Deconstructing print runs

"Numbers for Revised, Fourth, Ice Age, Chronicles, Homelands and promo cards 
are guesses based on almost no information.  I think I'm in the ballpark, 
but don't bet money on it.  Distribution between starters and boosters for 
The Gathering is almost completely guesswork.  Other numbers are from net 
postings and such. No inside information was used to build these numbers."

-Stephen D'Angelo, Rarity Information at Crystal Keep, circa January 1996

It is hard to overstate the impact of Stephen D'Angelo and the Crystal Keep webpage. The D'Angelo Files and Crystal Keep were the go-to sources of Magic information in the early days of the web. In an era before wizards.com or The Dojo, there was D'Angelo and his collection of wisdom from the internet and beyond. And wisdom he had. 

Apart from his updates on card lists, rules and errata, he took a stab at estimating the print runs for Magic sets. Some of these had questionable official numbers, and some had nothing but hints or nods. But he gathered the information he could find, bundled it together on a webpage, and added a disclaimer at the top noting that he had no inside info and that a lot of it was "almost completely guesswork". He had some very solid estimates, but as the Internet grew D'Angelos numbers somehow got repeated as the singular truth over the web and beyond. Wonder where the 1,100 estimate of each Alpha rare came from? That gained traction with D'Angelo. The only printed reference to Alpha rares in a Wizard's publication was in The Duelist #3, where they noted 1,400 cards.

A later (1995) supplement to The Duelist called The Complete Card list would however imply that the number was closer to 1,070 rares. I think that is slightly off as well.

Writing this post was triggered by a comment from Scott Latham on my previous post on Unlimited. After re-reading Scott's own (excellent) article on the set, it started to dawn on me how most every resource online that talks about print runs comes back to D'Angelo's estimates, and that so much random stuff we've found over the years cast new light and perhaps better accuracy on some of the "accepted truths". And we can't have people being wrong on the Internet. So this post is my truth on 93/94 era print runs. I believe I have access to pretty much all the written sources D'Angelo had at his time, along with perhaps a handful more thanks to excellent books and some better information on the print sheets and printing process available. This will be deep nerd and very number crunchy. You have been warned. Let's go.

Update March 2021: Every now and then new information on these things get unearthed, and one spectacular moment of new insights occurred just months after I'd finished writing this post. Peter Adkison (the founder of WotC) personally reached out to Luc Mertens (the sales executive at Carta Mundi throughout the 90s) to clear up what they had actually received from them as the first Magic shipments were concerned. Luc compared notes with one Peter de Weerdt, who had taken over Luc's responsibilities after his retirement, and after digging through the archives they found the actual, exact numbers for Alpha and Beta. So while our guesses for those sets turned out to be very close, they were slightly wrong. In particular our presumption that there was a 50/50 split of cards between boosters and starters was our undoing. So screw you, mid 90s WotC Usenet NetReps,  for giving us bad intel on that a quarter century ago I guess ;) All in all, these are the numbers of each rare, uncommon, and common shipped from Carta Mundi for Alpha and Beta according to their internal documents:

Number of each Alpha rare: 1,008

Number of each Alpha uncommon: 4,530

Number of each Alpha common: 16,030

Number of each Beta rare: 3,025

Number of each Beta uncommon: 13,590

Number of each Beta common: 48,100

I've opted not to update or edit this post apart from this disclaimer. I think it can be interesting from an historical context to see our old sources and assumptions, how we did the calculations, and where we went wrong. Our assumption of a 50/50 split of cards between different packages was the core source of mistakes (a 60/40 split with 2.5 million cards would lead to a bit under 1,000 copies of each Alpha rare (rather than the frequently cited "just over 1,000"), which would have us reject the 2.5 million figure for the then far more likely 2.6). OK, back to the original post!

Alpha

Alpha most likely had 2.5 million cards printed. I base this on information from production manager Dave "Snarke" Howell in 1993, which was noted in e.g. official.comments.and.PR.1993 at the Magic ftp (marvin.macc.wisc.edu). This number was repeated by e.g. Tom Wylie in 1994 (rec.games.trading-cards.magic.misc; huzzah for BBS archives and old ftp servers), and in general seems to be the number in early official documentation by people who worked with the release. A couple of other suggestions are of note however. One is 3.4 million cards, which was implied in The Duelist #3 via their "1,400 of each rare" suggestion. That particular article was written by J.M. White, the then-editor of Scrye Magazine. Impressive role, but probably not one deeply involved in print runs, and I believe she was slightly off.

