Working for Peanuts

Posted in NEWS on October 8, 2007

By Aaron Forsythe

Set out, cards firmly in hand, message boards a-twitter. Check. It is finally—finally—reality. You see, I was the lead designer of the set called Lorwyn—a job that darned near killed me. It was, without a doubt, the most difficult and humbling thing I have ever had to do as an employee of Wizards of the Coast.

The process was a long one and my notes are spotty; I'll do the best I can to reconstruct the important bits but in the end my reality may not jibe with stuff you read later from Mark or Doug or Devin. If there's ever a discrepancy, it's probably my error as, like I said, the process left my head swimming and my memory a tad jumbled. All I know is that I started the process almost two years ago with just a team and a list of nonnegotiables...

Setting Forth

Vivid_Meadow The team I put together was an interesting one. It began with Mark Rosewater, and his main role was that of mentor. As the game's Head Designer, he had standards and practices that I was expected to follow, and he was there to help me should I run into inevitable ruts. On top of that, he can flat-out design cards. The next most-important person on the team was creative director Brady Dommermuth. I had worked on set designs before that made decisions on flavor that didn't knot well with the work the creative team was parallel-processing, and the resulting battles over whose side should win out were rarely fun or productive. Because I knew how tightly tied flavor and mechanics were going to be in a tribal set, I wanted Brady there to help us make the crucial decisions and to guarantee creative buyoff on whatever plans the design team laid. I must say it was a smashing success. Next up is Paul Sottosanti. Paul isn't on the "Magic team" per se (working on digital games instead), but he has played his fair share of Magic and is a very enjoyable fellow. Plus, he'd shown an aptitude for design on other games, and Magic is always looking for new designers. Paul did such a fine job chipping in on Lorwyn that he was asked to lead design the following set, Morningtide. Andrew Finch rounded out the team initially. Andrew hadn't worked closely with Magic in years—instead focusing on new business opportunities for the company and then later, like Paul, digital initiatives—but he wanted a first-hand look at how Magic went about making his sets. As it happened, other commitments prevented him from being on the team for longer than the first few weeks, so I filled that slot with Nate Heiss. Nate, a one-time pro Magic player, one-time columnist for this and other Magic sites, typically works on board and minis games, but he certainly had a lot to contribute, and he has been working on more Magic sets in the time since.

As for the nonnegotiables... It was determined long ago that Lorwyn—at the time codenamed "Peanut"—was going to be a tribal set. The theme had resonated so well in Onslaught that we felt like it needed to come back, and "Peanut" seemed like the right time. Of course, just as Ravnica had the specter of the ever-so-popular Invasion block looming over its head, Lorwyn needed to find some kind of identity for itself such that it wouldn't be considered just a copy—or worse, a "bad version"—of Onslaught. The other nonnegotiable, which was finalized just before the design process began, was that the year was going to be divided into two two-set blocks as opposed to a single three-set block. The codenames for the initial three sets—"Peanut / Butter / Jelly"—lent themselves to a natural fourth—"Sandwich." But over time as it became clear to us that the two mini-blocks were separate entities, we altered the name of the fourth set to give us two pairs: "Peanut / Butter" and "Jelly / Doughnut." Mark was adamant that the key to making the two-block year work was to have the two "halves" be connected somehow, subtle or otherwise, and it was up to me and my team to figure out how to do that (luckily Mark was on the team, so he could help solve his own problem). Of all the hurdles I faced working on Lorwyn, finding the big "hook" that would differentiate it in players' minds from other sets was the most difficult. And I don't think it has worked out quite right... yet. Lorwyn is, in many people's minds, a tribal set. Period. How will it ultimately end up different that Onslaught on a big-picture level? That will all be clarified as the year unfolds.

When I look back at all the time spent on Lorwyn and still how rushed everything felt as important hand-off dates approached, I find myself wondering, "What were we doing all that time? How did we spend all those months?" The truth is that, while the actual set we were working on was crunched, the other three sets benefited from all of our effort. The main hook for Morningtide was worked out. The theme and two of the keywords for "Jelly" (now known as Shadowmoor) were designed.

Populating the World

As for the specter of Onslaught... To me, the easiest way to give this set a different feel from Onslaught would simply be by choosing different tribes to focus on. With the "race / class" model firmly in place as of Mirrodin, I wanted to just use "races" as our tribes. No Soldiers or Wizards—after all, those are jobs, not races, and it is the races that would most allow the chosen types to feel different from one another.

