## Unveiling the real process used to choose the guilds in Return to Ravnica

This was in the back of my head for a month, but it surfaced just a night ago: How hard exactly had it been to find the correct split of guilds in the Return to Ravnica Block?

I am referring to this article from MaRo’s Making Magic column: *Return on Investment, Part 1*. Scroll to the second half, to the paragraph titled *The Guild (And the Other Nine Guilds As Well)*. It is about the selection process of the five guilds to appear in **Return to Ravnica**. Here, MaRo goes on rambling quite a while on the many aspects to be considered with this choice. I feel there is something very strange about his explanations, though…

Let’s take a look at the rules he is showing us:

**Rule #1) Each set had to have each color represented twice.**

(This is obviously the same as “…represented EXACTLY twice”, since each set uses five guilds.) This is necessary if you want to avoid gross imbalance between colors or guilds, which was understandably one of the most important goals.

**Rule #2) Each set had to have some ally and some enemy guild combinations.**

This was a “novelty” goal – MaRo explains that they did not want the very obvious split between a set with ally-colored guilds and one with enemy-colored guilds because they had already used that model in Invasion. I believe that restriction is unnecessary fanciful: Invasion block was well over a decade ago, and it actually consisted of a large set plus a small set featuring allied colors – and a smattering of shards -, and only a small set (remember that the difference between large and small sets was MUCH more pronounced then than it is today!) featuring enemy colors and a few wedge cards. It certainly never felt like a block consisting of an ally half and an enemy half, and it is also already distant memory, so I frankly don’t see the need to avoid this elegant and intuitive model for the Return to Ravnica block, but okay: Let us just accept this goal and move on.

**Rule #3) Any groups of four or three guilds that appeared together in Ravnica block could not all appear together again.**

You see, I slightly misremembered that rule when I pondered this problem, because I read it so that no three guilds which had already been featured in one set in the original block should be in the same set in the new block. As I will explain to you a little later, this would already have precisely defined the split of guilds – and in the exact way it was finally done, nonetheless! Honestly, I find it a little hard to believe that three sets from the original Ravnica appearing in one of the new expansions was ever seriously considered – wouldn’t this have been a much clearer rehash of the much more recent Ravnica: City of Guilds than the ally / enemy split? I cannot help but read this rule in another way than that it forces a 2/2 split on the four guilds from the original Ravnica expansion, and if so, there is no more room for additional design goals when deciding on the split. But MaRo goes on:

**Rule #4) All the fast guilds and all the slow guilds couldn’t be together.**

Well, yes, this reads like it requires really deep thinking! Aren’t you impressed with R&D at this point, juggling so many requirements at once? But we’ll see in a minute what this requirement actually means, even if we accept the possibility of 3/1 splits of the original Ravnica guilds…

**Rule #5) Start with two Dissension guilds.**

Now look – an even stricter requirement! But hold on – is there any problem fulfilling it? No, of course not – Rule #3 already determines that the Dissension guilds have to be split 2/1, so this is simply about the order of the two new sets, and since no other rule cares about that order, this one is a freebie… or isn’t it? Hold on again – do I read correctly that there were still two options of splitting in consideration, and that the final choice was based on WHICH GUILDS KEN NAGLE PREFERRED TO DESIGN (since he was selected to lead the first set’s design, but obviously not the second)? Now THERE is a professional decision-making process…

But now let us look at the options which actually existed. Keep rule #1 in mind – it is the deciding limiting factor. Then, let’s start with rule #2: There needed to be a mix of ally-colored guilds and enemy-colored guilds in both sets (well, obviously, once you have a mix in one set, you automatically also have a mix in the other). Let us not care about the order of those sets right now and thus define that the set containing the higher number of ally-colored guilds would be the first (we can switch the order later, but it is easier to talk about this way). Since there are 5 ally-colored guilds, this means that we will need to include exactly three of them in the first set: The 5/0 split is out by definition, and it is trivial that a 4-1 split isn’t possible if you need to follow rule #1, since four allied-colored guilds leave exactly two colors for exactly one guild here: The fifth allied-colored guild.

