Fixing Magic

Magic: the Gathering is a 20 year old behemoth. Each year, it seems to get more popular and make more money. Wizards of the Coast, the company that makes the game, must be doing something right, right?

Magic: Pretty popular. Image from Modern Myths.

Magic: Pretty popular. Image from Modern Myths.

It turns out they’re doing a lot right. But that doesn’t mean they’re doing everything right.

Magic is an extremely fun game, but it’s not perfect. A large part of its success comes from its position as the first collectable card game, a much loved and very profitable genre. That innovation made players willing to overlook some of Magic‘s less than ideal qualities.

Today, I’m going to take a look at one of Magic‘s biggest flaws and explain how I would try to fix it. In doing so, I’ll show how to approach challenging design problems and encourage you to take a critical, honest look at everything, even games you love.

Before I get into the details, I’ll cut the suspense: the problem I’m going to address is the land system, one of the fundamental pacing mechanisms of the game.

Shields Up

Magic is extremely popular, which is one of the reasons I’ve chosen to critique it. Almost all gamers know the basic rules, and many designers were inspired to start tinkering around with games because of Magic, so most readers should be familiar when I talk about it.

That said, many will get extremely defensive when I suggest that something they love so much could be improved, so I’m going to take a minute explaining myself.

First, this discussion will not change Magic. The land system is truly fundamental to the game, and while Wizards does occasionally update the rules of the game, they tend to only change fringe rules. Even then, they do so in the face of major public outcry. Changing core rules like the land system would drastically change the game, possibly requiring bans on hundreds of cards from Magic‘s history. Magic may not be perfect, but it’s good enough without these changes rocking the boat.

In fact, I know Wizards won’t address this problem because Wizards employees talk about the problem as a feature rather than a bug. They say the added variance is good for the game. While I agree that variance is important for Magic, I think these people, all very intelligent, are just towing the company line. Addressing the problem would be too much of a shock to the game.

Second, these suggestions are not tested. My goal with this post is to show how to identify a design problem, carefully analyze it, and brainstorm possible solutions to it. If I were really trying to fix Magic I would build prototypes to playtest.

Because the ideas are untested, some of them probably suck. That’s ok. During the brainstorming phase of design, you’ll come up with bad ideas. Never dismiss seemingly bad ideas too quickly. You never know when one that looks bad will test well, or when a dumb idea will inspire a good one.

Finally, if any of the solutions were adapted, Magic would be radically transformed. That might mean your favorite deck wouldn’t work any longer, which might upset you. Sorry about that. Every change has winners and losers, but that doesn’t mean that every change is bad. Sometimes, sacrifices must be made for the greater good.

The Problem

The game of Magic is huge, involving tens of thousands of cards and an ever shifting metagame. But each game of Magic is quite short, usually 10 – 20 minutes. That said, some games are even shorter; these games are more frustrating than fun.

They may be pretty, but get too many or too few lands and you'll have a disappointing "game". Image from Collector's Cache.

They may be pretty, but get too many or too few lands and you’ll have a disappointing “game”. Image from Collector’s Cache.

In Magic, there are two types of cards. One is spells. These cards actually do something, getting you closer to winning or hindering your opponent. But they can’t do it alone–you can’t play your spells without the other type of card, lands. Lands give you mana, which power your spells. But without spells, that mana is useless.

Neither lands nor spells do anything on their own. You need a mix of both for any hope of winning. Unfortunately, it’s not uncommon for your starting hand to be lacking lands or lacking spells. When your starting hand is all spells or all lands, you’ve pretty much already lost before the game even starts.

Magic does allow mulligans, so hope isn’t completely lost after you draw your starting hand. But the penalty is harsh: whenever you mulligan, you redraw one fewer card, and cards are one of the basic resources of the game. Hands start at seven cards, so mulliganing twice puts you at about two thirds the starting power of your opponent. Even with a better ratio of lands to spells, you’re still at a huge disadvantage.

My suggestions today will address this problem. No game should be decided before it begins; those games are not fun. But before getting into solutions, let’s look at what the land system does for the game.

Spoils of the Land

The Magic land system may have its problems, but it’s also fundamental to how the game works. Whatever solutions we propose, we want to make sure to maintain the following contributions.

The land system makes expensive, powerful cards slow and risky. Image from Wizards of the Coast.