Amrou Kithkin

The team quickly agreed on Elves as the green race because they are an integral part of Magic, they are familiar to all fans of the fantasy genre, and lots of players love them. On the same note I tried to dissuade the group from using Goblins again, citing that they used to be funny and cool until the Onslaught block made them an oppressive killing force. People specifically ask not to face Goblins on Magic Online, which means they definitely have a PR problem. I figured that if we tried to do Goblins again we'd either be compounding the problem or be making weak cards in an attempt to avoid compounding it. In their stead in red, Brady suggested the idea of "fire people" or "elemental people." Rosewater had long been championing the return of Merfolk as the blue race, and Brady—Merfolk's mortal enemy—was willing to play along. We needed some large creatures and didn't want to repeat the "Beast" overload that Onslaught had, so we chose Giants. For the white race, Brady mentioned Kithkin offhand one day, perhaps as a joke. There had been a grand total of one Kithkin printed at that point—Amrou Kithkin—but for some reason the idea really, really resonated with me. White needed its "Goblins" or "Elves"—a non-human race that players could latch on to, and Kithkin fit the bill. Humans inevitably came up, perhaps instead of Kithkin in white, perhaps as the black race, perhaps in all colors. But we still couldn't stomach the idea of making cards that said "All Humans get +1/+1," so we talked ourselves into keeping Humans out of the block altogether (except for changelings, but that came later). I imagine one day we'll buckle and make a "Human lord," but Lorwyn is not that day.

With that incomplete list—Elves, Merfolk, Giants, Elementals, and Kithkin—Brady went to the creative team and started figuring out what kind of world they could create that could play host to this weird variety of beings. It is in those meetings that the "Lorwyn" that you all are familiar with was born—a lighthearted world of Celtic / British fairytale. That choice of setting then informed our choices about what the other tribes should be. A fairytale world needs Faeries, does it not? Luckily we were looking for some kind of fliers to populate the set—a perfect match! Treefolk were suggested as another large race to complement Giants. But what of black? Zombies? No, this world was no place for the undead. How about Imps? Imps in Magic traditionally fly, and flying Imps are awfully close to how we hoped to portray the Faeries. So what if we made a new race, something from British folklore, like boggarts? Sounds promising! We talked that idea over for a while before coming to the realization that "boggarts" are essentially Magic's Goblins in disguise—same motivations, same temperament, same size. Could we do Goblins in black as opposed to red? That sounded reasonable to me—new Goblins for people to enjoy that wouldn't necessarily add to their overall power in older formats.

From there we locked tribes into primary and secondary colors—making sure each tribe was in more than one color was important to Mark as another way to distance this set from Onslaught; he spoke on the topic previously in his column Lorwyn at all Costs, and he elaborates on it today. After that, we decided to sketch out how each tribe would play. The difference in tribal identity is what I consider one of Lorwyn's biggest victories. You see, in Onslaught, the tribes just did what the colors did. Goblins did "red things." Wizards did "blue things." And so on. But, using the techniques that were so successful for us in Ravnica, we decided to give the Lorwyn tribes more specialized identities so that (a) the different decks they spawned wouldn't just be hordes of 1/1s and 2/2s attacking and blocking all the time and (b) the different game-play scenarios we'd enable would allow a wider variety of players to enjoy the set.

I'm not going to get into what we chose for each tribe and how each worked out—that's all grist for Mark's and Devin's future columns. But I will say that it made designing the set super fun.

Once we knew each tribe by name, colors, and play style, we then started making cards for them... in Time Spiral.

I've written about cross-block synergy before (Ending Block-on-Block Violence), and we still believe it to very important that cards from one block work well with previous and subsequent blocks. Gone are the days of Odyssey showcasing none of the same tribes as Onslaught, or Mirrodin being followed up by a block with zero artifact creatures. We wanted the Lorwyn tribes to be "prepopulated" in the Standard environment. To that end, Amrou Scout was made a Kithkin, Coral Trickster was made a Merfolk, and Desolation Giant was added to the "timeshifted" cards. Heartwood Storyteller is a Treefolk. Merfolk Looter is in Tenth Edition. Llanowar Mentor and his cycle of friends all make Lorwyn-relevant tokens. Yes, we were working on "Peanut" that far in advance. I love it when a plan comes together!