Another suggestion from The Duelist (in The Complete Card List from 1995) is 2.6 million cards, which was later repeated in the first Official Encyclopedia. This 2.6 million estimate has since become the most common go-to number, but looking at information from the actual production manager and other sources closer to the printing, 2.5 million would be the best bet. It is reasonable to assume that these cards were split 50/50 in boosters and starters, as it has been documented by NetReps that Beta was split in that way and Alpha is basically the first shipment of Beta. The rarity distribution in a booster is 1 rare, 3 uncommons and 11 commons; in a starter it's 2 rares, 13 uncommons, 45 commons.

Now we reach a point which I think might be a source of some small estimation confusion. While there are 116 rares, 95 uncommons and 74 commons in Alpha, you can't just divide the number of rare slots in packs with the numbers of rares in the set. The rare slots are occupied by cards from the rare sheet, and the sheet had 121 cards, not 116. Every sheet has 121 cards, and Alpha didn't have a separate land sheet, instead basics were spread out across the other sheets. So, given 2.5 million cards, 50/50 split between Starters and Boosters we'd reach these numbers (rounded to the closest 10):

Number of each Alpha rare: 1,030

Number of each Alpha uncommon: 4,300

Number of each Alpha common: 15,320

Total number Alpha basics: 837,300

Dave Howell also noted in August 1994 that "just over 1,000 rares" had been printed for Alpha, which aligns well with our numbers here. And I trust Dave Howell over J.M. White on these kinda things.

However, there's a small however. A number of Alpha rares were to be found in Beta starters. This would mean that they printed more Alpha rare sheets than they needed in the original 2.5 million batch. There's also rumors of Alpha "God packs", where a large number of the cards in the pack are rares (God packs surely existed in Beta; here's a video). While Alpha God packs might be nothing more than a rumor, there definitly were a handful Beta starters with Alpha cards. It was a rare thing though. If the odds were as high as one in a hundred, that would still only add an additional 10 or so copies of each rare (and I suspect the odds were far lower than that). With rumored God packs, Alpha rares in Beta starters, and any other possible collation problems, I still wouldn't count on the numbers to change with more than 10 copies for each card, and we're most likely talking single digits.

I guess we should also note that Alpha rares are very rare. It is likely the set that has suffered the most impactful collateral damage over the years. A decent percentage of the cards were likely thrown away after being used as product samples, discarded as marked cards when combined with newer printings, or simply trashed as a part of some "satanic panic" at the time. And in the weeks Alpha was for sale, it is not like there existed price lists or a proper secondary market for Magic; cards didn't come with an inherent price tag to make you wary of throwing them in the trash after failing to grasp the rules. But estimating how many of the Alpha cards that survived past 93/94 would of course be nothing more than guesswork. 

Beta

Limited Edition - that is Alpha and Beta combined - was supposed to be 10 million cards total. This was the expected print run, and this limitation was stated by Peter Adkinson early on. In early documents, we hence often find 7.5 million as the print run for Beta. In the Pocket Players Guide (1994) and The Duelist (1995) we also find the number 7.3 million, which was probably based on the presumption that Alpha was 2.6-2.7 million and the total number should add up to 10.

Shortly after Beta had been finished, Richard Garfield noted in an interview with the magazine Cryptych that there had been 61,200 copies of each common card printed between Alpha and Beta combined. This is a very specific number for these kind of things, and works out almost perfectly if we have an Alpha run of 2.5 million and a Beta run of 7.5 (the number adds up to 61,295 to be exact). This again points to a total run of 10 million cards - the number that Adkinson had promised early on.