The land system makes expensive, powerful cards slow and risky. Image from Wizards of the Coast.

Pacing. The land system sets the pacing of the game. Each turn, you can play one and only one land. Since each land contributes one mana a turn, and each spell requires a certain amount of mana, cheaper, weaker spells can be played earlier in the game while expensive, powerful spells can only be played later. Uncertainty about how many lands you draw also means you might not be able to play a land every turn, so you won’t know when you’ll be able to play expensive cards.

Since most actions (playing cards and sometimes using cards you’ve already played) take mana, lands also determine how many actions you can make a turn. While other games, like Dominion, make actions an explicit resource, actions in Magic are implicitly determined by land.

Consistency. At Magic‘s core is the color system. Each spell requires mana of one or more specific colors to play, and most lands produce mana of only a single color. Different colors have different strengths and weaknesses, so including cards of different colors in your deck tends to make it more powerful. However, it also makes your mana less consistent, since only a fraction of your lands produce each color. In this way, players have to choose between a consistent but weaker single color deck or an inconsistent but more powerful multicolor deck.

Suggested Solutions

Lands set the pace of the game and help balance power and consistency–these are good things we’ll try to maintain. The problem is that lands do nothing on their own, and other cards do nothing without them, so hands with too few or too many lead to non-games. Here are some ideas that might keep the good stuff but eliminate the bad.

Less Punishing Mulligans. Simplest is to change the mulligan rules. Instead of losing a card each mulligan, perhaps a player would only lose a card every other mulligan (7, 6, 6, 5, 5… rather than 7, 6, 5, 4…). Or maybe instead of losing his or her whole hand, a player could discard any number of cards and redraw one fewer (for example, keep 2 cards and discard 5, but only redraw 4). Or allow players to use another resource (perhaps life?) to mulligan, rather than only losing cards.

In Game Mulligans. Another simple solution is letting players dig through their decks during the game. For example, if a player can’t play a land for the turn, let her draw an extra card (or draw and discard a card). Or at any time, let a player discard two cards to draw a card.

Less Variance in Starting Hands. What if a player constructs a special small deck, different from the usual deck, from which she draws her starting hand? There is still the potential for bad hands, but the variance would be much less.

Separate Land and Spell Decks. A friend suggested this solution to me. Players would have different decks for their lands and their spells and would somehow get a choice between the two when drawing. Alternatively, have a main deck with two smaller land and spell decks, and have some sort of penalty for drawing from the specialized decks. Maybe a player would have to pay life, or maybe there would be a delay in using a card from a specialized deck. Either way, players would be able to get more of what they need.

Multiple Card Uses. What if you could use spells to produce mana? Each turn, instead of playing a land, a player could place a spell from her hand in front of her. Whenever she wants, she could discard the card to get an extra mana. Only being able to use the card once is a real cost, but it’s better than never being able to use it.

Next Steps

If I were actually trying to fix Magic (or fix a game inspired by Magic), the next step would be to test these solutions. I would test the options that could be prototyped most easily first, starting from the top of the list and moving down.

Unfortunately, evaluating the different alternatives would be challenging because it’s unclear how much variance you actually want. Variance is crucial to Magic: it gives weaker players hope, it punishes greed, it adds suspense, and it keeps each game unique. But how much is too much? I believe that the current system has too much variance, but I’m not sure how much is ideal.

It’s possible that after testing all of the solutions, you’d decide that the current system is the best choice, as imperfect as it is. Either way, it’s worth exploring to find out.

Improving Perfection

Magic is one of the most successful and beloved American games. 20 years after its initial release, it’s stronger than ever. But that doesn’t mean there isn’t room for improvement.

If Magic could be better by examining part of its fundamental infrastructure, what assumptions are you making about your game that are holding it back?

Previous Post
Leave a comment


  1. Why not give to each card in the deck the ability to be played as a land of it’s own color. Simple (too simple ?)

  2. Adam Wardell

     /  November 11, 2013

    Nice article, Teale! The “Shields Up” section really messed with the flow of it though, in my opinion. It felt a bit lengthy and verbose, and while I realize it was meant to be preemptive, it alluded so much to the changes that you hadn’t yet suggested in the article that it felt very out of place. I think it would better serve if you’d had it at the very end.