You're Welcome, Tarmogoyf

Not a Goblin!

The next step to giving Lorwyn its own identity was to make the concept of "tribal" mean more than what players were used to. I looked for answers in a question—a question we've all either asked ourselves or heard newer players ask: "Is Goblin Grenade a Goblin card? Can I search for it with Goblin Matron?" We all know the answer to that question is, sadly, "no." But does it have to be? Isn't Goblin Grenade exactly the kind of "Goblin card" a crafty Matron should be looking for? Can't Goblins have more in their arsenal than just more Goblins? Why not give each tribe some race-specific tools that all tie together? Create new kinds of magic—"Elf-magic," "Goblin-magic," and so on?

I want to make it clear that when we were coming up with this idea, we in no way wanted to make a new card type. All we were after was putting creature types onto noncreature cards, like this:

Land - Goblin

And that was, in fact, how we mocked up and played the cards initially. I loved how it all felt—we gave each tribe a "Matron" (later downgraded to the top-of-library Harbingers) plus several other cards like Lys Alana Huntmaster, Boggart Birth Rite, and Drowner of Secrets that let you make use of and interact with non-creature cards that had creature types. We even tried fleshing out a whole mess of cards that could be used by revealing a card of the relevant type from your hand; most of those didn't make it, but we are left with the brilliant Paul Sottosanti design of Goldmeadow Stalwart, the rest of his cycle, and the rare dual lands. Creature types on noncreatures was an amazing innovation and something I knew we'd want to use.

Of course, the best-laid plans often crumble in the face of actually having to implement them. Rules Manager Mark Gottlieb caught wind of our designs and immediately got nervous. Within a few days, he had given the design team a seven-page write-up about why our way didn't work and what we could try to do to solve it. Here's the beginning of that document:

Subtypes exist in strict correlation with their types. Except for instants and sorceries, which share subtypes, there is strictly no crossover between one type's subtypes and another type's subtypes. The game has been built this way for 13 years. What has this meant?

Since the correlations are unique, all subtypes imply their types. If a card is a Goblin, it is by definition a creature, so "creature" hasn't needed to be said.
--Example: Fever Charm says "Deal 3 damage to target Wizard." Damage can't be dealt to a noncreature permanent, yet this card doesn't say "Deal 3 damage to target Wizard creature." It doesn't have to. "Wizard," by definition, implies "creature," so this shortcut can be taken.

Due to these correlations, unique subtypes can be used mechanically. Look at how the Shrine subtype was used on the Hondens, or the Locus subtype on Cloudpost. In the discussion about Planeswalkers, using their subtypes as a means to denote identity and uniqueness seemed quite promising. (All Planeswalker cards that represented Jace would have subtype "Jace," and their uniqueness rule would be based on their subtype, not their name.)

Each subtype makes flavor sense for its type. Furthermore, the subtypes don't merely describe or classify their cards, they say what those cards actually are in the Magic realm. A "Creature — Goblin Shaman" is a goblin shaman in the same way that I'm a human gamer. Honden of Infinite Rage is a Shrine.

As soon as "Land — Goblin" is printed, these distinctions crumble.

He then goes on to explain various proposals that would allow us some of the functionality we wanted but not as cleanly. Massive errata on many, many cards was proposed. The use of parentheses in type lines was proposed. (But a new card type wasn't proposed.)

It just didn't seem worth it to us to rip out the wiring of the game rules just to make a dozen or so cards work. So we went back to the drawing board, and tried to mark that spells and lands were attached to tribes in a different way. We kept coming back to either using a special word or symbol for each tribe—so "Elf Instants" might say "Instant – Elven" or "Instant – Elf-Magic" or "Instant – [a symbol that looks like an Elf head]". Crazy, huh? Of course, to make this work, cards like Wren's Run Vanquisher would have to say "As an additional cost to play Wren's Run Vanquisher, reveal an Elf card or an Elven / Elf-Magic / [symbol] card from your hand or pay 3." Hardly elegant. On top of that, under this new system "Elf-Magic" cards wouldn't be backward compatible with older cards that looked for Elf cards, like Sylvan Messenger. Boo! It would be "Arcane" all over again!

