Half grade mathematics [Archive] - Net54baseball.com Forums

View Full Version : Half grade mathematics

Archive

03-01-2008, 12:29 AM

Posted By: Ken W.David, Agree with your first paragraph that the half-grade is just a premium bump (like A+), but not your second paragraph. If the precision of the grading measurement technique is only 0.5, then the overall grade can only be determined that precisely - to the nearest half of a whole number. This is the way it is done in analytical science. If I measure something between 7.5 and 8.0, but my precision is only 0.5, I have to make a judgement as to which number the result is closer to and go with it. Often, an uncertainty value would follow the result, but that doesn't really apply here.

Archive

03-01-2008, 10:38 AM

Posted By: Eric BrehmNo card is in 'perfect' condition, so perfection is just an abstract ideal that is used as a point of departure for assigning a grade. Defects which reduce the grade include those introduced at the time of manufacture, and those related to handling after a card enters circulation. Defects introduced at time of manufacture include uneven borders, out of focus image, and printing anomalies such as stray dots or blemishes, etc. Defects related to handling (or just age) include corner and edge wear, chipping, tobacco or wax residue, surface scuffing, color fading, discoloration of borders, scratches, creases, stains, ink and pencil marks, paper loss, tears and pinholes, and so on. The grading standards define how many defects a card can have to receive a '10', a '9', and so on. You start with perfection and then work down. The more defects, the lower the grade. An 8.5 would be a card that has too many defects to qualify for a '9', but not enough to receive a '7', so that puts it in the '8' range. But is has less defects than the average card in the '8' range, so it is recognized as one of the better 8's, and is thus assigned 8.5. How many cards will receive the new half grades, whether newly submitted or reviewed with an existing whole grade, remains to be seen. Because of the exponential drop-off in the probability of survival the higher on the scale you go, I would expect there will be fewer cards graded 7.5 than straight 7, and many fewer graded 8.5 than straight 8. In the lower grades, the distribution across whole and half grades may be more uniform.

Archive

03-01-2008, 11:16 AM

Posted By: davidcyclebackKen, I meant if PSA could know exactly where .5 was (no margin of error), rather than rounding to nearest available number. As we agree, PSA's .5 is really just a notation, and it can be left at that. Discussion of .4s and .6s is academic, as PSA can't measure that precisely. Though another academic question for thought is, is it legitimate for a PSA set registry set to have a average numeric grade of 5.55 or 7.23? What does such a number really represent? As one example, can an average grade be more precise than a single grade? Or, considering PSA's grading 'rounds' to the nearest number (6.0 to 6.5, 9.0 to 10-- there is no PSA grade of 6.2 or 8.7), is it legitimate to have a set graded at a grade number that doesn't exist and has not been similarly 'rounded'?

Archive

03-01-2008, 11:48 AM

Posted By: Eric BrehmOf course it is legitimate for a set to have a rating of 6.23, just like it is legitimate for the average height of a group of 20 men to be 5 feet 10.5 inches, even though no one guy has that actual height. PSA doesn't 'round' to the nearest number. They just use whole (and now half) numbers to designate a certain subjective level of evaluation of quality. There is really no mathematical meaning to it. A 10 is not really 10 times better than a 1, and is certainly not 10 times as valuable as a 1 in general. However, you do get 10 times as many points for a '10' in the registry as you do for a '1'. Your set ratings, however, also depend on the relative value/scarcity of the cards you have; cards which are more valuable (ostensibly, based on PSA 8 values) receive more weight in arriving at your Grade Point Average as well as your overall Set Rating. (The latter also takes your degree of set completion into account.) In cases where all cards in the set have the same weight, your GPA is simply the average numerical grade of the cards in the set, and your Set Rating is your GPA multiplied by your degree of completion.

