In terms of Constructed, Standard is the format we give the most consideration to in the development process. The reasons are pretty simple—it's the most popular Constructed format by sanctioned matches played, and it's more reasonable for us to test than Modern or Legacy. Most developers spend at least eight hours a week building decks, playing them against each other, and meeting to discuss what is happening in our Standard metagame. It is rare that during one of our weekly Future Future League meetings that we don't tweak a handful of mana costs or powers and toughnesses, especially in the first few weeks that a set enters the FFL.
The Balancing Act
One of the things that surprised me when I started working here was that development really doesn't work on balancing the individual decks the way I'd imagined. I'd expected a lot of work going into planning out which decks would be making up the real-world metagame, but that isn't the goal of the Future Future League. The goal is to have individual cards that are at around the right power level, and let the players in the real world build the metagame that those cards will allow. The reason for this is pretty simple—we are wrong about enough things, far enough in the future, that if we were trying to carefully sculpt the decks needed for a perfectly balanced metagame, then we would inevitably fail (probably in spectacular fashion), and it would take months to print a new card to help solve it. For example, the FFL was, in general, based around Wolfir Silverheart being the five-drop of choice, not Thragtusk. While I do think the Silverheart deserves more attention than it is currently getting, I'm also glad that our process doesn't revolve around balancing our card sets around Wolfir Silverheart. It means that when Thragtusk is the popular card, it doesn't crash the real-world metagame. Instead, it is different than we had predicted, but still healthy.
In hindsight, it may seem crazy to some that we weren't playing as much Thragtusk as the real world, but the decks in the Future Future League are rarely perfectly tuned. We may be able to get within 80% or 90% of what the real world puts up in its second or third week, but the time spent going "tall" on tuning a few decks provides less utility than going "wide" and trying out more and more decks. The cards we work on are constantly changing, so a lot of work on fine tuning would be instantly lost if we change a card that was a little too good as a 4/4 to a 3/4, or add a mana to the casting cost or an activated ability.
There are also a lot of decks we create in the Future Future League that never see play in the real world. Sometimes that's because we altered one or two cards, and the super-powerful version of the deck will never exist in the real world, and sometimes it was just that our untuned version didn't gain much from being properly tuned and is never able to break out into a top-tier deck. Relying on these decks existing in the real world for the metagame to be healthy is a dangerous road to go down. Making sure the cards themselves are balanced is a much safer way to ensure a healthy metagame can come into being.
Building a Healthy Metagame
As I mentioned before, we have found it generally best to create the building blocks of a healthy metagame, instead of carefully crafting what that metagame will look like in the attempt to create a perfectly balanced game. That doesn't mean we don't put a lot of work into those building blocks, however. This is a huge part of the job of all of the developers, and one that we have progressively gotten better at over the years.
Games like StarCraft have been lauded in the past for having a perfect balance among each of its factions, which has brought questions from people on why don't we work toward that goal? Simply put, StarCraft and other video games have a huge advantage in the balancing realm that we don't—they get to patch units that hundreds of thousands of collective hours of play reveal are overpowered or underpowered. We can't send out stickers to change a sorcery to an instant, or remove or add a mana to a card's cost. We're stuck with what we've printed. There are always going to be things we wish we could change after a set comes out. For example, if we could go back, we'd probably change Boros Reckoner to deal his retribution damage to another target creature; that would make any possible infinite loops with him being indestructible and lifelinking much more difficult. As it is, we have to trust that the metagame is robust enough to handle a three-card combo like that existing, which I believe it is. It would be nice if we didn't have to rely on the metagame to handle things like this, but having the looser balance that we strive for makes us far more resilient when interactions like this aren't caught in development, and I believe makes the game as a whole stronger.
So, what do we do to create a healthy metagame? Here are the basics:
1. Make sure decks have enough to do
Former developer Zac Hill wrote in an article last year about the "buckets" we use to describe different strategies. For those who don't want to read a whole article for the sake of a paragraph, I'll cut to the chase and tell you they are Aggro, Midrange, Ramp/Combo, Control, and Disruptive Aggro. Each of these buckets represents a core strategy that most Magic decks fall into. One thing we focus on while playtesting Standard is making sure there are enough decks to fill these buckets. We don't want to focus too heavily on making sure that the aggro Zombie deck has the right win percentage against, say, the Esper control deck. Instead, we want to make sure there are enough cards in the environment to make all of these decks possible, and that no one strategy gets way too much. If nobody in the FFL is building new control decks, or has any desire to play them when we have a tournament, then we need to look at what is happening and figure out what the problem is. It could be that the aggro decks are too powerful and we need to adjust some cards, or it could be that the control decks are missing powerful late-game cards to make them more attractive than the just-midrange decks. We strive for an environment where all the people in the FFL are happy playing different decks and enjoying the game play.