It was however noted by production manager Dave Howell (in "Bog Wraith and Boogeyman Halloween Edition FAQ" from October 1993) that they overshot the production a little, and a total number of 10.4 million cards were in fact delivered. This would put Beta on 7.9 million cards. As production manager, Dave Howell would likely have the most accurate info on these things, and there's really no reason for him to make that number up, as it makes the "collectible limited edition" less rare and exclusive, and it's a confession that they messed up their the plan to print 10 million cards. So I strongly suspect 7.9 million to be the right number. This puts the print run of individual cards like this:

Number of each Beta rare: 3,260

Number of each Beta uncommon: 13,600

Number of each Beta common: 48,420

Total number Beta basics: 2,594,160 

 

Unlimited

There are two places where I'd like to challenge common conceptions on Unlimited. The first one is something pretty much everyone admits to being guesswork; the distribution of cards between starters and boosters. Everyone tend to agree that Wizard's had moved from away from thinking the average player would buy something like one starter deck and four boosters as their Magic collection when Unlimited was made. It was clear that players were buying a lot more boosters than that, and as such Wizards knew they should shift more cards to boosters. And here is where someone (outside Wizards) guessed that the distribution would be 2/3 of cards into boosters, and 1/3 into starters, which was then repeated without much fact checking 'til the point it became "common knowledge".

And that's where we, the full nerds, take a deep dive into UseNet archives from March 1994, to find a topic regarding expected availability of Antiquities, where the discussion moves into people complaining about the current (non-)availability of Unlimited. And who would chime in if not the production manager himself, Dave "Snarke" Howell:

Ah, 27 year old UseNet threads.

"100,000 decks". That would mean around 6 million cards in starter decks. Maybe slightly less, as 100,000 seems like a rounded number and if it were more he probably would have stated "more than 100,000 decks" to make his point harder.

The second thing I'd like to challenge is the full print run. I've read 40 million cards on a lot of random sites, but I have no idea where it came from. The earliest estimates by D'Angelo is 30-35 million (March 1994). Sometime later, he changes his guess to 40 million. Could it simply be based on this random UseNet comment from June 1994?

I don't know who this guy is, but he doesn't strike me as one particularly deep into first hand info. So in this case I'd actually like to turn to The Duelist Complete Magic Card List (1995) and The Official Encyclopedia (1996) as references. And they both tell me 35 million.

You can read this online at archive.org. Internet is pretty sweet.

I know I've criticized numbers from The Complete Magic Card List and Official Encyclopedia earlier in this post, and if someone can give me any good reason that the number is in fact 40 million I'd be happy to change my mind. But as it stands, we have a production manager saying that the printing got truncated to make room for other products, and two official sources (along with early estimates from D'Angelo) putting the completed run at around 35 million. Then we have one loose comment and a bunch of circle-references on non-official websites stating that it was 40 million without apparent basis. 35 million seems far more likely.

And then there's math. Popular opinion among Carta Mundi nerds held that the ration of printing rare:uncommon:common sheets for Unlimited was approximately 1:4:15. If we print one starter (2:13:45) and five boosters (1:3:11), the ratio averages out to 1:4:14.2; close enough to call home about. And given that we know we should have around 6 million cards (maybe slightly less) in starters from a statement from Dave Howell, that would put the total run at some point just below 36 million cards, with 1/6 of the cards in starters. All in all, things work out if we have a print run of 35 million cards with 1/6 of the run in starters and 5/6 in boosters, and get kinda weird otherwise.

With all those word and numbers said, I believe these are the print numbers for Unlimited (rounded to nearest 100):

Number of each Unlimited rare: 17,700

Number of each Unlimited uncommon: 58,700

Number of each Unlimited common: 212,900

Total number Unlimited basics: 11,390,300 

 

Arabian Nights

Ah, finally an easy one. Every early (and late) source points to Arabian Nights having a print run of five million cards. It only came in booster packs, so no fuss there either. And it only had two print sheets. Each booster contained eight cards; two from the uncommon sheet and six from the common sheet. So all we have to do now is count the cards on the sheets and apply some basic math. (As 5,000,000 isn't divisible by the print sheets, it is likely that the actual number was something like 4,999,720 or 5,000,204). In the end, we get this (rounding to nearest 100) :