Archive

03-01-2008, 12:03 PM

Posted By: davidcycleback"There is really no mathematical meaning to it. A 10 is not 10 times better than a 1" Then how can you perform meaningful mathematical calculations with the the PSA numbers? If a 10 is not 10 times better than a 1, then how can a set registry say "(10 + 1) / 2 = 5.5"? The accuracy of the answer in that equation is dependent on 10 being 10 times bigger (or better, in the case of grading) than 1. The crux is that PSA's numeric system and our normal system for doing mathematical equations are distinctly different. They can't be interchanged or mixed and matched-- as collectors erroneously try to do. As they share numeric symbols ('9,' '7'), they superficially resemble each other and lead people to erroneously believe you can mix and match the numbers in a calculator or chalkboard equation. As the above paragraph demonstrates, PSA's '10' doesn't represent the same value as the normal '10,' so you can't simply plug into a mathematical equation a PSA 10 as if it were a normal 10. If PSA used letters or other non-numeric symbols to just as accurately represent the same grades (A = best, etc), the hobby wouldn't be trying to mix and match the systems (After all, you can't type an 'A' into your calculator). It is because PSA adopted standard numeric symbols the people incorrectly jump to the conclusion that a PSA 10 is synonymous with a 10 on a calculator or in a Algegra 101 book. The two '10's or two '.5's may look interchangable, but they aren't, as they come from different numeration systems. PSA itself has said that the new .5 grade is not supposed to represent .5 in the normal mathematical sense.

Archive

03-01-2008, 12:20 PM

Posted By: Eric BrehmYou are correct, given that you have two cards in your set that have the same scarcity weight, you will get the same number of total points for a 10 and a 1, as you will for two 5.5's. But the fact is, in either case the average quality of your set is 5.5. Is having a 10 and a 1 better than having two 5.5's? Perhaps it is, certainly it more expensive, but your grade point average of 5.5 does give a good indication in either case of the average quality you've got. (Yes I know, you can drown in a lake of average depth 5 inches; that is the problem with averages.) On the other hand, if there were a set consisting of four equally weighted cards, and Guy A has one 8 and is missing the other three cards, and Guy B has completed the set with four 2's, then Guy A has a GPA of 8.0 and Guy B has a GPA of 2.0. They both have overall Set Ratings of 2.0 however -- Guy A has GPA of 8.0 * 25% complete = Set Rating 2.0 and Guy B has GPA of 2.0 * 100% complete = Set Rating 2.0 also. Which guy has the better set? Guy A has better quality on the average, but he just got started, and Guy B has lower quality but he has managed to complete the set. Seems to me the two guys have done about equal jobs in putting together a quality, complete set. Having said that, I agree that performing mathematical operations on the grades is probably a bit simplistic and perhaps misleading -- perhaps higher graded examples of the same card should get more weight in computing the GPA than lower graded examples because they are more valuable. (This is aside from the question of the card weights PSA applies to account for relative card value/scarcity between different cards, where a scarce card in 8 gets more weight than a common card in 8.) Applying another layer of weights based on the relative value/scarcity for each individual card in each grade would probably unnecessarily complicate the system. Say you were to get 5 times as much credit for a 9 as you get for an 8, because 9's are 5 times as expensive as 8's. And suppose you have two cards, one in 8 and in one 9, that each receive a weight of 5.0 for a 9 and a weight of 1.0 for an 8. Then you would get 45 points for the 9 (5 X 9) and 8 points for the 8 (1 X 8). You would have 53 total points, and your GPA would be 8.8 (53/6) rather than 8.5, if equal weights for the grades were used. I'm not sure such a scheme would lend any improvement to the current intuitive meaning of the GPA rating.

Archive

03-01-2008, 05:32 PM

Posted By: steveFor example, take an "old" PSA 7 graded card. In order for it to get the 7 it had to range between 6.5 and 7.5 in the graders opinion. Less than a 6.5 it would be in a 6 holder, greater than 7.5 it would be in an 8 holder. Now with the half-grades, a 7.6 thru 7.9 will be in a 7.5 holder. An 8 thru 8.5 will be in an 8 holder. And an 8.6+ will be in an 8.5 holder. In summary, an "old" 7.5 through 7.9 is sitting in an 8 holder. Now, with the new system, it would not be in the 8 but rather a 7.5 holder. In conclusion - high end for the grade cards will now only be a 7.5 instead of an 8. High end for the grade cards already got their bump with the old single digit system. steve

Archive

03-01-2008, 05:39 PM

Posted By: BcDThe parent department of PSA,PCGS (coins) already breaks grades down to the .001's! sure cards could follow that if the grading keeps fine tuning! BcD <img src="/images/happy.gif" height=14 width=14>

Archive

03-01-2008, 07:07 PM

Posted By: AnonymousIncorrect. According to all statements from PSA, a 6.99 (if they could grade so precisely) would be called a 6 because it did not meet the minimum requirements for a 7.

