ark Rosewater recently revisited a decade's old column entitled "When Cards Go Bad." The development team that I manage and work alongside bears the lion's share of responsibility for the power level of cards and so we greatly influence one criterion by which many players measure whether a card is good or bad.
Thundermaw Hellkite | Art by Svetlin Velinov
As we roll around into the holiday seasons full of giving, I'd like to turn Mark's original article upside down and talk about the good. The glass half full. Be warned, I'll try to play around and within Mark's initial article structure and see where we end up. Along the way, I'll hit on some of our more contentious cards, like Pack Rat and Thragtusk, which fluster many players who find them to be too strong. The topic as hinted at in the title might seem a bit silly, but let's look at why we make good cards, since after all sometimes "good" cards go "too good."
Let's start off with a subsection title that should be a hint as to some of the levity to follow:
1) All the Cards Cannot be Bad
Mark highlights early in his initial article that everything is relative. He walks through a scenario where, if players selected the game's best 1,500 cards for a given Constructed format, only a few hundred cards would actually see much play. There will always be cards that are better, or have better synergies, than others.
It would be a fairly silly enterprise for us to be spending a lot of effort trying to make bad cards. We are much more interested in making appealing cards. Early in development of each set, we make a list of all the cards we think we'd like to be good. We know that only a certain percentage of cards are likely to create an impact on Standard, so we hone in on cards we've had fun with in playtests or cards that look like they'd be fun if they were a "thing" in Standard. We also tend to favor novel effects. We listen to the feedback and numerical results from our rare polls. Fortunately, there are a lot of things we deem fun, and we try to ensure that it isn't always the same subset of effects that are getting more love than others.
We monitor these cards most closely and make sure they are in fact leading to fun games and fun environments. We cut cards, weaken cards, move other cards up into the "good" section, and iterate.
There is a certain expectation for where the spectrum of power level will fall for cards. This is largely dictated by our experienced players, who are in touch with the power range of cards from Magic's history, both in recent memory and from its beginnings.
You could imagine a world in which the power level of cards in Magic was much different. Picture this alternative universe in which Divination costs . Maybe you'd play it if the All-Stars were Squire and Chimney Imp. We could likely balance things in this world but there's a decent chance this world wouldn't be as much fun.
It is our responsibility to find a sweet spot for the overall range of power level of cards that make up Standard. There is also generally a power level for any given game engine that is going to be best for a game's target demographic and set of effects. We realize some players appreciate a speed and power outside of what is offered in Standard or a typical new draft set, and we encourage these as a breath of fresh air to what we primarily focus on. Modern, Legacy, and Cube are meant as good alternatives.
For that matter, you should experiment with your own creations of formats. I find it amusing that R&D has recently been toying around with a format that probably sheds insights on this very subject matter. Several members of R&D have spent many hours exploring a format they call "Gatherer-Terrible," where they can only use cards rated by the community on the Gatherer card database as 2 or less out of 5. I haven't explored the format, so I can only speculate what reigns supreme here.
We make good cards because doing powerful things can be fun.
2) Different Cards Appeal to Different Players
In this section of Mark's initial article, he explains how we have many different types of players and many different formats, both Limited and Constructed. This holds as true as ever. Players keep inventing and popularizing new formats. And we do the same. A regular pass we make during the creation of files for a set is to make sure there are enough cards that will excite each of our various player psychographics. But it goes beyond that. Hopefully, each card in a set is serving a particular role. It is aimed at a player, at a format, to help balance colors in Limited, to make sure there's enough artifact removal in Limited, to make sure the card's not so good that it will flow to the player we want it to go to in a draft, etc. That's not to say all of the cards are powerful.
While "powerful" is likely to correlate with "good," I'll note that "good" can mean a great many thing to various players. What is your favorite card that is a weak card? Have you ever thought in those terms? While a Johnny can enjoy a powerful combo card, trying to make a marginally powerful card be more than just that is the type of puzzle that often makes a card good to him or her, perhaps even more so than something clearly "good" to most.
If you are more concerned with power level, don't worry. Much of the discussion in this article will focus on that. We are making many good cards, although some of them may not be for you. Your personal mileage may vary. I'll also note it is challenging to get it exactly right for each of these formats and players.
We make good cards because we have many formats and types of players to appeal to.
3) Diversity of Card Powers is Key to Discovery
Mark's section here speaks of the importance of learning how to evaluate cards and the process of discovery more generally in Magic. This relates in many ways to the first section above. Mark focuses on the early parts of the learning curve and how bad cards can tie into that. We want this process of discovery to progress to the endgame. We have many players who devote countless hours toward building decks, reading strategy articles, and exploring every facet of Magic.