Arabian Nights U2: 20,700

Arabian Nights U3: 31,000

Arabian Nights U4: 41,300

Arabian Nights C1: 31,000

Arabian Nights C4: 124,000

Arabian Nights C5: 155,000

Arabian Nights C11: 340,900

Antiquities

As with Arabian Nights, the Antiquities print run is not really a contested number. All sources point to 15 million cards, or maybe slightly over. Packs have eight cards, where two are from the uncommon sheet. I didn't count the different pictures on the Urza Tron lands as different cards here btw, if you want to do that you could divide the C5 into three C1 and one C2, and the C6 into two C1 and two C2. Anyway:

Antiquties U1: 31,000

Antiquties U2: 62,000

Antiquties U3: 93,000

Antiquties C1: 93,000

Antiquties C4: 372,000

Antiquties C5: 465,000

Antiquties C6: 558,000

Legends

Not much to say here either. Barely know why I'm even spending time on these expansions, as the opinion of the internet hive mind appear to be pretty much identical to my own calculations. It is notable that Legends is a ridiculously large expansion though. It has more different cards than Beta. It also has more cards than Arabian Nights, Antiquities and The Dark combined. The vast number of different cards on the highest rarity (121 different rares) give individual rares in Legends very small print numbers.

Anyway, 35 million Legends cards give us something like this (rounding to closest 100 as usual):

Legends Rare: 19,300

Legends U1: 57,900

Legends U2: 115,700

Legends C1: 212,300

Legends C2: 424,600

The distribution in Legends boosters would imply there exists three times as many uncommons and eleven times as many commons as rares. Math checks out.

The Dark

Ok, so this one gets a bit more tricky. Two separate releases from The Duelist in 1995, along with a November 1994 press release (by Carrie Thearle) placed the print run at 62 million. Scrye noted 63 million. The Duelist #3 (1994, in the article that quoted Alpha almost 40% off no less) placed the run at 65 million, as did some early discussions on rec.games.deckmaster before the official release. 

On March 31 1994 the following policy release went out:

"Wizards of the Coast would like to offer an explanation concerning the
placement of orders for The Dark. As you are probably aware, WotC had
originally told all distributors that if orders were placed by a certain
time, WotC would guarantee that those orders would be completely filled.

However, after all those orders came in, WotC became very concerned. It
became apparent that if such a huge amount of product were released it
would ruin the collectability of The Dark. Magic: The Gathering
contained approximately 300 different individual cards. The Dark
contains approximately 100 cards. Yet orders of The Dark outstripped
orders for Magic: The Gathering, Limited Edition by more than 7 to 1.
This meant there would have been 21 times as many cards in The Dark as
there were in Magic: The Gathering, Limited Edition. Thus, an uncommon
card printed in The Dark would be 26 times more common than an uncommon
card from Magic: The Gathering, Limited Edition. In reality, no cards
would have been uncommon at all.

WotC has always felt that our Deckmaster line was more than just a
game. The opportunities for collectors are obvious, and we felt that it
would be a disservice to those collectors for us to blur the lines
between common and uncommon cards to such an extent. For that reason
WotC reduced distributors' orders of The Dark."

21 times as many cards as there were in Limited Edition (given an Alpha+Beta run of 10.4 million cards) works out to 218.4 million. The number of U1s in The Dark would then be 25.2 times as many as the U1s in Limited Edition. Math checks out. So it appears they had orders for 220 million cards, give or take, and decided against printing that many. 

In June of 1994, Tom Wylie noted "First, I know that the print run for The Dark is more than four times the print run for Antiquities, which was a comparably-sized expansion. Also, the cut [in store pre-order allocations] was made way back in late March or early April--we are not making any additional cuts." This implies that The Dark was planned to be more than 60 million in the early summer of 1994. Another NetRep around that time is noted to have quoted the set to about twice the size of Legends (i.e. 70 million cards).

Then, in late August 1994 (later the same month as the set was released) we find this clarification post from Tom Wylie:

I believe this is the actual number. It aligns well with plans to not print to demand (interpreted around 220 million), but having some wiggle room to adjust during the summer as new stores pop up. It smoothly evolves from "more than four times the size of Antiquities" and "around twice the size of Legends" earlier in the summer, and this number comes directly from Wizards closely after release, not as an estimate before the presses had cooled off. With that into account, I believe the November press release number was slightly wrong (it might well have quoted an earlier number on the Magic ftp), and when that number was repeated in the 1995 Duelist it snowballed into general acceptance.