    Sorry my comment is mostly criticizing your writing! I agree with what you have to say and it’s good food for thought. I’ve actually played a variant of Magic that is as Rémi suggests, with spells doubling as lands. It was quite fun, but it had a number of negative ramifications.

  3. Brian Jones

     /  November 11, 2013

    Interesting take on a semi-flawed game mechanic!

    I’ve been playing Blizzard’s Hearthstone CCG lately. They automatically increase the land-equivalent resource by 1 each turn up to a total of 10. I really like this variation because it makes it easier to plan around your resources, and you don’t get caught with your pants down with the too much/too little problem.

    I like the suggestion in the “Separate Land and Spell Decks” clause as well. It seems to me the main point of these types of games is the spells, not the resource management. Making it a choice between the two seems a valid strategy, especially in the case of Magic. It gives multi-color decks a better chance at being deployed easier, although single-color decks may have a slightly better advantage now that land flow can be consistent.

    Clearly a lot of ways to go about it, but I do agree that Magic’s system is somewhat flawed by being a combined deck resource.

  4. Sean McCarthy

     /  November 11, 2013

    Hey, I realize this is just a thought exercise and that you’re not actually advocating these changes. Nevertheless I found the content of your analysis and suggestions to be extremely disappointing – the kind of stuff I’d expect from random people on a Magic forum.

    Let’s talk about “The Problem”. You say that the problem is that sometimes, games are lost before they start because of what cards you get. That is a simple fact of any game with randomness, if you consider the randomization to have occurred at the beginning of the game. This point of view feels natural because in practice, you shuffle the deck at the start of the game and not thereafter. But it is equivalent to, say, a game of backgammon. Would you say that backgammon has a problem because it’s possible for there to be no way to win the game against reasonable play given the dice rolls that will occur throughout the game?

    I don’t think it is a problem for a game that sometimes luck is bad enough that someone playing perfectly can lose against someone playing merely reasonably. In fact I consider that a point in favor of the game – it’s important for the outcome of the game to not be in doubt simply based on the relative skill levels of the players.

    What is a problem is that sometimes you don’t get to make any real choices. For example, if your deck ends up being shuffled such that all your lands are on top every time, you get to make a few choices at the start (how many times to mulligan) but after that you’re just drawing and playing a land every turn. (OK, you do also have the option to not play a land as a bluff at some point, but it will feel pretty hollow!) Similarly, if you get so few lands that you can’t play any spells, your only choices are going to be what card to discard every turn, which doesn’t make for a fun experience either. So IMO, the problem is not that games are decided before they start, but that you don’t get to DO anything along the way. Of course, a saving grace here is that these games tend to be really short and you can always concede to move things along. If you play Magic, a very small fraction of your time is spent on these lame games.

    On to solutions.

    Most of your proposals show a lack of understanding of the game. Let me give a simple example and you can extrapolate to your proposals. Suppose we add a rule that every game, you get three free mulligans before your mulligans start reducing your hand size. How will this affect the game?

    The answer is threefold:
    1) Combo decks will become a lot more powerful, because they have extra built-in chances to find their combo. But they will become powerful only in a way that is based on luck at the beginning of the game. This will increase the number of games that are decided before they begin.
    2) Other decks will slightly reduce the number of lands they play with, increasing their power and variance. We lowered their variance by giving them extra mulligans, and of course they will decide to spend it on increased power.
    3) Games will take significantly longer to play, especially tournament games where players actually shuffle enough to randomize their decks.

    One other very common suggestion that people make (and that other TCGs have used) is that you can play any card face down as a land. This may just be personal preference, but I find that solution to be quite boring. The result tends to be that you plan to build your deck to play X lands, 1 every turn, and then stop. Therefore the sequence of card costs of the cards you end of playing goes e.g. 1, 2, 3, 4, 5, 5, 5, 5, 5…. I find Magic’s system where each successive land drop is more likely to be delayed 1+ turns to be more exciting and to result in more interesting strategies.

