Gaming According to Garfield

When Dr. Richard Garfield was announced as being a core member of the Artifact team a lot of people were really excited. The father of all trading card games was going to be joining up with one of the most celebrated game studios to work on a flashy new project! Personally, I was intrigued to see how things would shake out, given the stories I have heard of Dr. Garfield’s brilliance in game design. Surely his unique perspective on gaming would be instrumental in making Artifact into the title that we all hope it can be.

But as I was thinking about this, a question dawned on me. What is Richard Garfield’s philosophy on gaming and game design? I have bumped into some of his talks and writing a couple of times, but I don’t think I had ever sat down and studied his thinking. If Dr. Garfield is one of the main developers behind Artifact, then it might be useful to take time to figure out how he thinks, right? He is also different than some other game designers since he has given a number of talks on abstract topics related to game design. This makes it is a lot easier to peek into his thinking compared to someone like IceFrog.

My goal in this piece is to summarize Garfield’s thinking about gaming and design. Some major topics covered include his take on balance, randomness, “metagames”, as well as in-game economies. I am going to be writing about all of these topics in the context of “what does this mean for Artifact”. Sure, Dr. Garfield’s thoughts on balance and RNG are interesting, but how might these ideas actually play out? Before we get into Dr. Garfield’ thinking specifically, let’s discuss a broader question: why should gamers care about game design?


Why Game Design Matters

The average moviegoer, book reader, and music listener knows essentially nothing about what makes a movie, book, or album “good”. Most people know virtually nothing about screenwriting, literary theory, or music production. While this point might sounds elitist, I don’t see anything wrong with enjoying media in this way. The secret to great entertainment is that you don’t need to know why it is entertaining to enjoy it.  That being said, many people get enhanced enjoyment out of analyzing content at a deeper level. Personally, I love movies, but I don’t tend to watch them unless I have someone to talk about them with. I get most of my entertainment from the opportunity to cross-examine the characters, acting and plot progression of the films I watch. While someone else might not appreciate a movie the same way as me, both are just paths to the same goal of enjoyment and escapism.

While I think this is true of almost every form of media, gaming is different. The overarching purpose of gaming is still escapism and enjoyment but there is a more immediate goal: winning. The better you understand how a game works, the rules interact, and how balance is achieved the better you can exploit these to improve your chance of winning or advancing.

I want to be clear here: there is nothing wrong with being a casual. If you just enjoy firing rocket launchers across the map to see what happens, or stick with the one strategy you know and love, go nuts! There are lots of ways to enjoy games! That being said, if your goal is not simply to enjoy a game, but master it, this casual approach is not going to cut it. Understanding why games work the way they do can help improve your chance of winning.

This push to understand games has taken on different forms in different communities. I’m sure I don’t need to tell you about the tens of thousands of articles and videos that teach players about every facet of every new release. I love this kind of content, even for games that I don’t play! While this drive to improve exists in every competitive game, I feel like card gamers take on a slightly different approach to what I have seen in other communities. While many people are interested in understanding what makes a card or deck powerful, this often leads to more abstract discussion about balance, theory and game design. There are many complex reasons why card gamers are inclined to have a more analytical view of gaming, such as the structure of card games and the history/culture of Magic the Gathering. Digging into why card games are so fertile for this theorizing is beyond the scope of this article, but I think this genre has a structure that inspires players to think like game designers. This thinking encourages you to approach the game from many different direction, pushing you to think about the way resource systems work, how different strategies fundamentally interact, and the role chance plays in the outcome of games. Lessons learned through this kind of analysis can then be directed to improved deckbuilding, strategizing, and technical skill.

Hopefully this gives a good “big picture” as to why game design matters for gamers generally, and especially why it matters for card game players. Understanding how a game works can help players analytically dissect games in a manner that improves their chance of winning. With that out of the way, I’d like to talk about Dr. Garfield, but I have a couple things I’d like to “disclaim” first…


Before I move into the meat of the piece, I want to drop a few disclaimers.