2. Give decks room to evolve
Once we know the decks in the format have enough cards to compete, we also need to make sure they have room to move around. In most metagames, what this comes down to is making sure there is enough breadth in the card pool to allow decks to adjust week by week to better combat other decks. We don't want decks that have enough cards to fill out a seventy-five, but just that. There should be different ways to build decks to fit better into the real-world metagame.
One of the most noticeable ways this occurs in the real world is within control decks, and the question of tap-out control versus counterspell control. Traditionally, tap-out control is better against creature-based strategies than counterspell control, since most control decks tend to board out their counterspells for more removal in Games 2 and 3. In some metagames, this leads to people who start with their decks in that configuration and board in counterspells (or maybe not even that). This will result in suppressing aggro, since it tends to prey on control, and prop up combo and ramp, which are weak to counterspells. During this period, many aggro decks will move away from having a lot of low-drop creatures to having higher-casting-cost spells to capitalize on the lack of counterspells from control and being stronger against midrange. At this point, the person bringing the control deck with counterspells to a tournament will be advantaged in the format and will have a better chance of winning. So the cycle of life continues, and the metagame evolves from week to week with decks adopting new strategies to combat what other decks are doing.
The beauty of this whole dance is that it isn't something we have to artificially create. It happens naturally—we just need to give it the room to exist. Ideally, we don't want to try and force people's hands at anything if possible. The real world is much better at crunching numbers, and trying out new things, than we are. Instead, we have learned over the years what tools it needs to exist, and we just need to supply those. The real world also comes up with far more interesting and fun decks than we do, given the time. The game of Magic that is just players spending their time trying to figure out the decks we seeded for them isn't the same game we are playing today.
3. Insert safety valves into the format
The last thing we have to make sure of is that no one deck or strategy can get out of hand. This often means putting narrow cards into the format that are generally not quite good enough in a vacuum to see main-deck play, but are amazing if the metagame goes too far in any one direction. Slaughter Games, for example, isn't very powerful in most metagames, but if many decks rely heavily on one card, then it quickly becomes a sideboard card, or even main-deck card, to punish those strategies.
Perhaps the most common form of this is graveyard hate. One of the first major Standard tournaments using Gatecrash was won by a deck that comboed out using Angel of Glory's Rise returning numerous Humans, including Fiend Hunter, Huntmaster of the Fells, and Cartel Aristocrat, setting up a loop that would end with the player at an arbitrarily large life total and a theoretically unbounded number of Wolves in play. This was a combo that came up during our FFL testing but didn't prove to be too strong to deserve anything being taken from it. We like it when decks like this exist, assuming they aren't overpowered. They add a lot to the interesting aspects of the game and are more than worth the risks associated with combo. While this was a deck that we caught in our testing, it's possible we could've missed it. This is one of the reasons we put safety valves in the format—so when we do miss one, the format has the tools to deal with it without us banning a card or printing some overpowered hate a year later (at which point, the damage is already done).
If a graveyard deck turns out to be too powerful, then players will move more toward playing cards like Deathrite Shaman in their main decks and cards like Rest in Peace or Tormod's Crypt in their sideboards. Similarly, we generally have more than enough artifact, creature, and enchantment removal in the format that people will have places to go if those cards become too powerful.
There comes a point in the development process that is pencils down and we have to send the cards out to editing, localization, imaging, and—eventually—printing. That is the point where we have to trust that we've done what's within our power to make the set fun and balanced, while executing on the vision of the design team.
Working almost a year ahead of the real world is something that takes some time to get used to. It's kind of like shooting arrows at a target, but you can't use your previous shot to gauge the wind and instead have to use the arrow you shot yesterday. Luckily, we've gotten pretty good at it. Certainly, there are cards we find out much more about between that time and the time when the set actually gets in your hands, but it's often not until the cards hit the streets that we really know exactly how the set we developed will play out. We have our favorite cards and hope to see them get as much love as we think they deserve, but we've come to expect the unexpected. In that regard, we are just as excited as most of you are to see what changes come out of the Pro Tour, because it gives us better data and helps us to make better sets in the future.
If you have time this weekend, check out Pro Tour Gatecrash, but even if you don't, I hope you spend some time with the set and enjoy it as much as we enjoyed putting it together.