This is a version of an article that appeared in the January 2013 issue of Game Developer magazine.

When Magic: the Gathering first entered the gaming scene back in 1993, the mere idea of a game based upon an ever-evolving pool of collectable cards was just a zygote of an idea. Richard Garfield and the rest of Wizards of the Coast knew the game had real potential, but no one really knew how the game experience would really play out.

It’s not surprising that they got some things wrong. Their limited playtesting was not nearly enough to find all of the convoluted strategies that players would devise, and they had no historical data to look at problem spots. Sophisticated analysis of the game did not yet exist—they did not know (or fully appreciate) how powerful drawing cards would be in their game, and thus printed a cart that allowed a player to draw three cards for one mana.  And many of the rules were written to be ambiguous, so new expansions that introduced new rules brought in unexpected conflicts, and made it clear that card rules language needed to be much more structured and unified than it was previously.

They also underestimated their own popularity. They expected players to buy a deck and a couple of booster packs. Hardcore players started to buy booster packs by the case. Rare cards that Wizards of the Coast assumed would only show up once or twice in a deck ended up being highly sought after, and soon devoted players were packing 4 of each (the legal limit) in their decks, and destroying their less-invested opponents in the process, often in a couple of turns. The value of the best rares shot into the stratosphere, creating a legitimate aftermarket for cards.

Wizards has succeeded beyond their wildest dreams, reinventing the board games industry (and in the process saving America’s gaming and comic book stores) in the process. But it was clear that Magic had some bad structural problems that would need to be addressed. Fortunately, Magic had a winning core game design, which gave them the resources and time they would need to fix these structural issues. Magic was a game that had a long life after ship, and their game designers took advantage of this to great effect.

Finding the Design Space

Most games have a life after they initially ship. For most board games and single player video games, expansion packs are the norm. Thanks to living in our age of connectivity, this trend is now buttressed by downloaded content (DLC) for games on online platforms. Some games, though, are uniquely suited for this sort of life after shipping. Collectable card games, such as Magic: the Gathering and Pokemon, are built hoping to capitalize on this as a core business model.

Massively multiplayer titles, such as World of Warcraft and Star Wars: the Old Republic, are also games that beg for a post-launch strategy. Most MMOs opt to offer players occasional downloadable content (often for free, but sometimes for a small charge), and some buttress this with an occasional expansion pack. Again, the business model is central to the approach here; subscription models want to keep players engaged and happy for long, continuous periods of time, whereas games that depend on microtransactions are content with people stopping back for a brief visit, to purchase and consume the new content.

All of this is obvious, of course. What is less obvious is that designers need to identify their post-launch realities. If successful, where is the design space that allows the designer to expand the game? Is this a free expansion you expect everyone to acquire? Is it paid for? Is it premium expansion content that will go on a store shelf, or is it DLC? If it is a competitive game, are you expanding in ways that are mostly cosmetic, or are you encouraging players to invest in making their characters more powerful? If characters are getting more powerful, is there actually room in the math for that?

Pure Escalation

In World of Warcraft‘s first three expansion packs, Blizzard simply made the numbers bigger. Players wanted more levels, more and cooler gear, and tougher challenges. However, there were some complications. For one thing, the stats of players had already been increasing—Blizzard had continued to release new raid content, and added new, improved gear for each tier of content they added to lure players into that content.

This meant that people right off the boat in new expansion areas started in a completely different place. Those who had been raiding would play through a significant chunk of the new expansion content without getting any gear upgrades at all, whereas those who hadn’t would find “standard” quality items comparable to raid gear off of trivially accomplished quests, in a desperate attempts by the designers to close that gap.

But there was a more insidious problem: The power curve in the game was exponential in nature, which meant for the math to work naturally, the gains between levels needed to grow bigger and bigger. As of this writing, a top weapon in the game now offers +947 Stamina and deals 1996 Damage Per Second, and this number easily grows to five digits when combined with abilities and modifiers. By comparison, Perdition’s Blade (a top drop off of a Ragnaros from Molten Core, an early raid) did 61.7 DPS, and has no stat bonus (other gear he drops offers no more than +22 stamina). The number of hit points that a raid boss today needs to have to survive 25 players dealing out that much DPS over the course of a ten minute fight is staggering, and the numbers involved are large enough that the development team posted a blog to open discussions with the players about how to bring these numbers back into a sensible reality.