  • I am in no way shape or form “officially” summarizing what Dr. Garfield thinks. I would absolutely love to interview him about game design, but I’m not exactly “high impact” enough to set up that kind of thing. This is not an “authorized” piece in any sense.
  • I have spent a couple days researching and writing this article, but Garfield has spent several decades thinking, speaking, and writing about gaming, so I am not going to capture everything. I am also trying to summarize the thoughts of someone else based on a few chunks of content I could pull together. Most of my sources are about 5 years old, and he could very easily change his mind in that time. (If you have more recent sources I would love to see them).
  • In many cases I am going to be writing about generalities. Games are a complex topic, and some of the best games are good because they break some rule or convention in interesting ways. If I say “Randomness is good to have in games” and you comment with “Would chess be better with a bunch of dice rolling?” I am just going to respond with Twitch emotes (I’m thinking SwiftRage or FailFish, but I haven’t decided). I am also going to be focused on what Garfield defines as “orthogames” - games played with 2 or more players with clear winners and losers, and a fixed end point.
  • These topics can’t be fully separated from one another, or from other subjects related to game design. I will do my best to stay on topic, but these topics bleed all over the place. I have done my best to focus only on the subjects that seem the most relevant for now.
  • I am not going to preface everything by “Dr. Garfield thinks…” or “I think Dr. Garfield thinks…” I will try to make it very clear if I am interjecting some of my own thoughts. For example, there are elements of each section that speculate on how Garfield’s philosophy might apply to Artifact. This is my own inference based on my study of his thinking.
  • My main sources are some talks you can find on Youtube (talk 1, talk 2) the textbook he wrote (written with Skaff Elias and George Gutshera), this ‘manifesto’ shared on his Facebook page and a couple written interviews (interview 1, interview 2). I do use some quotes and direct references in the text, but not everything is going to be sourced to the level you would expect in something like an academic journal. If you disagree with my understanding of Garfield I would ask that you find specific sources to back that up. I’m also totally willing to be proven wrong! Obviously if you just think Dr. Garfield is wrong, you don’t need any source for that.

With all that said, let’s get to it!

Balance, Collapse, and Rock-Paper-Scissors


We take it for granted that competitive gaming should be “balanced”, but we should take a moment to examine that basic assumption, and discuss why this is the case. First, what is meant by “balance”? Essentially, we want there to be a variety of viable strategies in a game. If a game is designed such that everyone is encouraged to always play the game the exact same way the game is going to get really boring really fast. Obviously there are some games with a fairly limited range of strategic possibilities that have been fairly successful, but they typically have a shorter lifespan and are less replayable compared to games with more diversity in play styles.

So, we want different kinds of strategies in our game, but it is not enough that different strategies are possible, they all need to be competitive. There may be some segment of a player-base that is willing to play strategies they know are “fun” or just “bad”, but most players prefer strategies that give them a good chance to win. It might sound easy to just offer a variety of approaches, but if one strategy is just better than everything else, you enter a state called “strategic collapse”, where multiple potential strategies are dominated by a more powerful strategy. If the “best” strategy has the best match ups across the board it will push every other possible strategy out of the game. This leads us back to square one.

The most common solution to this problem is using some variant of “Rock-Paper-Scissors balancing”. Everyone knows the principle of Rock-Paper-Scissors: A beats B and B beats C while C beats A. “RPS-like” constructs are actually quite common in competitive games if you look a bit, though it is not typically implemented in such a cut-and-dry fashion. Take for example card games, where aggro is typically beaten by midrange, midrange is beaten by control and aggro keeps control in check. There have been many exceptions to this, but it is a fairly reliable rule-of-thumb. Here’s another “RPS-like” example for RTS games: ranged units beat flying units, flying units beat melee units and melee units beat ranged units. You can probably think of a lot of examples in games you play. RPS-balancing can help stop strategic collapse since it is impossible for any one strategy to ever dominate all others as it will always have at least one natural predator.

