Sports Card Analytics

Cuts Like a Knife – Part 3


Graded 1986-87 Fleer is like the reserve currency of the basketball card collecting world.  Of the 9,000 listings sold in eBay around 4,000 of them were in PSA slabs.  BGS and SGC accounted for around another 1,000 total.  Graded 1986-87 Fleer is one of the most liquid card assets a collector can have.  Over 30 years after release it is still in high demand, especially if the card is considered to be in outstanding condition.  A sample of the 24 most expensive graded card listings from pwcc_auctions over the last 3 months brought in a total of over $240,000.  There is a lot of money changing hands for some very popular cards.

In the last 90 days over 9,000 listings for 1986-87 Fleer basketball cards have been purchased on eBay.  Over 2,000 of those listings have sold for over $100 and 500 of those have sold for over $1000.  On average one 1986-87 Fleer listing sells every other day for over $5000.  Yes, in 90 days over $1 million worth of 1986-87 Fleer cards have sold on eBay. 

How does that compare to the number of completed listings for some other iconic Basketball card sets? 

1961-62 Fleer:  Over 1,200 listings, 64 over $1000

1980-81 Topps:  Almost 2,000 listings, 30 over $1000

1996-97 Topps Chrome:  Over 600 listings, 23 over $1000

2003-04 Topps Chrome:  Just over 1,000 listings, 118 over $1000

Graded 1986-87 Fleer is like the reserve currency of the basketball card collecting world.  Of the 9,000 listings sold in eBay around 4,000 of them were in PSA slabs.  BGS and SGC accounted for around another 1,000 total.  Graded 1986-87 Fleer is one of the most liquid card assets a collector can have.  Over 30 years after release it is still in high demand, especially if the card is considered to be in outstanding condition.  A sample of the 24 most expensive graded card listings from pwcc_auctions over the last 3 months brought in a total of over $240,000.  There is a lot of money changing hands for some very popular cards.

If you’ve been following my blog for the last few weeks, you know that I’ve been intrigued by the recent developments brought to light on Blowout Forums about card trimming, especially with cards that have been trimmed but end up in graded card holders.  There is compelling visual evidence on some serial numbered cards that suggest trimmed cards do reside in PSA and Beckett holders.  Joe Orlando, President and CEO of Collectors Universe, Inc. (the parent company of PSA), posted a brief article in 2004 where he discusses card trimming.  Mr. Orlando claims the “awareness level of his experts” is why people “have faith in the PSA process.”  Those that trim cards and sell them online are potentially committing mail fraud.  Mail fraud can come with a penalty of imprisonment of up to 30 years for each offense

I suspect there are a combination of things that must be true for a trimmed card to end up inside a graded card holder and then made available for purchase.  First, the individual responsible for trimming the card must expect there is some financial advantage to be gained by trimming the card.  Secondly, the trimmer must be willing to take the risk of being found guilty of committing mail fraud and having to pay a penalty or serve time in prison as a result of their actions.  I can’t explain the rationale that makes one do something criminal, so I won’t focus on explaining that element.  I can provide some thoughts and insight into the financial side of things.  A trimmed card, generally, will be valued at a fraction of the price of an unaltered card.  For example, this 1986-87 Fleer Michael Jordan card designated as altered by PSA sold for $580.  Let’s say, hypothetically, that particular card before alteration met the grading criteria specified by PSA as NM-MT (8).  A copy of the same Michael Jordan card, graded PSA 8 and with similar centering, recently sold for $1820.  The alteration in this scenario accounts for a reduction in value of around 68%.  Now, let’s run the scenario where the owner of the trimmed card successfully managed to get the card past the experts at PSA and into a PSA 9 holder.  A recent sale suggests a PSA 9 Michael Jordan rookie card brings in about $4000.  The card as a PSA 9 is now valued at a premium of about 120% above the PSA 8 card.  Let’s say our hypothetical card trimmer submits 5 of these trimmed Jordan rookie cards to PSA.  If 2 of them end up in PSA 9 holders and the other 3 are slabbed as altered, then the card trimmer may break even financially after grading fees and selling fees.  To make this exercise financially worthwhile, the trimmer needs to ensure better than a 40% success rate of obtaining a PSA 9.  Is the “awareness level” of the experts good enough to catch more than 60% of trimmed cards?  I would hope it is, assuming there is nothing unscrupulous going on with PSA.  If the trimmer can’t succeed 40% of the time I suspect the financial feedback loop would inform the trimmer that this is not a fruitful exercise.