We were just about ready to scrap the whole idea, but I couldn't give it up. It was too important to the set's identity. I said, "Let's keep them in, keep playtesting them assuming they work, and pray Gottlieb can come up with another solution."

Gottlieb wasn't thrilled to hear that he was being tasked to re-solve a problem he had already spent a lot of effort solving. But to his credit he eventually did come up with something that worked, something inspired by his own opening paragraph—a new card type. If a new type was created, that type could share its list of subtypes with the "creature" type and just be added onto existing cards, just as "Artifact" can be added to "Creature" or "Land." Under the new plan, types retain their specific correlation to subtypes, which is the lynchpin for making the rules hold up.

Gottlieb was aiming to name the type something unique and bold, like "anima," but I was nervous about calling too much attention to what was essentially a bookkeeping device, and hoped we could use something more low-key, perhaps something already part of the player vocabulary, something like "tribal."

The errata necessary to enact the plan was minimal, and Future Sight gave us a great place to debut the new type—that set already had a noncreature card that "counted as a Rebel card in all zones" independent of all the stuff the Lorwyn team was working on. I know the initial reaction to "tribal" was that it feels like a supertype and not an actual card type, and I agree. But the rules needed it to be a "type" proper, and I hope I've explained it at least well enough for you to take my word for it.

Treasure: 100 Points

Looking over Lorwyn now, I'm pleased with a number of good decisions I made regarding the design and structure of the set—at least I can say the ideas were good enough to survive development. I wanted exactly eight gold cards in the set—one legend per tribe, and those cards turned out pretty hot as a whole. I came up with the Command cycle, as well as the Vivid lands, and the beginnings of the "Cloudgoat Ranger" cycle of Deranged Hermit knock-offs. Good stuff. Of course, I made plenty of less-than-good decisions as well, all of which were slowly weeded out in design and/or development. There were mechanics in the set at various times called "assist" and "mojo" that were rightfully dropped or saved for later. The rare dual lands were at one time both tribal and legendary. But the biggest barking-up-the-wrong-tree mishap I can lay claim to is the ill-fated "treasure" mechanic.

The origins of the treasure mechanic begin way, way back in design, just after the setting of the set had been established. Once we knew we'd be in a pastoral fairytale world, the creative team prodded us to see if we could alter the set's makeup to reflect the "nicer" nature of the plane. Could there be less removal in general? Fewer giant monsters? Less emphasis on attacking? We actually tried accommodating most of these ideas at various levels in early design, but we eventually returned to normal on most accounts, realizing it just isn't worth altering the fundamentals of gameplay just to help flavor.

In looking for ways to make the game less about fighting, I tried to think of other things creatures could be used for in this world. My mind settled on "adventuring." Why can't creatures explore stuff, discover new lands, amass some loot? That's what I'd do if I lived in a fairytale world! So I looked for ways to make the concepts of exploration and treasure work using cards. Here's what I came up with in a nutshell:

As CARDNAME comes into play, add three cards to your treasure pile. (Your treasure pile is made from face-down cards from the top of your library.)
T: Add W to your mana pool.

Kithkin Explorer
Creature - Kithkin Scout
T: Explore (Reveal a random card from your treasure pile. You may pay that card's converted mana cost. If you do, put it into your hand. Otherwise, return it to your treasure pile.)

Those two cards are both (a) parasitic as all get-out and (b) flavorful as all get-out. It seemed worth it to me.

When I first added it to the file, there was a genuine buzz. People really liked the idea of it. The word "treasure" had a ring to it that just sounded like a natural fit to both Lorwyn and Magic as a whole. Using creatures to dig rather than fight—and then the fun moment of seeing what you found—was really novel.

The mechanic had problems, sure. We eventually took "explore" off of specific creatures and just made it a property of treasure—any creature could explore. That made the games less swingy based on who could draw their explorer and keep it alive. Of course, suddenly games were all about drawing tons and tons of extra cards. We had to power the lands down, first making them come into play tapped, then lowering the number of cards they gave you from three to two to one.

We tried having one big "treasure pile," we tried having each treasure land have its own stash that you had to keep track of; creatures could be sent to explore specific lands. We made the lands tribal so that you could search for them with Matrons.