“Wait a minute Neon! Rock-Paper-Scissors? Isn’t that a terrible game? Why should games be balanced like that?!” Good point! A lot of people have an adverse reaction to the idea of RPS-balancing, since Rock-Paper-Scissors is essentially just a game of chance. Well, there is no rule that “rock” needs to beat “scissors” 100% of the time in our game, or that our system should be limited to just 3 elements. It is possible to add a lot of complexity to the ecosystem if we want to. Let’s take two potential metagames for Artifact as an example. The first has just 4 possible decks: mono-red, mono-black, mono-green and mono-blue. Each of the arrows between the bubbles has a % associated with it, which represents the win rate of the deck at the start of the arrow over the deck at the end of the arrow. Black, red and green are in a simple RPS configuration, while blue is off to the side, and has a 50% match up against everything. This metagame looks extremely boring since so many of the matchups are 1-sided, with the exception of blue players. This is an example of bad RPS balancing.

Screen Shot 2018-04-18 at 10.30.07 AM.png

Now let’s take a more complex metagame. There are 6 decks rather than 4, and while there are a number of 65/35 match ups, this is the most lopsided you can see. All these decks technically have the same win rate against the field (assuming uniform distribution), but the metagame is going to feel more diverse given the range of decks, and this metagame could give rise to a lot more innovation and change as opposed to the more simple Rock-Paper-Scissors balance we see above.

So what does this mean for Artifact? I think we should expect some amount of RPS-balancing, and you should be prepared for this. Obviously in my examples I just used color identity to describe the various decks, but you can imagine these are defined more by hero composition or deck style. Hopefully this explanation is convincing that RPS-balancing is not inherently bad if implemented properly. This type of balance is actually quite healthy as long as the metagame is sufficiently diverse and decision-making within this meta is interesting.

Also, I should mention here that Artifact is likely not aiming for an infinitely diverse metagame. In one of Dr. Garfield’s talks he mentioned that most healthy metagames have about 4-8 popular decks. Obviously having fewer than this could easily lead to strategic collapse, imbalance, and/or staleness, but having an extremely complex meta has its own problems. This may include being difficult to navigate and a higher barrier to new players. Personally I would say that I agree with this, since an overly open meta is very difficult to “attack”, but at the same time I understand the appeal to metagames where anything is possible. There is also a lot that could be said about the barrier between popular decks and fringe decks, and how this is a very difficult line to draw. I have no idea where this equilibrium will ultimately fall, but you should not necessarily expect things to work as they do in DOTA where almost all heroes are viable.

In all of this I have written like decks just function autonomously, but that is obviously not the case. Let’s take a moment to discuss the interaction between skill and balance.

Skill and Balance

When discussing the win percentage of each deck we essentially spoke like the decks just played themselves. In reality, we know that this is not how things work, and that the player piloting the deck is often more important than the contents of the deck itself. This has interesting implications in balancing, as the balance of win rates between different strategies can actually be skill dependant. A match up that is “deck A” favored in the beginner experience level can become “deck B” favored in the expert metagame. In one of Dr. Garfield’s talks he describes the difficulty of this phenomenon, focusing on one specific game in which the starting character struggled in low experience brackets, but performed really well at higher experience levels. This lead to a groupthink where this free character was perceived as “bad”, and that the game was forcing you to buy the better characters to advance, when in reality you just needed to “git gud scrub!”

It is difficult for me to tell how intentional this style of balancing is from the outset of a game. Game designers can easily identify which strategies are easier or harder to play, but the exact question of how this will shake out in practice must be really very difficult (if not impossible) to predict. I do think, though, that we could easily see this “experience dependant power level” phenomenon in Artifact, since Valve uses this philosophy in DOTA 2. You can actually look at DOTA 2 database websites and see the win rates of different champions at different skill levels. Take for example Wraith King and Bounty Hunter (I’m using stats from April 2018 for anyone who is reading this in the future). Wraith King is a fairly tanky character with relatively few activated abilities, and whose ultimate automatically resurrects him when he dies. This makes him pretty easy to play, since your gameplan is pretty simple, and his ultimate doesn’t require strategic execution. As a result, Wraith King wins about 57% of his games at low experience play, but as the competition gets better he falls to about 50% at the highest skill levels. By contrast, Bounty Hunter is a character that uses stealth to set up surprise attacks on the enemy team, and benefits greatly from coordination from teammates. Bounty Hunter’s win rate starts at 45.8 in low experience games, and climbs all the way up to 53.4% in high rank games.


