Welcome to Praetor Week. This week will be about discussing the mythic rare cycle of legendary creatures that are running New Phyrexia. Designwise there isn't all that much to say about them beyond what I said last week, so I'm going to use today's column to do something I don't do often in this column. I'm going to explain a design problem to all of you that we haven't solved yet. What does this have to do with Praetor Week? Quite a bit, because the topic I want to discuss is something all the Praetors share in common: the supertype legendary.

First quick aside of the column. (Oh yes, there are more to come.) The problem I'm going to talk about today is what I refer to as a "problem of luxury." The game works just fine as-is and can continue to work with the issue I'm going to talk about. One of the things R&D tries to do, though, is to constantly find ways to incrementally improve the game. Today's issue falls in this camp. In short, R&D feels what we have is good, but it could be great. Today's column is about recognizing an issue that R&D thinks we could do better.

"True Story."

Before I discuss the issue with the legendary supertype, let's start by having a little history lesson. Okay, faithful readers, let's hop into the WABAC Machine and set the dial for the summer of 1994. Magic wasn't even a year old yet when its third expansion, Legends, was released. Designed by Steve Conrad and Robin Herbert, Legends took the game in a very different direction. Steve and Robin based their set off of their Dungeons & Dragons roleplaying sessions. Many of the cards referenced specific characters that they or friends had played. To capture this sense of a specific person, Steve and Robin came up with a new creature subtype: Legend.

In addition to try and capture the complex flavor of these legends, Steve and Robin introduced another revolutionary idea to the game: multicolored cards. These characters were so special that they were the only ones to use both the new subtype and the multicolored mana costs. Obviously the focus on them can also be seen in the name of the set.

A year later with the release of the next large set, Ice Age, the Legend subtype returned. It was only on four cards—two of which were multicolored—but it was clearly the first sign that the subtype had some potential. The next set after Ice Age was Homelands, which had fourteen legends in it.

Remember that Legend at the time was a subtype with mechanical baggage, not an actual mechanic. The importance to this difference was that there was no restrictions on what creature subtypes could be used. If a designer wanted a subtype, he or she just used it. This meant that Legend was free to be used whenever.

I feel the importance of Homelands to Legends was a push on the idea that having characters was a valuable tool, one that more sets should use. Homelands might not have been a great success mechanically, but it did demonstrate that flavor could play a larger role than it had in the past. From Ice Age forward, the Legend subtype became a staple that showed up in every set, although more in some sets than others.

"Haaaave you met Ted?"

The next big change in the Legend subtype came with the expansion Champions of Kamigawa. R&D had long disliked that there existed two subtypes that carried rules baggage, Wall and Legend. To solve the Wall problem, R&D created the defender keyword and retroactively gave every creature with the Wall subtype defender. (Even today, it's considered a given that a creature with the Wall subtype has defender.) They then stripped Wall of its rules baggage.

Wall had an easy fix, but what were we going to do about Legend? We toyed with the idea of doing essentially the same thing with Legend as we did with Wall. Imagine if every Legend had the keyword "legendary."

Besides the connection to the subtype, Legend had another problem. When Legend was introduced, here is how it worked mechanically. Any person could play a Legend provided that that Legend wasn't already on the battlefield. If it was, that card was stuck in its owner's hand. (They could cast it if they wanted to, but the newest one would immediately be put into the graveyard, so there was no incentive to do so.)

This issue came to great prominence during the Mercadian Masques block because the dominant deck was a deck known as the Rebel deck (making use of the Rebel mechanic—the mechanic itself wasn't named, but all the cards with the ability worked the same). The key card of the Rebel deck was a Legend named Lin Sivvi, Defiant Hero.

Lin Sivvi, Defiant Hero

The card was key to the deck so when two rebel decks played one another, the first person to get Lin sivvi out had a huge advantage. R&D felt that this gameplay was particularly bad and were eager to find a way to improve it. One of the suggestions that came up was to have the second copy destroy all copies in play. This way if your opponent gets the card onto the battlefield and you draw yours, at least you could use your copy to neutralize your opponent's.