Let’s build out this scenario a little further.  We’ll continue with the Jordan rookie as the test case.  Let’s say the card trimmer has a PSA 9 Jordan rookie card, valued at $4000.  The card has a slight chip on one of the edges or a tiny white mark on one of the corners.  If the trimmer cracks it out of the case and trims off the edge with the white mark or removes just enough to make the corner look sharp, it may have the appearance of a gem mint card.  A PSA 10 Jordan rookie currently sells for around $21,000.  Let’s do that same math as before.  If the card goes from PSA 9 to PSA 10, then the trimmer increases the value of the card by a whopping 425% above a PSA 9.  If the card is slabbed as altered, it’s a loss of about 86%.  The big disparity in price between a PSA 9 and PSA 10 makes this a much better proposition for the card trimmer.  For the trimmer to come out ahead financially they need to get 20% of these hypothetical trimmed Jordan rookies past the expert graders.

The Jordan rookie is such an expensive card.  Maybe the benefits of achieving a PSA 10 as opposed to a PSA 9 isn’t as great for other cards?  Let’s scale it down to a card that is of much lower value, the 1986-87 Fleer Darrell Griffith #42.  Doctor Dunkenstein was a heck of a player.  He ranks #4 on the Jazz all-time scoring list.  No one is going to confuse him for Michael Jordan though.  A 1986-87 Fleer Darrell Griffith in PSA 9 sells for around $19.  As a PSA 10 the same card sells for closer to $100.  I find it interesting that the same 5x multiplier between PSA 9 and PSA 10 applies for both Jordan and Griffith.  If a trimmer submits 4 Griffith cards and sees 1 receive a PSA 10 and the other 3 caught as trimmed, then the trimmer comes out close to even financially.  The thing that has a bigger effect in this scenario is whatever the grading fee is for the customer.  If part of a large submission, accompanied by discounted rates, then the grading fee becomes less of a consideration.  So, whether the card is fundamentally expensive or cheap, the trimmer needs to have trimmed PSA 9 card turn into a PSA 10 about 25% of the time to make it at least tolerable financially.

The third requirement for a trimmed card to end up in a graded card holder as unaltered is for it end up in a graded card holder with a grade assigned.  PSA and Beckett rely on humans to determine authenticity, an unaltered composition, and detect and rate all flaws of a sports card.  Many industries in the world rely on humans to do 100% visual inspection.  There was likely a 100% visual inspection on at least one element of the car you drive, or the airplane you last trusted to keep you aloft in the sky, or the piece of medical equipment that may have saved your life.  However, 100% visual inspection is not 100% effective if performed by humans.  We’re beings who can get easily get bored or distracted or have a bad day.  Sometimes our minds wander.  Maybe we have a hangover, or we didn’t sleep well the night before.  A study by Jospeh M. Juran, a quality management guru, placed the effectiveness of 100% visual inspection at 87%.  Others place the effectiveness closer to the 80 to 85% range.  That means somewhere between 13% and 20% of the time we can reasonably expect a fault to get past a single inspector.  These studies help to confirm that inspection is by no means foolproof.  I don’t believe that the experts at PSA or Beckett are superhuman and infallible.  Many industries serving much more critical applications than grading sports cards suffer from this same inspection inefficiency.  Human error isn’t the only possible reason that a trimmed card could end up in a graded card holder, but for the purposes of this exercise we’ll continue to assume human error is the source of trimmed cards having grades assigned.

The PSA website does detail their grading process.  PSA reportedly attempts to address the ineffectiveness of visual inspection by having, at a minimum, two graders reviewing a card.  If the two graders agree on the card grade then the card is assigned that grade.  Assuming a 13% ‘failure rate’ for identifying a flaw for each grader, we can estimate the probability that two graders will make an error on the same card.  Let’s say a trimmer submits 100 trimmed cards for grading.  Grader ‘A’ performs as expected and makes an error on 13 of the 100 cards.  Grader ‘B’ then independently looks at the 100 cards and makes an error on 13 of the 100 cards.  I ran a Monte Carlo simulation to determine that on average with a 13% miss rate the two graders will make an error on the same card around 2% of the time.  So, for a submission of 100 trimmed cards, we might could expect that on average 2 of the trimmed cards would not be identified as such by either grader.