Going up or down almost 10% win rate may not sound like a lot, but that is a pretty big swing if you consider these heroes are part of a 5-player squad. This kind of data also reinforces the importance and impact of data collecting programs. It is very hard to get a clear picture of match-up data without these specialized data collection tools, given that subtleties like the experience level of the player is essentially for interpreting data. Valve has generally been a fairly open company for data aggregation programs, as I understand it. I hope there is already someone out there thinking about how to make a program like this, as it will be a great contribution to the community.


When talking about skill, it is impossible to escape talking about complexity, as the two are obviously linked. There are many people who think the complexity of a game is tied to the depth of the game, or that simplifying the complexity compromises gameplay. This debate has been particularly pronounced in the Magic community, surrounding the design philosophy of “New World Order” which has been advanced within Magic development. There was a time circa the Mirrodin-to-Lorwyn blocks where the complexity of commons in Magic had really jumped up dramatically. This proved a barrier to newer players, as ever card they encountered seemed to have multiple abilities, or complex interactions with other cards. In response to this, Magic’s development team implemented a design philosophy known as “New World Order” that helped control the complexity at lower rarities. Some members of the community had a negative reaction to this change, claiming that simplifying cards would “dumb down” the game.  

In this interview Dr. Garfield is asked about “New World Order”, to which he responds “To me that is Old World Order!” meaning he had always approached design this way when he was working on Magic full-time. He goes on to explain that he thinks Magic can be a lot of fun with very simple cards. Complex rares and mythics should be thought of as a “spice” that is mixed in with the simple “meat and potatoes” that makes up the majority of the set.

It is pretty easy to imagine that Artifact would embrace this philosophy. The fundamental game mechanics of Artifact are fairly novel compared to games currently on the market, meaning there is no need to differentiate itself with needless complexity on the actual cards. There is a lot of room for simple and elegant designs that give rise to interesting gameplay. This should also be a warning to anyone who is expecting all of Artifact’s cards to be loaded with confusing rules text. This isn’t necessary for making a compelling game.

Mathematically Balancing

Screen Shot 2018-04-18 at 1.28.08 PM.png

With all this talk about stats and data, this begs a question that might occur to some: why don’t Valve and Dr. Garfield just come up with a giant mathematical model for balancing? There are a lot of smart people who do lots of mathy-stuff in Valve, and Richard has his PhD in mathematics, shouldn’t it be possible to come up with a formula to balance a “child’s card game”? Well, not really. Dr. Garfield actually told and interesting anecdote about attempting this approach in one game, and basically found it was impossible. As he was designing cards he would add lines of code to his equations to adjust for different variables, but he found that after a while it seemed like every new card was adding a new line to his code, and the whole project was just causing more problems than it was solving.

In fact, it seems like Dr. Garfield pushs against the premise that there is such a thing as the “right” balance of games. He specifically talked about Dominion, and said that he has played the game changing the point value of some of the cards (specifically changing Provinces from being worth 6 to 5), and said the game was still totally fine. Obviously there are “wrong” ways to balance things, but working towards some kind of theoretical “perfect” would rarely improve a game.

For now it seems as if balancing still can’t be done on an a priori basis, and you will need to rely on the same old exercise of game designers using a combination of intuition and playtesting. Of course, since there is no “ultra-formula of mathy-super-balance” guiding the process, we can probably expect that Artifact’s balance will not be perfect out of the gates. Dr. Garfield has made a few comments on patching, so let’s take some time to chat about that.


Patching and expansions are just normal parts of modern game design, especially in the realm of computer gaming. It probably is not a shock to you that Dr. Garfield is comfortable with this practice, and thinks it adds to games. The one comment that I specifically wanted to point out and discuss came from an interview with Game Informer. I am actually going to give you a screenshot of the exchange, since the specifics seem important. JB is Jeep Barnett a Valve employee, SE is Skaff Elias, a friend of Dr. Garfield who seems involved in Artifact, and RG is Richar Garfield.