All this taken into account, the most likely print numbers for The Dark are:

The Dark U1: 155,000

The Dark U2: 310,000

The Dark C1: 403,000

The Dark C3: 1,209,000

...

Jesus, this took a lot of time to dig through, cross-check and calculate. Got proper lost in UseNet groups for a while there, and my The Duelists don't have ctrl+f. This somehow felt like the fourth or fifth most time-consuming article I've ever written. I again blame Scott and the Brothers of Fire for pushing me down the rabbit hole to actually try and fact check these old numbers.

I would love to end here, but feel I should add a couple of notes on Revised, Fallen Empires and 4th Edition.

 

Revised

Revised is very hard to get a good grip on, but as it is part of the card pool for a majority of the world's OS groups and I believe the current popular estimates (500-600 million) are off by a factor two or three, some notes are warranted. In April 1994 (when Revised was released), Dave Howell mentioned that 50 million cards had been printed for the set at the time. As Carta Mundi had a capacity of around 15 million cards per month back then, and Revised started its print run in February, this makes a lot of sense. Carta Mundi increased their capacity quite a bit over the next year, and by mid 95 they could produce 90 million cards each month. But in the time span between summer of '94 to summer of '95 a majority of card production had shifted to Jyhad and Fallen Empires, and after that 4th Edition. We also know that 100 million Revised cards had been printed by August 1994, but how many came after that?

Well, in WotC News in The Duelist #3 (fall of 1994), Wizards noted that 300 million cards total had been sold at that point. That number would include most of the 175 million cards from the early expansions and core sets, along with the first drops of Fallen Empires and the first six months or so of Revised. In the summer of 1995 we find this post on UseNet from Marc "Sparky!!" Schmalz:

This 500 million number would include all of the expansions up to Fallen Empires (though FE had a good number of unsold cards in stock still at this point) and every core set up to and including Revised. FE is also very hard to put a more exact number on, but official statements include 321 million, 340 million, "about four times The Dark" (i.e. about 300 million) and "approximately three times the size of all previous Magic expansions combined" (i.e. approximately 390 million). How much of these card that were sold before they ordered 4th Edition is anyone's guess, but judging by the speed with which the game grew back then and the sales numbers for the expansions before it, I'd be kinda surprised if it were less than 150 million. There were a lot of FE cards everywhere in early 1995. 

All that taken into account, I would put the print run of Revised at around 200 million cards, give or take some tens of millions. The "around 500 million" estimate we find at e.g. Mtg Wiki is not really possible; it defies logic both in regards to what little data we have from WotC on the subject, and the capability of the printers at the time.

Fallen Empires

Hah, already covered the basics here while discussing Revised. Fallen Empires was somewhere between 300-390 million cards, with the two more exact official statements being 321 million and 340 million (both of them from 1995 in different publications of The Duelist). 340 was the latter number of the two, so maybe that one is the most correct? I am not digging any deeper on this one, and regardless where you decide to place your trust here, just know that a lot of Fallen Empires was printed.

4th Edition

Ok, last one. I guess the most common estimates for 4th are somewhere between 500 million to 700 million. I'm not sure where those numbers came from. We do know that pre-orders were to the tune of 900 million cards, though it is of course possible that WotC didn't want to fill the pre-orders (like they decided against with The Dark, and in a sense Fallen Empires when they stopped those presses in January 1995, though it is notable that the white bordered 4th wasn't intended to be "collectable" and limited the same way as black bordered sets were). The closest thing I can find to an official estimate is this post from the summer of 1995:

I don't know if this is hyperbole or something, but I read this as "we are printing over one billion 4th Edition cards". And I mean, considering the small number of expansions and long period of time between 4th and 5th, the increased capacity of the printers, and the surging interest in the game, I'd consider a billion (or more) 4th Edition cards to be a very reasonable estimate. So I believe the Magic Wiki is currently off by a factor two or so here as well, just in the opposite direction compared to Revised.

...

Now I'm actually done. Jesus, this was a bit nerdy even for me, and I once wrote a post about Richard Garfield's PhD thesis. But people were wrong on the web, and I had to correct them using printed media, 90's UseNet posts and math. Also the web I guess.

'Til next time, keep safe and enjoy the season responsibly.