    There’s an obvious and easy solution to the problem which doesn’t even require a rules change, and that is to print more strong cards which can serve multiple purposes; i.e. either as mana, or as cheap cards, or as expensive cards. Indeed, there are a ton of mechanics in the game which allow this. Seriously a ton. There’s nothing preventing Wizards from increasing the number of these that they print and therefore reducing the amount of mana screw variance in the game, and they know this. There is no conspiracy where they are “toeing the line” when they claim that they think this level of variance is good – it is not at all hard for them to “address the problem”. They can change things simply by printing new cards, which is what they do anyway. They actually do think the current amount of variance is good. (Well, as far as I know – I haven’t played Magic or followed it much for the last few years.) Personally I think it could stand to be a little bit lower, but I realize that the vast majority of their audience is more casual than me, and I suspect they’ve put a lot of thought and research into this topic over the years.

  5. Hi Sean. Thanks for your well thought out and lengthy comment. Forgive me for not being able to address everything you say.

    Based on your response, I believe I may have explained the problem poorly. I’m not claiming that the outcome of games with randomness should not be determined before the game starts, but I do claim players should never feel like the outcome is determined before the game starts. Game design is about psychology and the emotions of your players, not about math (though math is a very useful tool). The problem with Magic is that games feel like they are hopeless before they begin, not that the decks are shuffled before the game starts or that optimal play will sometimes lose to suboptimal play.

    I believe your analysis is making two common mistakes that less casual players make when it comes to game design.

    First, you are presuming to know how changes will impact player experience and player behavior. You might be right (in fact, I discussed the changes likely requiring bans and affecting how good certain decks are). That said, you could just as easily be wrong. Playtesting is extremely important because it shows how a game will actually work. Letting yourself think you know how something will turn out is a great way to rule out potentially legitimate solutions before giving them a chance.

    Second, you are letting particular implementations of solutions limit your imagination (including, apparently, a solution I never even proposed). I mentioned that some of the solutions I proposed might suck, but it’s still valuable to consider them because they can be jumping off points for better ideas. Taking a list of solutions as complete, especially when you have preconceptions on how the solutions will work without testing them, is a great way to limit yourself.

    If you’re serious about game design, I encourage you to work on opening your mind about these two things. You’ll find you have a lot more possible solutions available, which will tend to result in better games.

  6. Martin JT

     /  November 12, 2013

    How about a guaranteed minimum number of lands, spells in your starting hand, with luck-of-the-draw after that?
    1) Shuffle lands and spells into separate decks at the start
    2) Deal yourself x cards from the lands deck, and y cards from spells
    3) Shuffle the two decks back together
    4) If x+y < 7, deal more random cards to fill up the hand
    5) Play with the all-cards deck from then on

    I am not nearly experienced enough with MTG to know what x and y should be for this change. I'm also not sure if the problem you describe extends into the later game with unlucky draws, but in that case the player is probably more likely to feel like they could have done something different to prepare for this bad luck.

    Fun post, Teale!

  7. Krynn

     /  December 12, 2013

    I am glad i’m not the only one that is fed up with Wizards inability to fix their game. the only reason it hasn’t been fixed yet is they are afraid to lose money/momentum for their new cards…

    A conclusion that ive drawn in the past year or so is that part of the reason Magic is so popular is its another form of Gambling, and therefore can be truly addictive… which might be why Wizards is never going to change the Mana system.

    I actually like your solution where you could play colored cards as lands! that would be fun to test. and i believe it would lead to much more evenly matched and involved games… i can’t even express how frustrating it is to take your time, and money to a event and get mana hosed all day… its actually what has made me stop playing competitively… i also think that the “bug” causes more cheating…

    I would like to run another one of my ideas that i’ve played with a few times across you and see what you think…

    My other problem with the game is the escalation of rare cards.

    Thats why i suggest a Points system for building a deck.

    Basically, set a point limit for the Event. Say… 50 points

    Mythics are worth 5 Points
    Rares are worth 3 Points
    Uncommons are worth 1 Point
    Commons are worth 0 Points

    Its my feeling that this system would lead to new strategies and more even games.

    Let me know what you think! Thanks! Krynn

    • Hi Krynn,

      Your idea sounds very interesting! (Though, sadly, again against Wizards’ business model.) I’ve often thought that limiting all cards to 4 in a deck feels arbitrary. In the past, I thought about having each card have a maximum number specified, so rarer cards could have a lower limit, but your idea seems like it could work too. I feel like Android: Netrunner does something like this… does anyone know the specifics of how that works?

  1. Today in Board Games – Issue #85

Leave a Reply

Your email address will not be published. Required fields are marked *