R&D had talked about making the change but didn't know how or when to do it. Then along came Champions of Kamigawa. The set had a gimmick where all of the rare creatures were Legends.

Quick aside. I often talk about all the wonderful ideas I have that I had to fight through the onslaught of naysayers to get printed. This is an example of one of my ideas that I wish someone would have stopped. You see, while I wasn't on the design team for Champions of Kamiagwa (or any of the sets in the block), I was on the Champions of Kamigawa development team. One of my recurring themes in development was that I didn't understand what the central theme of the set was.

Eventually I was told it was about Legends, to which I replied that it is hard to have a theme that players can't see. The only way for players to even have a chance, I said, was to make all of the rare creatures and some of the uncommon ones Legends. It turns out that, one, that wasn't enough (our understanding of the importance of "asfan," the actual frequency of cards in booster packs, has grown) and, two, taking something special and making so many of them that a lot of them can't be special is a bad, bad plan.

Anyway, Champions had decided to throw its proverbial hat into the Legends ring so we all felt like if we were going to make a change to Legends, this was the time. We opted against the "Wall solution" as we knew that Legends had a tendency to get wordy so finding an answer that didn't chew up text box space seemed important. That is how we got to the idea of a supertype. The basic supertype had been introduced the year before so we knew that the technology existed.

"Legend" became "legendary," as we like supertypes to be adjectives as it sounds more natural stuck in front of card types, most of which are nouns. (You use "instant" in a Magic context enough and it starts to sound like a noun.) We made the change in Champions of Kamigawa and ever since that is how legendary has worked.

"Challenge Accepted."

Now that I've talked about how we got here, it's time to talk about the problem at hand. It short, R&D is not fond of "legendary." To be clear, we love having cards that represent specific characters. We understand the desire for the game to have notable personalities and we wholeheartedly support it. Obviously, legendary creatures are also a huge part of the Commander format and, as this summer's Magic: The Gathering Commander shows, we are fully behind that format as well. Our issue is specifically with how the legendary supertype works.

What don't we like? Let me explain:

Problem #1 – It's an All-Downside Mechanic

Each year we do market research to see what players thought of the previous year's cards. As part of this study we look at keyword mechanics to get a sense of what was popular and what was unpopular. This data helps us in two ways. First, it specifically lets us know what players think about particular mechanics. This allows us to use data to help decide whether or not we might want to bring that mechanic back. Second, it allows us to see trends.

The following five mechanics did not fare all that well in surveys: Echo, fading, phasing, suspend, and vanishing.


What do these mechanics all have in common? They are what we refer to in R&D as "downside mechanics." That is, they allow a player to get a more powerful spell at a cheaper cost but at the price of something else. Maybe the creature disappears every other turn or just goes away altogether after a few turns or possibly takes multiple turns to even get to the battlefield.

Tolarian Drake
Shade of Trokair

Time has shown that downside mechanics poll poorly. But it doesn't stop there. Not only do downside mechanics fare poorly, so too do most downside cards. The biggest reason upkeep costs, as an example, have dropped off significantly is that players don't like them. There are exceptions, of course—players will accept almost anything if the card is powerful enough—but as a trend, it's clear: most players do not like downside.

Waning Wurm

That's the first problem with legendary. It's all downside. Now, I'm talking mechanically here. There is plenty of flavor upside. In fact, there is so much flavor upside that we've just sucked up the mechanical downside for years. Nonetheless, legendary is doing something that we've been trying to do less of in recent years.

Another quick aside. Just because I say R&D is down on downsides doesn't mean that they're going to disappear completely. There is a time and place for downside cards (and, to a lesser extent, downside mechanics), but we choose to use them very carefully and only when they serve a specific purpose. I just want to stress that downside isn't going away, it's just been, well, downsized. And this all happened years ago, so if you like modern Magic the level of downside is about where it's going to stay.

Problem #2 – It Creates Dead Draws

An important aspect of game design is something called tempo. The idea of tempo is that you want the game to keep moving towards its conclusion. Part of doing this is making sure that something happens each turn that infuses the game with some quality that helps propel the game forward. In Magic, the key device used to propel tempo is the draw. Every turn, you get a new resource that will allow you to do things you might not have been able to do before.