Exploring went from instant speed to sorcery speed. Unpaid treasure went from being shuffled back into the pile to just staying on top of the pile to just going to the bottom of the pile.

Creatures were made that added treasure to piles, creatures were made that died to the treasure pile instead of the graveyard. Heck, one Giant even threw treasure at people for damage. All of the artifacts in the set had some tie to treasure, some triggered ability that went off if someone found them buried in the ground. (For instance, if a creature dug up a Moonglove Extract, it did two damage to that creature and was put into the graveyard. Oops! A trap!)

I had such high hopes for the mechanic, I really did. It was a big part of the set and really gave it a uniqueness that made a lot of people smile. But none of that feel-good stuff could save it from the fact that it was, at its core, a bad mechanic.

Treasure required a ton of wordy, parasitic, linear cards to work right. It was essentially degenerate in constructed, and made people want to build decks full of land and cheap cards as opposed to big cool ones, but we needed the converted mana payment on there so that people would care about what cards they found. As developer Erik Lauer was fond of saying, "My treasure is... a Plains! Yes! I'm so happy it wasn't some stupid Dragon!" It bogged down Limited to the point of absurdity, and it made weenie decks into card-drawing machines in Constructed. Every time an opponent led off with a treasure land, you just rolled your eyes. The sad truth is that being fun once or twice isn't good enough for a Magic mechanic. They need to be fun a hundred times over, and treasure failed at that.

I was stubborn regarding the mechanic and kept trying to make it work, long after people like Lauer and Matt Place and Mike Turian felt it should have been taken out of the set completely. I eventually gave up, but not before wasting at least a month of everyone's time clinging to hope.

Treasure eventually left the set, long into the development cycle. The hole created by its absence was eventually filled by clash and evoke—fine mechanics in every sense.

Mosswort Bridge

All through the treasure process, Mark Rosewater was continually suggesting that I step back from the mechanic and try to isolate the thing I found most fun about it to see if I could salvage something along those lines. My answer was that having a hidden stash that you can work to unearth was very fun—a sort of new direction the game hadn't taken. He said to think about just that part, and I did that once the mechanic was axed, stewing over how to capture that feeling in a way everyone could be happy with.

I came up with what would become hideaway. The hideaway lands, like Mosswort Bridge, are meant to capture some of the "hidden treasure" feeling of the original mechanic. But now, instead of just having a creature explore, each treasure comes with a mission attached—a mission you need to complete to access the treasure. And they make for some enjoyable game states: The Bridge is out; the card nestled under it gently taunts your opponent with the allure of secret information. Suddenly the game has two objectives for you—one, kill your opponent; two, get ten power of creatures in play. Feel free to prioritize those however you want. Similarly, your opponent, assuming he's afraid of whatever might pop out of the Bridge—for free!—has an incentive to keep you off of ten power as long as possible. It's all very fascinating.

Great Designer Ken Nagle calls Mosswort Bridge his favorite card in the set. I consider that a victory, and I hope some of you will come to enjoy the mini-games associated with the hideaway lands in the coming weeks.

So Much More

Those are some big stories, but there's so much more to tell. Thankfully there are regular columnists on this site that will have no shortage of material in the coming months. I haven't even touched on planeswalkers or champion. Changeling Week is coming up—that group of cards certainly has some interesting stories attached to it. Clash went through some serious revisions. Mark already talked about evoke. Wow, this set has it all.

With this article, I can finally put this set to rest from an R&D standpoint. I feel like Bilbo Baggins dotting the last period in his book, closing it up as the credits roll. As for credit, I want to thank Mark Rosewater and Bill Rose for their guidance, Brady Dommermuth and Jeremy Jarvis for working with me to make this whole thing come to life, and especially to Devin Low and his developers for letting me stay so hand-on as we sorted out the various messes. I'm thankful that I got to contribute so many cards—Oona's Prowler, Gaddock Teeg, Galepowder Mage, Incremental Growth, Forced Fruition, Ashling's Prerogative, and more—during development when I wasn't even on the team; it really makes the set feel like its my own.

As for my favoritecards... I'm going with Boggart Loggers and Hunter of Eyeblights. They aren't all that good, but one of them is kinda funny and the other is a flavorful little game all wrapped up in one card. And as long as Magic remains funny, flavorful, and fun, I'll be around.