted by Jesse_O
I hope you do not take my critique personally.

Absolutely not, more input from more folks gets me to look at the system and make it better.

To your various points,

- Yes people still spam their categories, but less than search terms. This is why they have to be looked at regardless

- I have been doing the sampling myself. You would be surprised at how accurate a 2% sample can be. Think about political polling. They can survey one thousand people out of millions and manage a small margin of error.

- The 2% will build upon itself over time as well. We can consider our past samples as valid data to help out. I'd like to see it at 10%, but that's a bit more time than I have (that equates to looking at twelve thousand listings on the overall available). The past month, 2,500 were looked at. They are looked at random. For example, If I have it filtered to 50 listings per page with 2500 pages total, I'll look at one full page and skip by 50 pages for a random selection of listings.

Your spot checks for god country and hulk netted you a 56% average, but that was by search terms. Search Hulk 181 within collectible comics, filter to categorized as CGC, and look at the results. Several are not 181s, but we don't care about that; its wether or not they are categorized correctly as CGC books. 96.5% accuracy right there for CGC books. Only ten out of the pile of 289 were not CGC books.
Post 27 IP   flag post
Moderator Jesse_O private msg quote post Address this user
@shrewbeer
"You would be surprised at how accurate a 2% sample can be. Think about political polling. They can survey one thousand people out of millions and manage a small margin of error."

Yet we all know how well THAT worked out two years ago!!! LOL!!!

Sorry, had to say it.

Now that I understand your methodology, I appreciate this report even more. How do you sort them? eBay automatically defaults to best match. If you start at the beginning and go from there, the worst matches (presumably raws) would be at the absolute rear of the list. If you want, I will offer my services to you. Perhaps if we coordinate on sampling CGC listings (so we can follow technique but not duplicate each other), we can improve the sample size and get an adequate size.

Again, I advise caution in interpreting this data until there is a large data base to go by. In my opinion, one full year is an absolute minimum, but two years worth will really start to tell the story. And I'm willing to help you do the grunt work.
Post 28 IP   flag post
COLLECTOR shrewbeer private msg quote post Address this user
@Jesse_O cool, I'll be in touch next month when it's time for another run of the data.

No need to sort the listings by best match or anything. You can see every single book listed without doing any search at all.

Good joke though lol. Politics sucks
Post 29 IP   flag post
Collector Homer private msg quote post Address this user
I think your data does have merit, the problem taken out of context it is the personality of the sellers who are selling CBCS, CGC or PGX books on Ebay. There are many aggressive dealers using CGC to move books - flip, cash, and repeat. Thus a larger sell through ratio for CGC. While some CBCS sellers or collectors do the sit and wait, no risking with auction, and waiting for the right price is lengthy. The poor turn around time by CBCS has self eliminated themselves from the dealers who are moving books as flip, cash, and repeat.
Post 30 IP   flag post


"Forum Overlord" bah ha ha ha... JustThatGuy private msg quote post Address this user
Personally, in my opinion, strictly my opinion, I don't really care how the market does and who's selling more. I collect both CGC and CBCS. Yeah its a terrible concept, I know. I am not a dealer or slab dealer. Not that I haven't sold books before, I have sold a few just to buy better ones. The amount that I've sold to buy is like .01% sold to 99.99% buy. I love the hobby. Before CBCS, I used CGC and I still do. They both offer something that the other one doesn't( and both offer everything that PGX doesn't). This is not a pissing contest, each side is simply doing what they can to strive for the best to what they offer for their customers. So boys, beer up. And if you are in Texas, beer up cause you know you can't do anything outside cause its raining.
Post 31 IP   flag post
Collector moodswing private msg quote post Address this user
So I came across a video describing this and the YouTube said it isn't much of a deal. I interpreted the deal to be their discount tier pricing like modern at $12 per slabbed and if it gets a lower grade it will only cost $10. He states it is the tier pricing plus the extra $10. Little confused here. Could someone clarify what the deal actually is. Makes no sense getting a modern slabbed since it would cost more with the tier discount price plus the $10.
Post 32 IP   flag post
Ima gonna steal this and look for some occasion to use it! IronMan private msg quote post Address this user
Quote:
Originally Posted by shrewbeer
Quote:
Originally Posted by IronMan
But anyway...for sell through to be helpful and have meaning we need stats for the same sellers that have copies of the same books in the same grades at the same prices and same selling format at the same time by the different grading companies....