So, rule #1 and rule #2 together already define a 3/2 split of ally-colored guilds – I feel MaRo should have mentioned that, but I guess he left it out on purpose, since this trivial deduction takes away a little glory from the oh-so-complex decision-making process…

Let’s go on: There are two distinct models for a 3/2 split. Look at the color wheel: It shows a ring of friendly colors. Three color pairs consisting of friendly colors have either to be in straight succession (leaving out two adjacent friendly color pairs) – let’s call that model A – or they have to be divided into two adjacent pairs and one with no connection to them (leaving out two other disconnected pairs of friendly colors) – model B. Put another way: Model A has the friendly color pairs from both sets ordered 3-(2); model B has them ordered 2-(1)-1-(1). (The numbers in brackets show the pairs which are left for the second set.)

It doesn’t take long to realize that model A does not work with rule #1: There is a single color not included in any of the three featured ally-colored guilds (the one combining the two adjacent guilds from the other set, obviously), so that color still needs two appearances in the two enemy-colored pairs – but its enemy colors are, by necessity, the ones already used twice to link the three adjacent friendly pairs together! (For an example: If the three adjacent friendly pairs are GW, WU & UB, then Red needs two enemy-colored pairs to appear in to satisfy rule #1 – but its enemies, W and U, are each already featured twice.) To satisfy rule #1, Red would have to be paired with its allies, leading away from the 3/2 split and back to the all-allied / all-enemy split forbidden by rule #2.

So, our only hope left is model B: Two adjacent ally-colored guilds forming a shard (meaning they are linked by a central color), and a third ally-colored guild made up by the enemies of that central color. To fulfill the requirement of rule #1 is now possible in exactly one way: By using the two enemy-colored guilds which neither contain that central color (obviously, since it would appear in more than two guilds otherwise) nor both of its friends (because that would form a triangle of three guilds using three colors two times each, leaving the other two colors no partners to pair with aside from each other).

So, we have exactly one model actually fulfilling rules #1 and #2, and it is completely defined by its choice for a central color:

The two ally-colored guilds containing that color;

the ally-colored guild featuring the enemies of that color;

the two enemy-colored guilds each featuring exactly one of that color’s friends.

Since that is the only model satisfying the conditions of rules #1 & #2, we are already down to just five possibilities: One for each possible choice for the central color. Funny how MaRo didn’t mention that…

I will now list all possible splitting options according to rules #1 & #2 (the set containing more ally-colored pairs comes first here, but of course that order can be switched):

White as the central color – Selesnya, Azorius, Rakdos, Golgari, Izzet / Boros, Orzhov, Simic, Gruul, Dimir. (Yup, that is the split we ended up with!)

Black as the central color – Dimir, Rakdos, Selesnya, Simic, Boros / Golgari, Orzhov, Izzet, Azorius, Gruul (Out due to Rule #3 because all three Guildpact guilds are together in the second set.)

Green as the central color – Gruul, Selesnya, Dimir, Orzhov, Izzet / Golgari, Simic, Boros, Azorius, Rakdos (Out due to Rule #3 because all three Dissension guilds are together in the second set.)

Blue as the central color – Dimir, Azorius, Gruul, Boros, Golgari / Izzet, Simic, Orzhov, Rakdos, Selesnya (Contains three guilds from the old Ravnica in the first set, but this might be allowed.)

Red as the central color: Rakdos, Gruul, Azorius, Simic, Orzhov / Izzet, Boros, Golgari, Dimir, Selesnya (Out due to Rule #3 because all four guilds from the old Ravnica are in the second set AND all three Guildpact guilds are together in the first set!)

Now let that sink in for a minute… These were the five ONLY options to satisfy rules #1 & #2 – and THREE of them fail to satisfy rule #3! Actually, as I mentioned before, if you apply a stricter interpretation of rule #3 which makes more sense to me (no three guilds from the same old set should be together in the same new set), we would already have a clear winner! But wait… didn’t MaRo mention another rule (rule #4) AND say that the final choice came down to Ken’s personal preferences? Yes, he did, and that means he is not telling the truth: Either rule #4 had no bearing whatsoever on the decision-making process, or Ken didn’t have a choice.

