Highly Logical Outcome

As promised last time out, today we look at the press sheet array of the 1967 Topps Baseball high numbers.  This series set me on a path of unswerving nerdiness almost 40 years ago as I attempted to decipher a grainy photo of an uncut sheet and correlate it with the price guide short print designations of the day.  It would be many years before I realized Topps used a two slit (or sheet) press sheet encompassing 264 cards, with 132 cards per slit in a standard sized card array, 12 rows of 11 cards per slit.  132 card slits are called uncut sheets in the hobby, which is correct but doesn't account for all 264 cards, which I refer to as a press sheet. Within these parameters, cards sometimes became short or over printed as Topps changed the arrays from one slit to the other.

Why this happened is open to speculation but by the time 1967 rolled around I don't believe it was done to entice the kids to buy more cards looking for subjects that were suppressed in production. Rather, I think it had something to do with how they filled their various packaging configurations, at least in theory. Plus, I'd wager Topps really didn't want a lot of overstock or returns of the high numbers.

Many of the old price guide SP and DP notations were based upon the "tabulation" method where cards were observed as packs or vending boxes and cases were opened.  Depending upon the sample size, this was either accurate, or not.  I believe the idea of the 1967 Brooks Robinson high number card (#600) being super short printed - an idea which still somewhat persists to this day despite overwhelming evidence to the contrary - was due to the tabulation of a single 500 count vending box's contents.  

Well, there is still a wild card in the mix, which is a row or card's position in a sheet.  Corner cards,  those in bottom rows (but not necessarily tops) and the occasional random spot on the sheet do seem to have distribution issues sometimes and the 1967 high numbers are certainly affected by this. Topps must have had a way to segregate and pull certain cards, something that they had been able to do since at least the early 1950's, when various disputed player contracts with Bowman caused certain cards to be pulled due to court orders.

Putting that random production method issue aside for the moment, when this two slit brainstorm finally took hold, I searched for the second 1967 high number uncut sheet .  And I searched and I searched and I searched. Every example I found just showed the same sheet I had already deciphered as a young buck, like so:

It's an old scan but the basic row setup, with each letter representing a row and with the first, or head,  card in each identified, is:

A     Pinson

B    Ferrara

C    NL Rookies

D    Colavito

E    7th Series Checklist

A    Pinson

F    Red Sox Rookies

G    Orioles Rookies

B    Ferrara

C    NL Rookies

D    Colavito

E    7th Series Checklist

Then one day, over ten years ago, Friend o'the Archive Keith Olbermann sent along a partial section of a sheet with a different array-huzzah!  You can tell it's taken from the top left corner of an uncut slit:

That A row headed by Pinson is a match to the top of the other sheet but then the array changes.  So now we have a sheet that goes:

A    Pinson

F    Red Sox Rookies

A    Pinson

So a little odd but not 100% unexpected as the semi's seem to feature a 44x3 and 33x4 array and those Pinson rows get us to four. What to do now?

Well, I did an eBay count a couple of years ago and got this average count per row over 2,539 cards, with an overall average of 33 cards per subject:

A    77    Pinson

B    22    Ferrara

C    25    NL Rookies

D    22    Colavito

E    29    7th Series Checklist

F    34    Red Sox Rookies

G    26    Orioles Rookies

Row E has a slightly higher skew due having Brooks Robinson, a popular slabbing subject, in it and the checklist also being printed with the semi-highs (#531) but that Pinson row is such an outlier.  So where does this lead us? Well I thought here, using this pattern for the "other" sheet:

A    Pinson

F    Red Sox Rookies

A    Pinson

F    Red Sox Rookies

G    Orioles Rookies

B    Ferrara

C    NL Rookies

D    Colavito

E    7th Series Checklist

A    Pinson

F    Red Sox Rookies

Orioles Rookies

Or:

Five impressions:

A    Pinson

Four impressions:

F    Red Sox Rookies

Three impressions:

B    Ferrara

C    NL Rookies

D    Colavito

E    7th Series Checklist (contain Brooks Robinson)

G    Orioles Rookies  (contains Seaver Rookie)

But when you look at the PSA pops (from June 17, which total 42,165) it smooths out, just like with the semi's.  I'll save you all the math but when you factor out Hall-of-Famers, the Robinson and some other more widely collected cards, you get an average count per card of 405.  These are the per row averages using PSA's figures:

A    450

B    382

C    393

D    412

E    397

F    414

G    390

Ferrara is the lowest pop card at 293 and the White Sox Team is the highest at 530, factoring out all the "popular" cards but there is no discernable pattern, it's totally random.  The lowest pops are all over the place, as are the highest ones. The Seaver rookie leads the way among the glitterati, as you might expect, with 3,540 slaberoni's. Maybe the Ferrara was prone to damage or it's just not a card that's graded a lot, possibly due to centering issues. It occurs to me a production issue midway through printing could have occurred, requiring a quick fix of some rows on one of the slits, but good luck figuring that out if it even happened.

I once correlated the known SP and DP information as of December 2011 in a post and came up with 11 cards that didn't quite jibe among my source materials (i.e one source having an SP designation for cards from a specific row while another having the row as being full of DP's); all 11 cards that eneded up without overlapping SP/DP info were in the G Row. I still can't explain why the cards in this row caused set collectors the most reported difficulty (not counting the expensive Seaver card) other than confirmation bias playing a part, nor can I explain the relatively high A Row count, which is two standard deviations away from the mean where none of the other rows are more than one standard deviation away, although B is close on the short side and has the the lowest overall pop count average per PSA!

Summing up, the eBay dataset I used a decade ago was probably not robust enough. The PSA data suggest to me that rows A, D & F were printed 4x each and rows B,C, E & G were printed 3x each across the full 264 card press sheet, although A appears to be a bit of an outlier still (5% chance that it's random).  But it's not a guarantee and there could still be a 5x row, a 4x row and five 3x rows but there's too much noise I think to dissect this any further with the data at hand. I will say whatever old SP and DP data certain guides had in the past seem to be have been based upon incomplete information at the time. And it just feels like that Pinson "A" row is a fiver!