Archive

03-01-2008, 08:05 PM

Posted By: davidcyclebackPSA said a 6.5 means the card is an extra good example of a six. A .5 essentially means a '+.' They could have just as easily used a .4, a .3 or a + to mean the same thing. So the .5 isn't intended as an exact mathematical measurement, but a notation that card is a premium example for that grade. PSA's numeration system isn't something you'd want to use to perform actual mathematical calculation with. If PSA could literally determine where .5 was, that would mean they would know where .4 and .6 are. You can’t count to five without counting to four first (and you can't count to four without counting one, two and three). You can't know where is exact center without knowing exactly what is on each side. If PSA could really determine where is dead center, that would mean they could provide 90+ different grades.

Archive

03-02-2008, 05:14 PM

Posted By: Scot ReaderPSA is only bumping about 3-5% of cards submitted under the half-grade service. Moreover, most collectors are only re-submitting cards they think are particularly strong for the grade. Given the rigor of the half-grade standard, cards receiving a "point-five" grade cannot be thought of as simply at the midpoint between two integral grades. The PSA half-grade is better thought of as a qualitative, aesthetic metric without an exacting quantitative basis.

Archive

03-03-2008, 07:16 AM

Posted By: Eric BrehmI agree with Scot; as I said before, card grade numbers should be viewed only as labels for subjective condition qualities, not as literal numbers to be manipulated in a traditional mathematical sense. But in terms of Set Registry arithmetic, and the effect of the new grading system on the set rankings (for those who care about such things), it does make sense for less than 50% of cards that originally received whole grades under the old system to get a half grade bump. Example: suppose for the sake of argument there is a set where all cards in the set have equal weight (i.e. they are equally scarce/valuable). Guy A has 100% PSA 8's with a set rating of 8.00 Guy B has 67% PSA 8's and 33% PSA 9's with a set rating of 8.33 Suppose Guy A re-submits all of his 8's for review. What is the effect on his set ranking as a function of the percentage of his cards that receive the bump? If Guy A received the bump from 8 to 8.5 on half of his cards, his set rating would jump from 8.0 to 8.5, and he would pass Guy B's 8.33 score in the ranking. But if Guy A only received the half grade bump on 10% of his cards, his set rating would only increase from 8.00 to 8.05 (.90 X 8.0 + .10 X 8.5 = 8.05). He would still trail Guy B in the rankings. If he wants to catch Guy B, he is going to have to buy some PSA 9's. That is probably the way it is going to play out, so those who fear that they will be compelled to resubmit their cards in order to retain their current place in the Set Registry rankings should probably not worry about it. And similarly those who think they might have an opportunity to move up significantly in the standings will probably be wasting time and money if they try to accomplish that by re-submitting their cards for review.

Archive

03-03-2008, 08:30 AM

Posted By: JimBI have always assumed since Day 1 with PSA that a PSA 6 was everything in the 6.0-6.99 range. A PSA 7 was everything in the 7.0-7.99 range, etc. That is the impression I got from reading their grading criteria. I never presumed that a "6.6" would get a 7. JimB

Archive

03-03-2008, 08:34 AM

Posted By: Eric BrehmJimB -- correct, that is why I explained in my earlier post that you start with the abstract ideal of condition 'perfection' and then work downward through the scale, dropping one whole grade (and now half grade) with the number of defects of various types you discover on close inspection. As soon as you discover the slightest thing that gets you out of the 7 range, even if in theory you have '6.99 quality', you drop to the 6 range. But maybe now you'll get a 6.5 instead of a 6.

Archive

03-03-2008, 11:13 AM

Posted By: davidcyclebackI think the creation and application of a professional grading number scale for trading cards is one part applied mathematics and one part voodoo. And the third part is marketing.

Archive

03-03-2008, 12:08 PM

Posted By: Eric BrehmIt is marketing driven by (or obfuscated by) voodoo mathematics. ... at least as far the Registry competition aspect of it goes. And of course many, like me, do participate in the Registry but don't particularly care where they are in the rankings (provided my set shows up in the Top Twenty on the first page). <img src="/images/happy.gif" height=14 width=14>