You're overlooking sample size here. These sales are in the thousands, and in the case of CGC, in the tens of thousands. When you get that large, any outliers are lost in the mean.


That's just not true. It doesn't matter how large your sample size is (unless it is 100% - the entire universe) if the groups being sampled are over or underrepresented. This is why political polls - which you mentioned - are carefully weighted for demographics (race, sex, region, political affiliation, age, cell phone and landline)to match up with general population. The better polls also look at voting history - people that voted the last election get more "weight" than people that did not.

For all you know, your sampling may be over weighting CGC in moderns. It might be over weighting large dealers. Or flippers. Perhaps CBCS sellers - on eBay - tend to be newer sellers, perhaps they tend to sell older, slower moving books. Things entirely possible. The only way one could flip hot new releases in CBCS slabs would be to pay for 2 day modern. Otherwise CGC Fast track is going to be the preferred grading service for that crowd. And that appears to be a very large crowd - on eBay.

Pull your data exclusively from Comic Link, Comic Connect, Heritage. Maybe even MCS. You'll need to weed out bat sh*t crazy asking prices. But your data from eBay encompasses way too many variables that you cannot account for.
Post 33 IP   flag post
If the viagra is working you should be well over a 9.8. xkonk private msg quote post Address this user
Quote:
Originally Posted by Jesse_O

Yet we all know how well THAT worked out two years ago!!! LOL!!!

Sorry, had to say it.

Since it's about that time again, I'll just point out that the polls did a fine job two years ago. They were actually better than 2012. Some people definitely interpreted them poorly though. https://fivethirtyeight.com/features/why-fivethirtyeight-gave-trump-a-better-chance-than-almost-anyone-else/
Post 34 IP   flag post
COLLECTOR shrewbeer private msg quote post Address this user
Quote:
Originally Posted by IronMan
Quote:
Originally Posted by shrewbeer
Quote:
Originally Posted by IronMan
But anyway...for sell through to be helpful and have meaning we need stats for the same sellers that have copies of the same books in the same grades at the same prices and same selling format at the same time by the different grading companies....


You're overlooking sample size here. These sales are in the thousands, and in the case of CGC, in the tens of thousands. When you get that large, any outliers are lost in the mean.


That's just not true. It doesn't matter how large your sample size is (unless it is 100% - the entire universe) if the groups being sampled are over or underrepresented. This is why political polls - which you mentioned - are carefully weighted for demographics (race, sex, region, political affiliation, age, cell phone and landline)to match up with general population. The better polls also look at voting history - people that voted the last election get more "weight" than people that did not.

For all you know, your sampling may be over weighting CGC in moderns. It might be over weighting large dealers. Or flippers. Perhaps CBCS sellers - on eBay - tend to be newer sellers, perhaps they tend to sell older, slower moving books. Things entirely possible. The only way one could flip hot new releases in CBCS slabs would be to pay for 2 day modern. Otherwise CGC Fast track is going to be the preferred grading service for that crowd. And that appears to be a very large crowd - on eBay.

Pull your data exclusively from Comic Link, Comic Connect, Heritage. Maybe even MCS. You'll need to weed out bat sh*t crazy asking prices. But your data from eBay encompasses way too many variables that you cannot account for.


The point is that it doesn't matter. I'm not looking to start a new GPA site. I'm simply offering three categories of data to my group members, which they can interpret however they like. Total overall sell-through rates, total market share, and volume over time compared to themselves (ie how the company is in general performing in sales against it's own previous months).

If you want to know why each data point is how it is, or know for sure other conclusions one may draw from the data, that is entirely up to you to find out.
Post 35 IP   flag post
I bought a meat grinder on amazon for $60 and it's changed my life. kaptainmyke private msg quote post Address this user





Post 36 IP   flag post
596795 36 11
This topic is archived. Start new topic?