A second big problem with the legendary supertype is that it creates dead draws. What I mean by that is that if you have a legendary permanent on the battlefield and draw a second copy of it, it is useless to you, much as if you had drawn a blank card. (Yes, if the first copy gets destroyed you have a back-up, so it's not a completely dead draw, but it is more dead than your second copy of a nonlegendary card.) Lands can occasionally be dead draws in the late game, but we do a lot in design and development to give players things to do with their mana, so often a land isn't a dead draw late game.

The offshoot of this problem is that players tend to play fewer than four copies of their legendary cards. This is suboptimal in that it makes it harder to build around legendary permanents and makes drawing them happen less. Each of these things might seem small, but added up they start to have a significant impact.

Problem #3 – It Leads to Unfun Game Play

Let's say you like a creature and put it in your deck. The turn before you play it, your opponent plays the same creature. Let's examine the two scenarios, the first with a normal creature and the second with a legendary creature:

Scenario A (Normal Creature)
Your opponent casts the creature. On your turn, you cast the creature. Now both are on the battlefield. You each get to play with it. Perhaps the creatures will fight each other. Perhaps you'll race, each using them to hit the opponent. Perhaps you'll combine your creature with some other card in your hand to do something cool. In this scenario, you and your opponent each get to play with the card.

Scenario B (Legendary Creature)
Your opponent casts the creature. On your turn, you cast the creature. Both are put into the graveyard. Now no one plays with it. There's no interaction. There's no game play.

Let me stress that the kind of interaction we see in Scenario B is not all bad. The game does want to have threats and answers, leading to cards being played and then neutralized. The problem is that with normal Magic what gets neutralized gets moved around. The creature you Lightning Bolt one game often isn't the one you Bolt the next game. This variance allows cards to each get its time in the sun. Legendary permanents, though, can only neutralize one another. What this means is that they tend to spend less time on the battlefield, on average, than other permanents. This is particularly bad for cards that we want to shine a light on.

In addition, because the legendary permanents only work on one another, they tend to play more similarly from game to game. There is simply less variance in the play. Longtime readers will know that I feel variance is one of the most important aspects of making a game design fun.

Problem #4 – It Encourages People to Play Them as Answers

Another sign that the legendary supertype leads to bad game play: When a legendary creature gets good enough to appear in tournaments (and we do want our characters to show up in our game at the highest level), some players will start to put the legendary permanent in their decks as a means to deal with their opponents' copies of the card. This means that players are putting cards in their deck not because they want to use them, but because they need them as an answer to another card. This means that there are times even when you draw the card first that it can still be dead.

As you can see, the legendary supertpye has a lot of baggage. It doesn't play as well as we like and it creates numerous situations that we feel lead to unfun games.

"When I Get Sad I Stop Being Sad and Be Awesome Instead."

So here is the challenge that R&D has before it. The following must still be true about the legendary supertype:

  1. We want the supertype to refer to a card that is something unique. For creatures, that means that this is a specific humanoid or animal—a character.
  2. The execution of the supertype has to make some logical sense storywise. Legendary is as much about flavor as it is mechanics, so anything we do mechanically has to make sense flavorwise. Sometimes we are willing to gloss over flavor to get better game play. This is not one of those cases.
  3. The new rules for legendary have to make intuitive sense. When you explain the new rules, players need to feel that the explanation makes sense in that what it does feels like what it should do. I have spent numerous articles talking about the importance of mechanics feeling right. This is doubly so with the legendary supertype.
  4. The new rules have to be easy to explain. Solutions that hit the first three criteria but would take several paragraphs to explain aren't going to work. The idea has to be straightforward and relatively simple.

Most often I use this column to explain how R&D solved a particular problem. Today I wanted to demonstrate something different: that there are still problems that R&D is working through. For every problem we do solve there are two more hanging out waiting for a solution. Making Magic is a complex task, one that never really ends. I have no idea how long it will take us to crack this particular problem. Maybe tomorrow, maybe never. And that is an important lesson in how making Magic works: There's always a problem left to solve.

Join me next week when I get carded.

Until then, may you find solutions where others merely saw problems.