Every so often, we want to challenge your notions of what you've learned to be good and bad. We also want to change what play patterns are best. For a year in Magic, you may find, for example, that Standard is all about ramping into huge monsters. In the years following the dominance of a ramp deck, we are likely to take precautions so ramp isn't so good. Our Limited environments will also change from time to time over how you are supposed to prioritize various effects. The relative power level of creatures to spells has also been shifting, for example.
Primeval Titan | Art by Aleksi Briclot
It can be very frustrating to find that the rules you've set up in your mind about what is good and bad no longer apply. And we've see this frustration when we shift some of the power in a format too much.
Nevertheless, while we want certain things you learn to be true and stable, there are a great many things we want to shift from time to time. Many people are playing Magic because they like a good challenge and we want to keep challenging even our most advanced players.
Who knows, the next Ivory Cup variant we make could be the next big thing in a format coming to a game store near you. And the players who first recognize and embrace this shift away from something they've been trained to think of as bad will the first to thrive. It's on us, though, to make sure we shift to new powerful things that are still fun. Ultimately, this means we make plenty of good cards that we think will feel new and be hard to evaluate via comparison with past cards.
We make good cards to challenge our players to adapt.
4) Power Levels are Relative to the Metagame
In this section, Mark had talked about how cards can shine depending on the other cards in the card pool for a given card to combo with. Some cards certainly shine more than others from time to time. His focus was on cards that looked bad or hard to use.
Looking instead at cards that are meant to be good in the first place, we also still see dramatic shifts in power within a format. A good current example of a card varying from good to better in today's world is Delver of Secrets. It was a huge player in pre-Return to Ravnica Standard and barely a force at all in today's Standard. Meanwhile, Delver is very much a force in Legacy but not so much in Modern. Its power has been very context-dependent. Maybe it will rise again in Standard soon?
Meanwhile, we have other cards, like Falkenrath Aristocrat, that we thought would be an immediate hit. Other than making brief appearances in combo decks at Pro Tour Avacyn Restored, this Vampire was largely in slumber for a year and has only very recently begun to dig its teeth into Standard within recent Rakdos Zombie decks.
Finally, let's take a look at Thragtusk in this section, since we are somewhat focused on metagames and available card pools here. Thragtusk was very much meant to be a thorn in the side of Delver decks. We were looking to make a good card that would be a tool against that deck. Generally speaking, we know cards targeted like this will need to hold their own against other decks and be broadly appealing and strong; they can't simply do work against the targeted deck. Thragtusk does a lot. It does yet more with cards like Restoration Angel, which was even more abundant than we'd expected. There is always a danger with a card like Thragtusk that other decks will get caught in the crossfire. Fortunately, aggro decks are thus far more than still holding their own, finding ways to go big, finishing the opponent that turn, and/or keeping Thragtusk from blocking. Nevertheless, we have upcoming checks and balances for Thragtusk and so the cycle begins anew. It's always challenging since cards that begin to take over formats are themselves elusive to obvious ways of tackling them.
We make good cards to push the metagame along, to create new archetypes, and to suppress other archetypes.
5) Diversity of Power Rewards the More Skilled Player
Magic is going to reward players who can identify which cards are good and which cards are bad. Mark speaks of a world in which we could hypothetically make all cards of the same power level. In practice, this goal isn't remotely feasible. Nor would it be particularly desirable. There will always be better and worse cards and strategies, and it's all a matter of degrees and how long it takes players to put the puzzle pieces together.
Just how skill-testing is to identify good cards? I'll give a recent Return to Ravnica example. We have a lot of data about our Magic Online Prereleases, and thanks to Lee Sharpe, we gain an analysis of much of that data on a regular basis.
It's worth noting that our players have become a very discerning audience. Prereleases are the first time players gain experience with cards. Players have learned over the years, though, what typically makes for a good card and what doesn't. Take, for example, players playing Golgari in these events. What is just about the worst card I could think of playing in those colors? Urban Burgeoning. What percentage of Golgari players included at least one copy of this card in their deck if it was in their card pool? 4.4%. It was only "outdone" by Catacomb Slug, ringing in at 3.8% of Golgari decks that had the option to play it. The Slug shows a clear discernment of player expectations for desired stats along a creature curve.
Let's look at the other end of the spectrum among Golgari decks. Nearly all players, 99.0%, played a Vraska the Unseen if they opened one. Similarly, 97.7% played Overgrown Tomb and 96.8% played Jarad, Golgari Lich Lord. In all cases, these percentages are based on the player's choice of guild, regardless of what kind of deck the player actually built.
So where am I going with this? Well, there is a great spectrum of perceived power levels. Some cards don't as easily fall into categories of meeting expectations. For example, a big awesome monster or creature removal. There have been many players suggesting we ban Pack Rat in Limited.