The Beckett grading process is described here.  Beckett’s description of their process only mentions a single grader acting as a reviewer for each card.  Many collectors feel BGS and PSA are more or less equivalent.  But, the details of their grading process suggest that’s not the case at all.  Remember the error rate for a single inspector?  It was between 13% and 20%.  That is exactly the scenario Beckett describes as part of their process.  They’re running a process we might expect to miss at least 13% of the things they get paid to evaluate and catch.  With a minimum of two graders PSA theoretically drops that error rate to 2%.  If you’re a collector, advantage PSA for using two graders on each card.  Based upon this analysis, and assuming the processes published online by each company are accurate, it makes more financial sense for a trimmer to send their cards to the single grader at BGS than it does to send them to PSA where two graders will evaluate the trimmed card.

PSA does offer a financial guarantee of grade and authenticity.  I was unable to find any guarantee with regards to authenticity or grade on the Beckett Grading FAQ page.  So, if you happen to have purchased a trimmed card that was graded by PSA, you may have recourse for recovering your losses.  Buy one in a BGS holder and you may be out of luck.  Again, as a collector, PSA seems to have an advantage over BGS in this regard.  I wonder if PSA has an insurance policy that covers the cost of their financial guarantee of grade and authenticity or do they cover that cost out of an allocated budget?  In the Collectors Universe 2018 Annual Report there is no mention of an insurance policy.  On pages 31 and 32 you can find that Collectors Universe has what they call a warranty reserve.  The amount in that reserve as of June 30, 2018 was $862,000, which is intended to cover guarantees for coin grading and trading card grading.  If you’re interested in knowing more about the inner workings of PSA, you should read over their annual report sometime.  You will learn interesting things such as the fact that 16% of Collectors Universe’s revenue comes from just 5 customers.

Would $862,000 sufficiently cover the costs of PSA’s grade guarantee should there ever be a finding of large quantities of trimmed cards residing in PSA holders?  Over 1.7 million trading cards were graded by PSA in 2018.  From 2016 to 2018 they graded over 4.5 million trading cards.  Time for another quick math exercise!  Let’s say the average financial risk to PSA for having a trimmed card assigned a grade is $100.  Based on the Michael Jordan and Darrell Griffith examples, I am guessing that $100 is not an over estimate.  Find a trimmed Jordan and the claim should be for around $20,500.  That one Jordan card would consume 2.4% of the warranty reserve.  Find a trimmed Darrell Griffith and the claim is likely closer to the $100 mark.  Let’s also assume all $862,000 is available to be applied to guarantees associated with trading cards.  Trading cards only comprise about 37% of Collectors Universe’s business.  What percentage of the 4.5 million cards graded between 2016 and 2018 would it take to consume the entire $862,000 warranty reserve at a cost of $100 per guarantee claim?  That number would be 0.19%.  It doesn’t look like the Collector’s Universe warranty reserve would go too far in a crisis.

Let’s focus a bit more on why trimmed cards end up in PSA holders.  A 2% success rate for the card trimmer isn’t very high.  At a 2% success rate the reward of a PSA 10 does not warrant the risk of diminishing the value of a card by trimming it.  Even with the potential to resubmit repeatedly, a 2% success rate makes it a losing proposition.  But, of course, card trimming happens and these trimmed cards do end up in PSA holders.  So, what’s happening?  Unfortunately, I don’t have the answer.  Also, I can’t provide data supporting the frequency trimmed cards end up in PSA holders.  But, I can offer some speculative scenarios that could be possible.  I’ll pose each scenario as a question instead of running the risk of these being considered declarative statements.  Again, I don’t know the answer.  This section is all about brainstorming and running thought experiments.

1)      Perhaps trimmed cards in PSA holders is an uncommon occurrence?  This is definitely a possibility.  I have no data to support what percentage of cards in PSA holders are trimmed.  A statistical analysis could be performed to generate an estimate of the pervasiveness of this problem.  The study would rely on dimensional measurements of many cards that currently reside in PSA holders.  One would need to crack those cards out of the PSA holders and measure them.  The probability of a given card possessing the dimensions measured could be calculated by comparison to the mean and standard deviation of measurements coming from a population of cards that are known not to be trimmed.  For most of us this exercise is not practical.  Few would be willing to crack out a 1986-87 Fleer Michael Jordan PSA 10, or even a Darrell Griffith PSA 10, in order to measure it and find out there is a high probability it is trimmed.  That’s a lot of financial risk to take, especially since the PSA guarantee would then be void.  That’s exactly what makes this such a difficult and troubling issue.  There is no incentive to find out if a card is trimmed once the card makes it into a holder.  Detection of a trim job once a card makes it into a graded card holder is discouraged by pretty much every feedback system.

