How to Fix Play Design

Ryan Normandin
August 26, 2020
0 Comments

This week or next, Zendikar Rising spoilers begin, signaling the end of one of the most tumultuous years in Magic's 27-year history.

     

      

      

   

No discussion of this year would be complete without mentioning the countless cards that have been banned (alongside the 10 Companions that received functional errata) in Standard, Pioneer, Modern, Legacy, and even Vintage. This was the year that Wizards of the Coast seemed to forget what makes for good games of Magic, instead leaning hard into cheating on mana, restricting the opponent's ability to play the game, and building cards that lead to repetitive, drawn-out gameplay.

  

The implementation of Play Design in 2017 after Wizards missed Saheeli Rai/Felidar Guardian was heralded as the beginning of a new age of balance. That promise, unfortunately, has not been kept. The very first set that Play Design touched, Dominaria, brought Llanowar Elves and Goblin Chainwhirler, which would warp the format around themselves the entire time they existed. Guilds of Ravnica and Ravnica Allegiance were bright spots, but then we got War of the Spark, Throne of Eldraine, Theros: Beyond Death, and Ikoria: Lair of Behemoths, each bringing new catastrophes to the Standard format (and Pioneer, and Modern, and Legacy…).

  

Comparing the formats before Play Design to the formats after, it does not appear that, overall, Play Design has led to an improvement in competitive gameplay.

Of course, this is a difficult claim to make because it's impossible to know what the game would look like without Play Design. Maybe the team actually caught things far worse than Oko and Wilderness Reclamation. However, the fact that the current format was supposed to contain a deck with

is extremely concerning.

I believe that Play Design is both competent and passionate about making the best game that they can. Calls on social media for firings and accusations of personal failures are unproductive, unnecessary, and would not lead to better Magic. The people who have made the greatest number of unique mistakes have learned the most, after all.

This leads us to a problem in explaining the current situation.

Premise 1: Play Design shoulders most of the responsibility for competitive balance.

Premise 2: Play Design is competent and passionate about doing a good job.

Premise 3: Competitive balance has been poor for most of the time since Play Design has existed.

So what conclusion can we draw? It would seem that, if these three premises are true, something is internally broken at Wizards of the Coast, and/or building metagames is so complex that Play Design is understaffed.

What are some examples of something being broken on the inside at Wizards? Perhaps, as one of Hasbro's only profitable IP's, they are being squeezed for every penny, being encouraged in either a systemic or a personal way to rush and throw caution to the wind, to take riskier risks than they might otherwise. Perhaps Hasbro is focused, to the game's detriment, on the short-term gains rather than long-term stability.

If so, there's nothing that WotC can do about that, so let's focus on the other explanation, which they can do something about. Consider, as an example, “The Case of Siege Rhino.” 

In this article (highly recommended read), Sam Stoddard discusses the origin of Siege Rhino, and it all begins with First Response. Originally two mana, and then three mana, First Response was dominating the internal metagame in conjunction with Mana Confluence. When Checklands in M15 were swapped with Painlands, First Response only became more busted.

Stoddard details how even master deckbuilder Gerry Thompson continued to build decks with First Response even after it was pushed up to four mana. Was it still playable? Probably not at this point. However, as Stoddard says:

“…it's a lot easier to get onto playing a four-mana card that used to be three mana when that version was very strong. The card just looks a lot stronger to us because we've seen what it can do. In the outside world, nobody knows the card used to cost three, or even two mana, and it's much harder to think that the card could even be Constructed playable.”

Internal timelines meant that it was too late to make changes to the token-making enchantment. To fight the menace of First Response (which, remember, saw zero play), Wizards looked to see what they could do in Khans of Tarkir. Here is the original Siege Rhino:

Siege Rhino

WBG

Creature-Beast

When CARDNAME enters the battlefield, each opponent loses 2 life and you gain 2 life.

If a spell or ability an opponent controls causes you to discard CARDNAME, put it onto the battlefield instead of into your graveyard.