There are a few interesting points that can be drawn from there. First, the patching approach they are aiming for seems to be somewhat minimal. There are a number of games that are constantly patched, sometimes as often as once every couple of weeks, leading to slow but constant evolution in the game. It seems like the Artifact team is going to be a little more hands-off, and are comfortable with things being a little broken from time-to-time. This should serve as a cue to the community that begging for nerfs and buffs for every card they see as flawed is probably a waste of time. I know writing this is also of a waste of time, since some people will do that no matter what, but it is useful to note that these guys might be a little less susceptible to that kind of “BabyRage” than most.

As an aside, I should mention that although I am joking about players whining about balance changes, I am honestly fine with people doing it under the correct constraints. Giving constructive feedback about cards might not result in a direct change to your specific card of interest, but it could also result in changes to future cards. I am just not a fan of the constant cycle of “ban this, nerf that, everything is OP” I see in almost every gaming community, especially when it is just mindless bitching.

One specific element of this interview that struck me was the notion that there was no need to buff cards, and that the team expects to never buff a card… why? What is bad about buffing cards? For example you could find that a specific class of strategies is overpowered, and rather than nerfing 3 or 4 cards one-by-one, you could just buff one and achieve the same effect. In addition, I think having cards buffed feels better than nerfing. When you nerf a card it feels like you are taking something away from me, but when you buff something it feels like you are giving me something new to play with. Obviously buffing has its limits, and I have no fantasy about aiming every card as top-tier constructed playable, but why not buff cards on occasion if that is the most efficient way to achieve balance. Perhaps my appetite for slightly unbalanced metagames is just less than that of Dr. Garfield and crew, and I will admit that they are the professionals, but I find this curious.

Finally, let’s discuss the idea of releasing new cards in order to maintain balance. This idea might be more broad and complex than people realize. For example, players who are only familiar with Magic and Hearthstone might expect this to mean all card releases happen in the form of new sets or expansions. This is obviously going to be a part of Artifact, but it should be noted that some game companies (Direwolf Digital from my experience) has started releasing “Promo cards”. These are cards that do not belong to any one set, but are released in the form of play rewards for a limited time. You could imagine that instead of buffing an old card, they could just release a new card. This would even circumvent my criticism of “buffing feeling better than nerfing” since getting new cards is even better than buffing old ones. We do not have any news about content micro-releases like this, but it is clearly an interesting approach to both collection building and balance, and it is possible Valve could be planning to use some variant of this strategy.

I am also curious how future card releases might work for hero-specific cards. Are the Axe-specific cards going to be the only Axe-specific cards over the course of the entire game? Might we see different hero-specific cards released at some point in the future? How would that work? Maybe we see alternate “builds” of certain heroes, where they have a different mix of hero-specific spells, abilities and stats, and you may be limited to one copy of a specific hero per deck? This is just personal curiosity, and I have no answers at this time, but I do wonder.

Snowballing Versus Catch Up Mechanics

There is a neat little conversation in the Dr. Garfield’s textbook on the subject of “snowballing” and “catch up” mechanics. If you don’t know what that means, let’s use some examples from Monopoly, which I am sure almost everyone has played. If you are the first person in the game to assemble all the property of a given color (such as New York, St. James and Tennessee, which are all the orange properties) you are at a massive advantage in the game. This lets you begin building houses, which give you tons of rent, which you can then push back into building more houses and hotels. This is snowballing in action.

The flipside of snowballing is a catch up or comeback mechanic. In Monopoly many people play with the house rule where all money spent in Community Chest or Chance goes into a pile in the centre. This money builds up until someone lands on “Free Parking”, and that player gets all the money. Part of the reason why this might be so popular is because it gives the players that are behind a reason to stay in the game. If I am struggling, but there is a fat stack sitting in the middle of the table, I will stick it out for the chance to run hot and turn things around.

I am sure you have seen or experienced many games that have elements like this. The point that I wanted to bring up from Dr. Garfield’s book, was the notion that catch up and snowball mechanics are actually just illusions. In the Monopoly example the player who assembles the first set of properties is already at a wild advantage in the game at the moment they complete their set. The net worth of the player does snowball over the course of the game, but if you were going to plot “chance to win” over the course of the game, the main jump would happen when they complete their set. The Free Parking catch-up example also works. If you were to just assess the net value of all the players, the worst off might appear way behind, but when they score Free Parking their net worth spikes. If, instead, you were following the odds of winning, you would need to factor in the chance of hitting Free Parking as part of that calculation, meaning their position might look better than their net worth implies.