But, hey, just look at the one possible guild split which didn’t make it: The two guilds listed as the fastest by MaRo are in separate sets; just like the two guilds he lists as the slowest! It gets even better: There is also a split between the 3rd and 4th fastest guilds, as well as between the 3rd and 4th slowest guilds! Interestingly, this is also the case with the split which R&D finally settled on: Both options obviously fulfilled the requirements of rule #4 perfectly!

Now MaRo writes that once “all these rules were thrown into a blender, there ended up being only two options.” He conveniently avoids the fact that you already ended up with these two options even without applying rule #4, meaning that narrowing down the field of options to that point needed all the design skills of an 8-year-old with a slightly over-average attention span… you don’t need to know ANYTHING about Magic design to come to this conclusion; you just need to understand the concept of the color wheel and access to a list with the original guild distribution. Yes, it all comes down to being able to count to five and crossing off options! How impressed are you now with the rigorous, professional design process leading here?

But, hey, there still IS a decision to make – there still IS room for R&D to show off their design skills while deciding the correct split, even though it’s just a choice between two possible options. So, the tie-breaker was…

… KEN NAGLE’S PERSONAL PREFERENCE OF IZZET AND GOLGARI?

Well, yes, since rule #5 states that there should be two guilds from Dissension in the first set, and since Ken Nagle obviously was only to lead the first set (otherwise this makes no sense at all, since he would have been in charge for all ten guilds anyway), this leads to White as a central color as the only option left, and that is it: The only choice left after applying most basic mathematics was made according to which guilds Ken Nagle likes…

Honestly – is that what you would have taken away from this part of MaRo’s article after reading it? Were you not rather deeply impressed by the complex puzzle R&D had to solve to create the perfect play environment for you? Well, yes – I’m afraid that was exactly the point of this article, as well as of many others MaRo has written over the last few years: If R&D makes decisions which seem bad to you, it is only because they face problems so complex and intricate that you should just stand in awe of them, not criticize them!

If you let convince yourself of this, take the process of splitting up the guilds in the new Ravnica block as a counter-example, since it shows so clearly that the means R&D used here were actually incredibly shallow and arbitrary. And you should be happy that Ken Nagle didn’t rather fancy Izzet and Orzhov instead of Izzet and Golgari, because in this case – at least according to MaRo – Guildcrash would have contained three of the four guilds which Ravnica: City of Guilds featured, and I am convinced this would have felt a lot more like something already done than that all-ally / all-enemy split.

*(Oh, while I’m here, I would like to remind my German readers that the third part of my limited preview for Return to Ravnica is now online on PlanetMTG!)*

**Explore posts in the same categories:**General

**Tags:** analysis, arbitrary, decision, Design, guilds, Magic, Mark Rosewater, professional, Return to Ravnica, shallow, split

October 2, 2012 at 11:26 pm

Ken Nagle is also the guy who wrote inflammatory posts on tournament players back in the old SCG days and designs cards with the simple purpose of filling a hole in his own Commander decks, so what can you expect from that guy. Oh, and regarding rule #2): They also used the “Invasion model” in Shadowmoor and Eventide. Granted, that was not a multicolor block, but still.

October 3, 2012 at 12:18 am

Well, I have yet to see a card design explicitly labelled as one by Nagle, which I don’t hate. I am still disgusted by the bluntness and idiocy of Blood Tyrant…

However, this one time I haven’t been ranting about Nagle, but about MaRo’s tendency to overhype the complexity of Magic design.

October 3, 2012 at 10:45 pm

I didn’t think it through completely but I already thought by myself while reading Maro’s article that it couldn’t be THAT complicated to choose the right combination. In my opinion the only necessary rules are #1 and #4 because they provide a good limited environment. With the other they could have easily been less restrictive without making anybody angry…