2)      Maybe trimmers are ignorant of the losses they’re likely to take over time by submitting trimmed cards to PSA and a smattering of trimmed cards pass through the grading expert filters?  Never underestimate the capacity of humans to do irrational things.  This could be so.  But, the trimmer is also a schemer.  This personality types generally makes and executes plans that have been thought out at least on some level and then adjusts if things aren’t working.

3)      Could PSA’s practice of assigning at least 2 independent graders to each card no longer be part of their service?  Demand is high and lead times are running longer and longer.  What if only 1 grader were assigned to each card?  The 13% to 20% expected error rate by graders would then be much more in line with what a card trimmer would need to achieve in order to be financially successful.  This would also put the PSA process on par with their largest competitor.

4)      What if the graders at PSA are much less effective than research suggests we might find with 100% inspection?  Perhaps the workplace is incredibly distracting or there is something in the work culture at PSA that leads to disengaged employees.  Maybe whatever training the graders receive is insufficient and it doesn’t provide the skills necessary to spot trimmed cards effectively?

5)      Could the card trimmers have such advanced trimming technology that it’s nearly impossible to spot a trimmed card visually?  This scenario reminds me of the reports about performance enhancing drugs in sports.  In sports, it seems the drug developers are always at least one step ahead of the latest drug detecting capabilities.  If there is an advantage to be gained it likely warrants an investment.

6)      Could the card trimmers have developed technology that allows them to remove unaltered cards from PSA cases and seamlessly and undetectably replace them with trimmed cards?  In this case, PSA never actually sees the trimmed card.  The ‘switch’ is done after the slab is in the wild.  This one just feels unlikely.  When molecules are fused, as happens when plastic slabs are welded closed, they rarely are separated and come back together again without some trace of the process imposed on them.

7)      Is someone manufacturing card holders and flips with fonts that are identical to PSA and then placing their trimmed cards in the holder?  Counterfeit cards and other products have been around for a long time, so this is feasible.  However, to have the cert number check out the trimmer would need to find the cert number for a version of the card that possesses the same grade.  Then no one would need to notice that there are two cards out there with the same cert number.  If there is a 1 of 1 out there that has been trimmed and it’s in a PSA holder and the cert checks out, that would remove at least some of the potential around this solution.

8)      Could card trimmers be bribing graders to ensure their cards receive special treatment and end up in high grade holders instead of marked as altered?  Bribery is common throughout the world and considered just a regular part of business in some. Maybe there is a group of card trimming lobbyists contributing to the vacation funds of the graders?  Based upon salary reports at Collectors Universe on the website Glassdoor, I don’t suspect the compensation for graders is phenomenal.  Compensation shortcomings could be further exasperated by the high cost of living in the Santa Ana, California area where Collectors Universe headquarters are located.  Money is a powerful thing. 

9)      What if PSA (or even BGS) graders are instructed from the highest levels of the company to knowingly grade trimmed cards and place them in high grade holders?  The business models established by the grading companies doesn’t exactly discourage this type of behavior.  Demand for grading appears to mostly be driven by those looking to gain financially from receiving a high grade.  The more graded cards that sell for high prices the more attractive grading is to those hoping to use card grading as a vehicle for financial benefit.  The more people that want to grade cards the more orders the grading company receives, each one accompanied with a payment for the service.  It would be an effective marketing tool, but one that comes with a risk of getting caught and destroying any confidence held in the services provided.  Or maybe the risk to the company isn’t that big?  Remember scenario number one on this list?  It isn’t likely that someone will detect that a card is trimmed.  If the owner of the card can’t provide irrefutable evidence of trimming all of this becomes a matter of ‘expert opinion’ versus the suspicions of the owner of the card.  As the one who drafted the terms of the guarantee policy and the employer of the ‘experts’ the balance of justice likely tilts in favor of PSA.  Finally, how many companies have been accused of wrongdoing only to come out the other side effectively unscathed?  How are Facebook, Apple, Volkswagen, Wells Fargo, Samsung, Equifax, and Uber (amongst many, many others) doing these days?