This discussion is almost philosophical in nature, so let’s turn it to something more practical. It seems to me that Artifact’s creators don’t seem to be afraid of either snowbally effects or powerful catch up mechanics. Judging by the core mechanics of the game, it seems to me the fundamental rules lead to some snowbally game play. If you have a burly hero and a decent squad of creeps you will probably become really good at just mowing down anything your opponent puts in front of you. This allows players to just dominate a lane until the opponent takes some action to inteterfere. There are a lot of possible ways to build catch-up mechanics into the game, including powerful swingy spells like “Annihilation” which wipes out all the units in a one lane. Other examples could include prices on “secret shop” items being randomized within a range. You would have some low probability chance to pick up a sweet weapon at a heavy discount. This might add compelling comeback potential to the game.

“But Neon! That sounds random, and random is bad! Everyone knows that we want as little randomness in our games as possible!!” Well, in this case I have some bad news to break. Not only is randomness part of Artifact, it might actually make the game better!

Luck Versus Skill

Dr. Garfield thinks luck is misunderstood. Many people talk like everything is either a game of luck or a game of skill. Chess and checkers have no luck, and are clearly games of skill, so they are good games. Snakes and Ladders or Rock-Paper-Scissors are games of luck, so they are not good games. Well, things are more complicated than that. Take poker for example. This is a game that clearly takes a lot of skill to play at the highest levels, but luck is also a factor, so where does it stand? Dr. Garfield argues that luck and skill should really be thought of as two separate vectors, and games can have a mix of luck and skill. You can see a visualization of this perspective below, with Artifact placed in the range of “high skill, medium luck” which is where I expect most trading card games aim to be.

For the record, Dr. Garfield would object to my classification of Chess being a “no luck” game. Technically any game could be won via the “infinite chimps on infinite typewriters…” approach. Personally I think that is a stretch, but I will just take this to mean that I could beat Faker at LoL with the right amount of luck!!

So we have established that luck isn’t inherently bad to have in a game, or that low-luck games are not inherently better than high-luck games. But what does luck and randomness do for you? Why would you put it in? Actually, there are a lot of reasons.

  • Luck increasing variety – when luck is part of your game then your games are more likely to play out differently than one another, since you are not seeing the same pieces in the same order ever time.
  • Luck protects egos – losing sucks, and when you know you lost because you were just outplayed it really sucks. Having some randomness in your game protects players from the full force of feeling like they are bad. Maybe you lost just because you were unlucky, and your opponent ran hot? Obviously this does happen in real games, but bad play almost always plays a role.
  • Broaden audiences – match making is quite a challenge in games. If you are against someone who is a lot worse than you the game is boring because you are going to feel like you are taking candy from a baby. If you opponent is much better then you, you get the experience of being the baby whose candy is stolen. Your most interesting games are going to be against a small group of people that are approximately the same skill level as you. Variance helps broaden that range, since it gives you a better chance of wining against players that are much better than you, as well as lower skill players offering a challenge.
  • Catch up – up above I go off on the idea of catch up mechanics in games. It is pretty easy to imagine how variance can be used to give increased chance to players that are behind.
  • Skill + Randomness – something many people may not appreciate is the skill of dealing with randomness. Take for example the stories you hear where a player is in a bad spot, but plays perfectly so they can draw the one card to win the game. This specific style of planning is quite a skill! I recently did a podcast on the skill of playing around randomness, if you want to check it out. It is specifically surrounding the game Eternal, but many of the lessons cross over to card games more generally.

