I’ve considered this in the past, but then you’re still left wondering what the true variance between graders is.
Personally I just submit everything I have when there’s a special worth using. First sub ever was 142 cards, I’ve had a few in the 40-70 size, smallest was 23, and I just got back my largest at 172. I’ve never noticed any crazy variation in grading to be honest.
As far as cards I’ve cracked out and resubmitted, I have about 80% on those being upgraded, but the sample size is so low I wouldn’t put much stock in it.
With no evidence to support it, I feel that PSA typically doesn’t want to give you a submission with all 10s. My pregrading strategy is to aim for about 80% 10s. I’m definitely including questionable cards in my submissions because 1) I could get lucky and variance swings in my favor, 2) they make the better quality cards look even better by comparison, and 3) it gives PSA some fodder to hand lower grades out within my submission, if that’s actually their policy. I am happy with my results operating this way.
As someone who used to make a living managing risk my brain has never let write down a single number when I pre-grade. I’ve always done so with a two grade range. Eg worst-case it will score an 8, best-case it will score a 9.
Given that approach I can only recall one card I’ve graded where I was surprised with the results (I’ve probably graded less than 200 cards though).
With that being said I’m preparing a submission right now so I’ll make myself write down a single grade to see how accurate I am and report back.
The problem is the submission has 5 unique cards but 9 copies of each all in virtually identical condition so if I’m wrong about 1 I’m probably wrong about 9
Side note - If I were submitting to a company that hands out half grades I think this exercise would make my head explode
I don’t see the logic behind the idea that PSA does not want to give a submission all 10s. They have a grading scale and they no doubt want their graders to follow that scale as best as possible. Call me naive but I cannot envision a scenario where training of graders involves explicitly instructing people ‘oh by the way don’t go giving out too many 10s on a specific order’. I can barely even think of a reward for doing that to even consider outweighing the risk of that information getting leaked and the fallout that would involve.
However, I agree with the idea that maybe seeing near perfect condition cards over and over again can subconsciously make a grader start over-focussing on things and being more critical with their grades. I would imagine that PSA tries many things to avoid human psychology and bias from interfering with grading, but it would be near impossible to completely avoid little things like that.
Employees are paid not-very-well, since grading is a low skill job.
Employees presumably have some sort of quota, or at least some performance metrics by which they are evaluated. Every card has a certification number and it would be trivially easy to link that to which employee graded the card. It is known that PSA has cameras throughout their entire operation to “review the footage” and likely for insurance purposes, since they strictly deal with other peoples’ property and sometimes it can be quite valuable.
There is a difference between explicit work place training and work culture. If you are one of many workers in a low skill situation like PSA’s TCG bulk grading, you do not want to stand out. Give too high of grades? Stand out. Give too low of grades? Stand out. It costs the company time to go manually review your work to see if the employee is “applying the grading scale correctly.” It also makes more work for somebody (the bulk grader’s superior) to do, which makes their mood worse.
There are lots of social pressures that could exist, intentional or not, that would incentivize employees to sprinkle in some lower grades in most submissions. Anecdotally, that also seems to align with most peoples’ experience on getting their return.
Card grading is subjective. I think we can agree on that. Most of us who have submitted to PSA probably have stories of regrading cards that got 9s the first time around and then got 10s the second time around.
My issue is when you start putting together theories which basically say “PSA employees are giving some cards artificially low grades instead of evaluating cards in line with the scale so that they can adhere to a grade quota.” We don’t need conspiracy theories when simple subjectivity suffices.
Let’s say I submit 100 brand new, pack-fresh Japanese modern cards that I looked through and ensured were all in great (but perhaps not perfect) condition. When I get the submission back, I got 90 PSA 10s and 10 PSA 9s. I take those 10 PSA 9s, crack them, and resubmit them. 4 out of the 10 come back PSA 10.
Did PSA artificially lower the grades in the original submission, or were the original 10 9s simply cards that were on the line between a 9 and a 10, and arguably graded “correctly” each time?
You could make the same argument about cards that grade higher than expected, but nobody ever seems to argue that bulk graders have a pressure to give out 10s in every submission to keep their numbers in line, and that’s why they got all the 10s they did. No conspiracies about overgrades exist, everyone just happily accepts them because it benefits them.
I do believe that the bulk graders are pressured to give grades roughly on some distribution, which includes 10s to cards that maybe shouldn’t get them. I have bought cards that were already graded PSA 8, resubmitted them in bulk, and they come back a 10. That’s a pretty big jump to say “subjective.”
Without ever having seen it done, my impression based on how I pregrade, what I’ve submitted, and the grades I’ve received is that most mint cards are examined for <10 seconds, then given a grade 8-10 with a stochastic process. Like you said, it’s harder to measure overgrading, because people pay less attention to it. However, it does seem that “10 quality” cards only actually receive a 10 as low as 80% of the time. It seems reasonable that “9 quality” cards would be overgraded a similar up to 20% of the time too.
PSA extracts additional value from customers by having some randomness in their grading process, especially when it can be slightly spread around throughout a submission. Each grade is not independent. I think that is the key for people to understand. Grading is done in batches, and humans are generally not very good at objective measures, but are much better at comparative measures. It’s easy to tell that a 10 lb object is heavier than a 5 lb object, but very difficult to tell the weight if handed a random object and questioned.
Every card has a certification number and it would be trivially easy to link that to which employee graded the card.
Yes so why would you want to have cards on your resume of graded cards that are clear migrades? Submissions don’t stay together once they’re returned.
it costs the company time to go manually review your work to see if the employee is “applying the grading scale correctly.”