3/4 

Stoddard explains that this card existed to fight Liliana of the Veil, which had originally been slotted for a reprint in M15. When LotV was removed, Rhino was given uncounterability instead. When First Response continued to dominate, they bumped Rhino up to four mana (and apparently increased its stats and drain by one as well), gave it trample, and weakened what had been the four-mana Anafenza down to three mana.

This is how R&D created an (at the time) unpopular card that dominated Standard for some time. Players looked at Siege Rhino and thought, “Whoa, how could they have printed this?! It's so powerful!” They had no idea that Siege Rhino's development was tied up with:

    

None of these cards exist in a vacuum; metagames, by definition, are formed by how every card interacts with every other card. The Siege Rhino article was published in 2015, but five years later, it's hard to imagine that it doesn't still apply.

From the article, it sounds like Development at the time consisted of one group of people who constantly built decks and developed cards, and then went back and iterated to improve them. From the few articles that Play Design has printed, it sounds like the system works much the same. This system of development is inherently flawed, as it leads to inbred metagames that are working off old assumptions that are no longer relevant. For even the best developers in the world, it's impossible to throw off the biases and preconceptions that have been built up by working with cards like First Response that have been dominant for their entire lifespans.

So what would be a better way to develop cards? The primary change that I would make would be to split Play Design into separate, independent teams that see the cards sequentially with no input from the prior team.

Back in 2017, Mark Rosewater wrote that

“Play Design has four three-month segments, one for each play environment (with four Standard-legal sets released each year, there are four distinct play environments).”

This means that Play Design had three months to playtest and refine Zendikar Rising Standard. Personally, I suspect that this is not enough time, especially when the cardset is still changing. But Wizards releases products on a tight schedule, so timelines are challenging. Here's what my model would look like:

Weeks 1-3: Group 1 of Play Design takes the cards handed to them by Set Design and works to build the best format they can. They will iterate on the cards they received to balance the format. In all likelihood, they do not finish. When time is up, they hand the updated, modified cardset to Group 2 with no comments.

Weeks 4-6: Group 2 picks up where Group 1 left off and works to build the best format they can using Group 1's modified cardset. They, too, modify and iterate and attempt to create a fun, balanced format. When time is up, they hand the cardset to Group 3 with no comments.

Weeks 7-9: Group 3 does the same thing as Group 2, but with Group 2's cardset instead of Group 1's.

Weeks 10-12: All three groups reconvene, compare notes, and collaborate to finalize the set.

Is this time structure perfect? Almost certainly not; other than what they publish, I have no insight into how Wizards builds timelines, etc internally, and I'm sure this could be tweaked. The key principle, however, is successive review by Play Designers with no biases or preconceptions about the cards and metagame. The downside of this process is that there will be time lost at the beginning of each cycle as each group familiarizes itself with the cardset and works to build up a metagame from nothing. Yet this is also the strength; the current model makes it impossible for Play Designers to start over from zero because they have worked on the cards and iterated them for months, thus leading Play Design to miss things because they're too close to the project. 

In many fields, this is common practice. Authors give their work to beta readers, gamma readers, and editors because, among other reasons, they reach a point where they are so close to the work that they are entirely incapable of finding certain problems on their own. Software engineers do code reviews. Many people even do this in sending emails! They have a friend or colleague read something before hitting “send” because a fresh eye can catch things that an entrenched eye cannot.

Currently, Play Design has nine employees. Under my model, the groups would each have three people in them, which does not seem nearly large enough to build a format. Here, Wizards needs to do a cost-benefit analysis and determine how much value a good competitive metagame generates to balance hiring additional Play Designers (five per group seems like the bare minimum). 

With the dawn of MTG Arena, more non-competitive players are entering the competitive space than ever before. It is imperative for player engagement and retention that competitive metagames be diverse, fun, and balanced.

Ryan Normandin is a grinder from Boston who has lost at the Pro Tour, in GP & SCG Top 8's, and to 7-year-olds at FNM. Despite being described as "not funny" by his best friend and "the worst Magic player ever" by Twitch chat, he cheerfully decided to blend his lack of talents together to write funny articles about Magic.