What percentage of Golgari players receiving Pack Rat in their pool chose to play it? Better yet, take a guess at the order of cards in the set, from most-played to least-played, as a percentage of opening those cards in a card pool among Golgari decks. Your choices:
when you are ready to see the results, from most played to least:
The purpose of this section thus far is more to highlight the skill-testing nature of Magic rather than to absolve us of what we've created with Pack Rat. We hear loud and clear the frustration of our players. There is a very brief window to defeat this card, and only a handful of cards are likely to trump it either early or late in the game. It can also win the game regardless of what other cards its player might have drawn.
What can I say in our defense? While we want cards that will reward the more skilled player, both in terms of recognizing raw power and in recognizing what will synergize with those cards, we also want to make cards that are obviously powerful and appealing. We want to make cards that newer players to a format, Limited or otherwise, can figure out how to take advantage of and win with. And then keep coming back for more. An early Pack Rat does the winning part in spades.
However, based on the percentages above, it's arguable that in this case we didn't make it quite obvious enough for the players we actually most wanted to help. Pack Rat was a deliberate choice during later development to push the power of black in the set and to do so in a way that wasn't just another typical enormous monster or sweeper spell.
And yet, players selecting any guild choosing to play a Pack Rat in their decks at the Prerelease only won 55.4% of their first-round matches. It's a noticeable, and I can only assume significant, bump in performance as with any number of other bombs, but we feel this is well within the range of tolerance for imbalance. Yes, if we were to look only at the percentage of games where it was drawn and played early, it might be a very high percentage. Clearly, this is too high a rate for the tastes of many. But Pack Rat isn't always drawn early—or drawn at all.
Since we want to make cards to help newer players win more often now and then, we'd be curious how you might want them to win. In what fashion would you like to take your defeat? Just be careful your answer doesn't ultimately simply boil down to being outplayed or out-skilled or seeing that it happens less regularly, since that is at odds with the goal of the small percentage of cards we'd like to be doing this.
We make good cards both so the most skilled player will win, but we also make good cards so the most skilled player won't win too much.
6) People Like Being Given Gems
Mark's section discusses finding hidden gems and how rewarding that can be. He's right. The dialog created by hard-to-evaluate cards is great. It fuels community discussion. There's spirited debate each preview season as to whether a card is good or bad. Cards are seldom just okay in these discussions. It's fun to try to prove your case with new wacky decks or combos and to make predictions about power level based on what you've learned about Magic since the last set.
You know what else is also rewarding, though? Not having to mine for those gems. We certainly want many cases where things are as you expect, where you don't find out your new, exciting sparkly thing is a dud. There is something to be said for opening up a pack and thinking, "Yes, I got there! Look at that name and that art, and oh, that power!"
Sounds good, right?
While we want hidden gems, it's also important that thematically what our players want to be good is good. Before I worked here at Wizards, I'd led development for years on a TCG about superheroes. I assure you, it wasn't satisfying when responding to the question, "Who's the best character card in your game? Is it Superman, Batman, Wolverine, or The Hulk?" with the timid answer, "Well, possibly Ahmed Samsarra or maybe Dr. Light."
What should be awesome in our game? Dragons, Angels, Sphinxes, Demons, Vampires, Hydras, Planeswalkers, and legendary creatures, to name but some. You tell me what else. It gets trickier with spells, although I'm a personally a sucker for X spells and cards with awesome names and cool, straightforward effects. It would be awesome if such cards were among our best cards, at least more often than they have been at times in the past. We are spending more energy making this happen. For example, Zac Hill wholeheartedly poured much effort into making this happen with Thundermaw Hellkite and Sublime Archangel.
We make good cards to excite people about our current plane and its stories.
7) R&D is Only Human
We are still as human as ever. We have a collective knowledge passed along for almost twenty years on what is working and what is not. Yet, we make mistakes. We want many cards to be good. Yet, we rarely want them to be better than they need to be. That will just make it harder to outdo ourselves in the next set. We're always looking for small knobs to turn down on potentially overly powerful cards. We don't always discover the knobs to tweak such that we still remain confident the good cards will see play.
We also keep making changes all throughout our process. Any time we think we can make a tweak or a big change to make our set better, we usually do so. This means our sets aren't stable through much of our testing. This also means we are more likely to miss things. The alternative, though, is that we don't address possible weaknesses with the set. It's a delicate balance.
Maybe someday Magic Online will help us in our mortal ways. Its sentience appears a way off, though. I can only promise you that we are passionate about what we are doing. And we are listening to your feedback.
We sometimes make good cards inadvertently because we are imperfect.
We also make good cards because we like good cards, too!
Thanks for reading,