This is not even a complete list. If you want to learn more about this topic I would specifically check out this talk by Dr. Garfield, and you can also read an article about randomness in game design that goes into a lot more depth (once again, this is focused on Eternal, but most of the lessons apply to card games more generally). It is very hard to guess how randomness will be implemented in Artifact, since there are so many possibilities on how this could be executed. Some of the places we know randomness can already be found include drawing cards from your deck (without a redraw rule), placement of creeps, attacking direction in combat, and items that appear in the store. One of the things I find frustrating in Garfield’s talk on this subject is how there are no specifics on where randomness should go, when it should be applied and how much. How does Dr. Garfield know that a game needs more or less randomness? He does hit on the idea that games often benefit from getting less random over time, but that is not really answering questions for us now. This video released by Lysander last week was an interesting take on the subject, which I encourage you to check out. I obviously agree with the idea that RNG is good for games if used properly, though I have some disagreements on their thoughts on where RNG should/should not be used. I am interested to see exactly where Dr. Garfield and friends land, and then evaluate that in light of his talk.

Screen Shot 2018-04-18 at 1.51.32 PM.png


Let’s talk now about what a “metagame” is. I’m sure some of you are already rolling your eyes, thinking “I obviously know what a metagame is! I’ve been playing games for years, and I talk about metagames all the time in the DOTA/HS/MTG Reddit!” Well, the definition most hardcore gamers use to describe “metagames” is not what Richard Garfield thinks about when he talks about “metagames”. When gamers use the word it typically is referring to the mix of available strategies or what styles of gameplay are most popular. Dr. Garfield use of the term is much more broad, and includes all elements of the game that are outside of the actual playing of the game. Tournament organizations, discussion forums, and even fan-art are, in a sense, part of the metagame. It is actually funny to think that this article is part of the Artifact “metagame”, while the game itself is not even available!

The topic of a “metagame” in the sense that it is used here was discussed in Dr. Garfield’s textbook in moderate depth. What I found interesting about this concept was imagining what Artifact might look like if thinking about the metagame was incorporated during the design process. Often, games are made with only limited attention to metagame features. Magic is an extremely obvious example of this. When Dr. Garfield originally designed the game I expect he would never have imagined things like the Pro Tour, cosplaying, or podcasts all revolving around his game. These were natural outgrowths of a game that people were passionate about. Now, let’s imagine that Artifact is built with the goal of supporting this metagame from day one. What might that look like? How might they accomplish this? Obviously you wouldn’t want to compromise the strategic depth of the game for the purpose of propping up a metagame, but you might prioritize adding features like spectating, or dedicate resources pre-emptively to the competitive scene. Many of you have probably already heard about the 1 million dollar prize pool tournament that Valve is planning for Artifact. This seems like a bold approach to push the growth of a metagame from the beginning. One could argue about whether this pushes the metagame in the healthiest possible direction, but at the very least it jump starts things pretty aggressively.

One facet of the metagame that deserves its own separate section is game economics.

Game Economics

Most of Dr. Garfield’s writing has a tone of objectivity and openness to it. There is a lot more “These are things you observe in games” rather then “This makes a game good, and this makes it bad.” There are a lot of paths to create good games, and if you apply the correct combination of tools you might make a great game even if you violate some of the principles of “good game design”. The one area where I would say Dr. Garfield has a different approach is game economics. In fact, he not only has opinions on good and bad practices; he takes a strong moral stance on certain monetization strategies.

Here is a link to the “A Game Player’s Manifesto” by Richard Garfield. It is worth reading in full, but I can summarize some of the most important points:

  • Some games can be described as “Skinnerware”, which are essentially designed to prey on specific parts of human psychology.
  • Not every game player is equally susceptible to addictive properties of certain games, but some are extremely vulnerable, and they are easily exploited with Skinnerware games.
  • Many Free-2-Play games are paid for largely by a small fraction of the population that spends huge amounts of money. These people are known as whales.
  • There are a lot of misconceptions about whales. While some are fabulously wealthy, many are people limited means. When they dump tens of thousands of dollars into a game they are ruining their lives.
  • While there is nothing wrong with making money off games, it is wrong to deliberately structure games to hunt whales.
  • Not only is this morally wrong, it generates bad games. For example, adding purchasable power-ups to single player games is just bad game design since it means that either the powered or unpowered version of the game is the intended version. Only part of the player base will experience the intended version.
  • It is good for games to have a cap on how much you can spend. If you can buy everything then players will never be tempted to keep dumping money into the game. 