 Is any of this a problem?  Do collectors care if cards are altered?  The card grading industry was built on the premise that collectors care about the authenticity and condition of cards.  Let’s challenge that assumption for a moment.  Maybe collector sentiment has changed.  At one time almost no one would climb into a car with a stranger and have them drive them from place to place.  But, the ride-hailing industry has emerged and it is expect to grow to over $285 billion by 2030.  Ask people from even 10 or 20 years ago if audio surveillance in a home would be acceptable.  The answer would almost certainly be “no way.”  Today it can be difficult to find a new TV  that doesn’t have a microphone built into it and isn’t connected to the internet.  We’re actively placing audio and video surveillance technologies in our homes.  So, it wouldn’t be unprecedented to have humans change their mind.  Maybe condition and authenticity are concepts that are outdated.  The recent 1997-98 Metal Universe Michael Jordan PMG Green auction that sold for over $350,000 was an altered copy.  So, at least a couple of high-end collectors didn’t really care.  If we all collectively decide condition and authenticity does not matter then it does simplify things quite a bit.  Grading companies either go away or turn into something analogous to a picture framing service.  Maybe all we really want is to own something that looks like the Michael Jordan PMG Green?  Perhaps then an abundance of cards that look just like the original Michael Jordan PMG Green are produced by some entity and then are made available to the market and it satisfies the desires of collectors. At a fundamental level cards are just wood pulp, some ink, and maybe some metal foil that captures a moment in time and the creative abilities of a few people.  Maybe digital versions are enough to make us happy.  Some research suggests photos of items create the same positive memories as actually having the physical item.  No more premiums for scarcity means that anyone can own whatever cards they want.  Is that a terrible thing?  If there were no social constructs around what is valuable or desirable would you like your collection more or would you like it less?  I suppose it matters if you derive the most enjoyment from the rush of profiting or for the journey and relationships associated with collecting.

After writing this blog post, I am of the opinion that grading can provide a false sense of security to most collectors.  Even if we can assume everything done by the grading companies is completely ethical, there is a chance that the condition of a card is overstated, alterations were not detected, or authenticity was improperly attributed.  Now that I’m considering the probability of unethical behaviors and the risk of having an inappropriate grade assigned, I do not plan on paying a premium for a graded card in the future.  This especially applies to grades that inflate the price of the card significantly.  A card being in a holder sealed by a third-party grader presents a barrier to self-determination of the condition of the card.  By no means does that mean grading as we know it will go away.  As I mentioned before, it’s all about what collectors cherish and value.  I think bringing awareness to the possibilities associated with trimming and grading aligns well with the purpose of this blog.  My goal all along has been to provide you with data, analysis, and insights that can help inform you so that you can build your collection with more information instead of less.  Of course, it’s up to you to decide where you land on this matter and all other matters that I research and write about.

One final note, which is more data based than exploratory.  I did take some dimensional measurements on 1986-87 Fleer basketball cards.  If you’re buying a card from this set, or you already own one, you can use this data set to help determine the probability that the card was created using the original factory cutting process.  Of course, you can also inspect the edges to see if you can detect deviations as well.  Regardless, I wanted to post this data since I have it.  Perhaps it’s useful to someone along the way.

I measured 63 cards from the 1986-87 Fleer set in order to build this data set.  With more measurements the mean and standard deviations would shift slightly in one direction or the other but the samples size is sufficiently large to give us a reasonably good approximation of the variation we can expect across the full run of 1986-87 Fleer basketball cards.  There can always be deviations in the production process, but probability says they occur at a much lower frequency than what we see around the mean values.

In scanning some information online, I have found a variety of size limit tolerances mentioned by a number of individuals.  Some people think the card graders look for the cards to be within 1/64” (0.016”) of the average dimension.  Others have stated the number is 1/32” (0.031”) or even 1/16” (0.063”).  When grading cards there is the potential to make two kinds of errors.  One would be a false positive, which in this case would be determining a card is too small to grade even though the card was produced using a factory process (i.e. not trimmed).  This is called a Type I error.  The other is a false negative, which would mean the card is trimmed but the grader did not consider the dimension to be too small.  This is called a Type II error.  If I’m a collector purchasing graded cards I would prefer for the grading company to choose to make Type I errors much more frequently than Type II errors.  If collectors were willing to accept that 5 out of 100 cards cannot be graded because of size variations (Type I error) then the size limit tolerance for 1986-87 Fleer basketball should be around +/- 0.009”.  For a service that is supposed to induce “faith in the process”, that seems like a more reasonable range.  There is no inalienable right that says all cards must qualify for grading.  In fact, if collectors are truly looking for the most perfect copy of a card, shouldn’t they want the one that is perfectly sized as well?

Till next time – Jeff