Adding Complexity

Hardcore players who love a game frequently don’t want just more of the same game. This works in some genres, such as first-person shooters and RPGs, where players can be satiated with expansions with more maps and more story. More often, though, these players want more mechanics and rule tweaks—even those that love the game are bored with it after a certain period of time. But this, too, raises dangers, mostly of overcomplicating the game for casual players.

This happens rather frequently in board games; lots of board games get expansions, and in many cases, the expansion packs make the game better. Kingsburg took the opportunity to patch the rules, fixing a couple of issues that allowed more potential strategies to emerge. Game of Thrones’s second expansion fixed balance issues so fundamental to the game that they were made part of the core gameplay in the second printing of the game.

All of this complexity can make the game impenetrable to new players. Seven Wonders is a card game that was widely praised by the board game community as an excellent short and casual game, in large part because it was quick to play and easy to teach, which was ideal for playing with casual audiences and/or while waiting for latecomers to show up. They’ve launched two expansions since the original release, and those expansions have added tremendous depth to the game, but the game is no longer a casual-friendly game that can easily be explained to new players in less than five minutes. A bewildering array of new cards and symbols need to be taught, and even experienced players need to reach for the rulebook.

The issue is that the two audiences need different things. Hardcore players need new patterns to learn and new strategic elements to challenge them.  A new player, on the other hand, needs to pick of all rules, addendums and strategies in one massive information dump that the more experienced player has accumulated over years.

The hardcore fans of the game, of course, come at expansion content with all of their accumulated knowledge to date. Raid fights in World of Warcraft in the latest expansion are orders of magnitude more complex than the fights they initially shipped with, and experienced players welcome facing these more intricate challenges. However, this means that new raiders have a much more difficult learning curve to overcome than they did in that first round of content. This accumulation of difficulty is often forgotten, although Blizzard has found they need to make a significant effort to reduce this learning curve as much as possible.

Planned Obsolescence

Magic: the Gathering started in a different place. As they weren’t expecting players to chase and build all-rare decks, they discovered that the game was simply way too fast once players did so. Players could frequently win in three or four turns. Printing more powerful cards would accelerate the problem. Wizards of the Coast tried printing cards that were less powerful, and discovered, perhaps unsurprisingly, that they wouldn’t sell. They quickly hit upon the idea of a Restricted list for cards (cards that could not be used in tournament play), to weed out the ludicrous cards, but the play environment was still too fast, and worse, the sheer number of interactions that were available in the pool was becoming impossible to track and design for as new expansions rolled in.

What they did was to introduce obsolescence to their game design, with the establishment of “Standard” as the primary form of Tournament play. Standard tournaments allow only the inclusion of the latest 2 “blocks” of cards (where a block is comprised of up to 3 expansion) and the most recent “core” set to come out, effectively limiting the card space to a fairly manageable number. This had a lot of obvious upsides: By waiting a few expansions, the game could be slowed down as those older, more powerful cards disappeared from the sets. It also created an incentive for the players to continually buy new cards, and not just be content with their old collection.

Planned obsolescence had a lot of subtly deep design implications as well. Every block could have strongly different metagames; in one block, counterspells would be common, and in others, they’d be hard to come by. The designers could make major changes to the design philosophy of the game, as they did when they redistributed various spells to different colors in order to even out the usage of those schools. And perhaps most importantly (to those of us designers who fancy ourselves as mad scientists, anyway), designers could take bigger chances with weirder mechanics, knowing they would eventually rotate out of the pool.

Planning for Growth

This is not hugely applicable for those of us making digital games, except for one core, inescapable fact: Magic, a game that struggled out of the gate under the weight of many large structural mistakes, is now designed with expansion in mind. It is designed, at its very core, to grow, to change, and to reinvent itself with every expansion. This level of foresight serves it well: Magic will reach its 20th anniversary in 2013. The latest Magic Expansion,“Return to Ravnica,” is selling like gangbusters, and high-level magic players now play for thousand-dollar purses in sanctioned tournament play.

We no longer live in a world where any game is every, truly, finished. If your game is beloved by fans and sells enough to satisfy the suits in your building, there will be a desire for expansion content, for booster packs, for downloadable content, and even for cosmetic skins. Designers are well advised to plan for their success, and to know where, exactly, the game’s design allows them to go to satisfy these appetites.