Before moving on to an analysis of what this might mean for Artifact, I want to just touch on a couple things. Dr. Garfield doesn’t really expand on where the term “Skinnerware” comes from, so I will give you a brief primer on operant conditioning. B. F. Skinner was a psychologist from the mid-20th centaury that followed up on some of the work by Pavlov. Some of his most famous research focused on training pigeons to do different tasks by rewarding them in various ways. While the specifics of his experiments had some interesting implications for psychology, his research gave us some powerful tools that would be used to train people to do certain behaviors under the correct conditions. By Skinnerware Garfield seems to be referring to games that exploit Skinner’s psychological techniques into their monetization model.

Games that leverage lessons from Skinner’s work could often be characterized as being a lot more addictive than they are fun. If you want to learn a bit more about this subject check out this video by Extra Credits. In fact, if you like learning about game design, just go watch all their videos. Also, Garfield’s manifesto responds to a common defence of the practice of making Skinnerware, which runs the logic of “People spend what it is worth to them”. I recently listened to a podcast by Ezra Klein interviewing Tristan Harris (link) that touches on this subject more broadly in relation to social media and technology, and how our “revealed preferences” are not actually reflective of our true preferences. If your day is going a little too well, this interview will give you great chance to dwell on the ways technology is ruining the fabric of society.

So what about Artifact? To begin with, card games as a genre have some resistance to this problem. It is possible to complete your entire collection, meaning there is just nothing left to buy. In the past, I have criticized “F2P” card games that had bad F2P models, but at least there is a limit to what you can spend. Also, you should not expect the game to offer stupid stuff like power ups that you can buy. If you have played a lot of “free-to-play” games you might be able to buy (with IRL money) potions or spells that you can play even in PvP encounters. This is a really big turn-off to most hardcore gamers, so I was not expecting to see them, but it seems like Dr. Garfield is not interested in such things either.

Where I am most interested is how this related to tournament structure. From the sounds of it, Artifact will offer in-game tournaments essentially from launch. We know very little about the structure and regularity of these tournaments, but if implemented poorly, they could be a version of Skinnerware. I think for example of Magic the Gathering Online. The game is quite bad from by almost every metric, but one feature I find specifically distasteful is the relatively high stakes in real life currency in many game modes, as well as prize structures that are not particularly generous. Artifact will have a lot more polish than MTGO, so I worry that a well-executed version of a similar approach to competitive play could be exploitive. Given what I have seen in Dr. Garfield’s writing I am not that worried, but it is something to bear in mind.


Game design is a fascinating and deep topic that I enjoy discussing, reading and writing about. I have spent the last 2 years looking to Mark Rosewater (head designer on Magic) as something of the figurehead of the modern card game design seen across the industry. His thinking has been quite influential for card games designers that came after him, but Dr. Garfield is one of the few who have actually came before Rosewater. If I was going to keep writing about game design for Artifact I knew I needed to do a deep-dive on Garfield’s philosophy. Rosewater was an acolyte of Garfield, which would suggest a very similar outlook on gaming, and while there is substantial overlap, I feel like there are some interesting differences that I picked up on through this study.

Hopefully you all found this deep-dive on Garfield’s design philosophy was interesting. I really enjoyed writing it and researching it. If you are passionate about these topics you should honestly read his textbook. There are a number of interesting topics that I either had to skip, or did not go into with as much depth. The writing is also quite engaging for a textbook, as you can see in these quotes:

“Sports – Players can instantly react to the physical acts of others due to the wonder of the interface that is real-life physics.”

“Taste is not commonly a factor in games, except perhaps in sports if things go drastically wrong.”

I would also be remiss if I did not shout out to Skaff Elias as well. He was an early Magic creator, was involved in writing the textbook with Garfield (along with Robert Gutshera), and he seems to also be attached to the Artifact project in some capacity. I like to imagine that some of these concepts and ideas were hashed out over the decades these two have known each other. If you have any thoughts or feedback on the piece be sure to check out the Reddit thread to share them! Also, be sure to check out my guide to Artifact that came out last week. Until next time,