If there is some form of tracking of grades given by each employee, then it would be trivially easy to see a distribution of any person’s grade allocation distribution.
From a mathematical standpoint, if you’re intentionally lowering some grades from big distributions of good quality cards, then you’re going to trend away from the ‘ideal distribution’ (which would I suppose be the mean distribution across every employee) unless you start balancing that by giving out intentionally higher grades to submissions of poor quality cards.
Also, and I don’t know for sure about this so correct me if I’m wrong, but isn’t each card supposedly reviewed by another person before being finalised? So you would be more likely to get caught out for poor grading if you started arbitrarily lowering some grades.
This is just pure conspiracy. No way are cards anywhere from 8-10 simply lumped together and graded on a random distribution. If it were, grading would be impossible. As mentioned in this thread, there are people who are very good pregraders because they know the scale well and it is applied (mostly) consistently.
That being said, I fully agree with your assessment that PSA graders likely spend less than 30 seconds per card on bulk orders. Given this, I can’t seem to understand how you reject the very basic and plausible explanations that variance is determined by subjectivity, or these bulk graders that spend maybe 15-20 seconds per card missing flaws or thinking a flaw is there when there really isn’t one.
Don’t attribute to malice conspiracy that which is easily attributable to incompetence (and subjectivity).
A thought experiment, but let’s say I gave you 1,000 cards of the same card in various conditions, and asked you to spend as long as you needed grading them, and then you do the same process a few months later to the same cards in a shuffled order, how evenly do you think your two batches of grades will match up?
There isn’t a grading quota to meet or have to give a certain amount of a specific grades to every sub. That is such an absurd statement to make and just speaks to a person’s inexperience with grading and with Pokemon in general. I’m saying this in the most civil and non insulting way possible lol.
Just throwing in some anecdotal experience:
I’ve been collecting since 1999 and started grading WOTC holos regularly in 2011. In the beginning, I was extremely selective on which cards I sent to grade. Most 1st Edition WOTC holos were only about $20-30 in a PSA 9 and PSA 8’s almost always lost money. Even 1st Edition Base holos were under $100 in PSA 9 . . . I picked-up my Blastoise for $80. Those were the days!
Over the last 13+ years, I’ve graded thousands of WOTC holos, working on completing PSA 10 holo sets for my personal collection. In that time, I’ve experienced mostly consistent grade ratios . . . occasionally a weird submission (ten PSA 7.5’s in one submission). Also, the population reports for most of the difficult-to-grade cards have increased at a fairly consistent rate over that span. Some of these cards are cards I still need, so I’ve been regularly watching their population reports and their overall gem rates: most of these cards have had Gem Rates below 10% for as long as I remember . . . it’s not a new phenomenon. Most of these Gem Rates have gotten even stingier in the last 3-5 years, but I suspect that’s because lower-grades are now profitable too and people are willing to resubmit their 8’s and 9’s hoping to hit a homerun on a PSA 10.
Some cards are especially susceptible to holo print lines, making them almost impossible to grade. Often, these cards aren’t even popular Pokemon: Dark Magneton, Yanma, Misdreavus, etc. Occasionally, a few have snuck through and received 10’s undeservedly, but not often (even years ago). I do suspect that PSA is far more aware of card values now and are more thorough than they were in the past when these cards weren’t worth much though. Also food for thought: many of these same difficult-to-grade cards also have Gem Rates below 10% with CGC.
Occam’s Razor: Maybe, some cards are just difficult to grade . . .
That’s exactly what rational employees should do. The easiest way to keep your distribution stats good is to always grade that distribution. Pretend the scale is only 8, 9, 10 and the ideal distribution is 80% 10, 10% 9, 10% 8. If I give no more than 8 10s every 10 cards I grade then I’m forcing myself into the ideal distribution regardless. I’m saying the employees are poorly compensated, not that they’re stupid. Their best interest is best served by making their superior happy. You don’t do that by objectively following your company’s publicly stated goal. You do that by playing the internal metagame of whatever metric by which you are evaluated.
I’ll just say that from a psychological perspective, if you want to give off the impression of being a “fair” grader, you will want to not grade entire submission as PSA 10. Even if it deserves it. The same way asking a human to give you a set of random numbers, they will avoid repeating numbers in order to maximize perceived “randomness” even if that’s not how a true random function works.
When grading a bunch of 10 candidate cards in a row, what should be an independent assessment of each card becomes coorelated: “well this copy looks a touch weaker than the previous copy”. The “objective” grading scale is then calibrated to the quality range of the submission, leading to more critical grading overall.
I actually spike my submissions with a few dud cards, maybe some miscuts or things I’m ok with getting a 9 on or just cards I want slabbed even if they are PSA 6. Just to give the grader some psychological relief and to help recalibrate on the larger grading scale than just the 9-10 boundary.
This is pure speculation. No evidence that it works. But I’ve yet to get a submission back from PSA where I didn’t think it was generally correct. Especially in contrast to grading with CGC before their label change.
It was very weird. It was like the PSA 7-10 grade range was squished between CGC 8.5-9. CGC to PSA Cross-Grading Result ← the average difference for me was half a grade which is pretty huge, especially when you compare CGC 8.5 prices to PSA 9 or CGC 9.5 prices to PSA 10
Agreed, a note about new label cgc, I have noticed from me cracking and cross grade new label cgc to psa that its a
"take what it would get at PSA and add 1”
I don’t have a big enough sample size but majority(<20) of new label cgc cards I’ve graded at psa have been downgraded. (i bought the card and not the grade, with a bonus being picking up cgc 10’s for less than a price of a psa 9 or cgc 9 for less than raw price.)
