Anonymous said: Which do you like better : Dungeon Crawl Stone Soup or Nethack?

I haven’t really played Nethack. From what I’ve seen of it (Let’s Plays and the like) it seems to have a lot of funny exploit-like strategies, things that experienced players do without thinking that newer players would never think of. Crawl has some of these as well (stair-dancing comes to mind), but it doesn’t seem to have quite as many. These kinds of strategies bother me, because they’re ultimately kind of shallow. Experienced players just do them automatically, while inexperienced players fail to do them and are punished for it.

I prefer DCSS for this reason, although I’m biased because I’ve played DCSS. I like DCSS for the sheer variety of distinct characters you can play and magic that you can have access to, and I’m not entirely sure how Nethack matches up against it by those criteria.

Recently I’ve been playing Brogue, which is much prettier than either Nethack or DCSS, and worth a try if you’re into screens full of colorful ASCII symbols.

Anonymous said: Can you give me an interesting thing to ask? I can't come up with one on my own.

You should ask why game design is an important skill for policy-makers and people in management positions.

Lucky Noob

As a medium, games are unique for the variety of experiences they provide even within a single game.  When you play a game for the second time, you can expect it to remain exciting, interesting, and engaging.  This expectation presents game designers with a challenge: how can my game provide interesting variance of experience and remain exciting indefinitely?

Variance Tools

Dice are a very popular variance tool.  They are very flexible in the degrees of variance they can provide.  For example, a normal six sided die will produce a random integer between 1 and 6, with an equal likelihood of each outcome.  Adding a second d6 (a shorthand for the number of sides on a die.  ndx means n dice with x sides each) almost doubles the range to between 2 and 12, but changes the weighting so that the most likely result is 7.  As more dice are added, the range becomes wider, but the percentage of the range where most of the possible rolls lie decreases.  Each die adds its possible values to the range, as well as making the dice pool as a whole more consistent.  Additionally, dice with fewer sides have a smaller range, and therefore roll closer to their average value.  Taken together, these properties allow game designers to control a random value’s reliability and possible range.

Three different probability curves achieved with only d6s

But the true value of dice lies in their ability to produce tension.  In tabletop games, dice rolling is loud, wild, and physical.  The scattering of the dice across the table when they are rolled can be very climactic, if at times a little aggravating as miscast dice must be retrieved from the floor.  Each die roll is an independent event, meaning previous rolls have no impact on its outcome and the result can never be certain.

Dice are used often enough that a large family of dice-based variance tools has been developed over the years.  One of the more interesting techniques I’ve seen is “exploding” dice, in which the highest possible rolls on a given die cause that die to be rerolled, and the value of the rolls to be added together.  For example, if a six is rolled in a d6, that die “explodes” and is rerolled.  If the reroll is a four, the resulting value of the die is a total of ten.  Rerolls can explode as well as original rolls, so it is technically possible to roll any positive integer value that is not a multiple of 6 on an exploding six sided die.  This exploding dice system greatly affects the range of possible values and player perception without greatly affecting the average value of a given roll.  Dice explosions are exciting events, as values skyrocket past player expectations, but the average value of a given roll isn’t majorly affected by it.

Cards are another old standby, with their main selling points being manipulable variance.  A deck of cards can have more copies of one card and less of another to manipulate the chances of drawing both cards.  If more cards are drawn, the chance of a predictable result is higher (this is a theme - the more individual random events you have, the more consistently they will perform).  Designers can allow players to rearrange the cards in a deck, or shuffle it at various times to create an unknown ordering.

Over the course of a game, if a deck’s contents are known, players can speculate more accurately about cards that might be drawn from it.  Drawing a card from a deck is a dependent event, meaning prior draws affect each subsequent draw by changing the deck’s contents.  While there are enough cards that uncertainty remains, this can be very exciting and rewarding, as players can adjust their strategies to the possibilities the deck still contains.  However, as the deck shrinks, the guessing game becomes more of a predetermined certainty, and the limited possibilities that remain may make playing onward pointless as the outcome is already determined.

Modern card games have taken advantage of a crucial facet of cards and decks: they are physically manipulable.  This means they can be bought, sold, and traded as individual entities as well as groups, and a deck can be easily assembled from any group of similarly backed cards.  By contrast, it is much more difficult for a player to assemble a die.

Random Number Generators (RNGs)
RNGs use hidden lists of arbitrary numbers to determine the results of a given event.  One advantage of an RNG is the complete control a designer has over it.  RNGs can simulate dice, deck shuffling, and coin flips with ease.  The RNG’s greatest advantage is its ease of access to great computing power.  An RNG can generate thousands of die rolls and have them all put to use right away.  Clever designers don’t confine themselves to simulating dice when they have access to such a tool.

Computing power is especially useful in making elements of chance more palatable.  The human intuition when it comes to independent events is truly horrible.  There is an expectation that dice shouldn’t roll the same thing twice or more in succession, or that there shouldn’t be a long run of die rolls without a certain result coming up.  These intuitions are brutally misleading, and lead people to be upset by random events.  A clever random number script can respect these intuitions by making them somewhat true.

Riot Games’ League of Legends has put this into practice with their critical hit system.  With a pure chance system, players could score multiple critical hits in a row on their opponents, even with a low % chance of doing so.  Conversely, a player could also suffer a long dry spell without any critical hits.  When players complained, Riot changed the critical hit system to match player’s expectations while in effect maintaining the old percentile based system.  The newer system lowers a player’s critical hit chance after successfully scoring a critical hit, and raises that chance whenever a critical hit is not scored.  The degree of these adjustments is different depending on the player’s critical hit chance statistic so that the overall hit percentage remains the same as it appears in a character’s stats.  Essentially, this system acts as a deck of cards whose contents are constantly being rearranged.

Snake Eyes
Randomness can be a great obstacle to a player’s enjoyment of a game.  It thwarts players’ intuitions, making them feel as if the random element is unfair.  People can feel victimized when a dice rolls something it “shouldn’t” roll, when their opponent gets two critical hits in a row, or when their Magic: the Gathering deck gives them the worst possible draw.  The rationale is, “I couldn’t have planned for this.”  In these instances, the random elements can punish players for making reasonable decisions based on the information they had access to.  It isn’t a good idea to play as if all the most unlikely negative events are bound to happen to you.  In most chance-based games, doing so will cripple your ability to act.

Overused variance can make players feel powerless.  When dice determine the outcome of all player actions, it can be impossible to make informed choices.  Without the information to properly make decisions, the player is robbed of agency and their choices become meaningless.

It wasn’t very long ago that I thought randomness was an irredeemably bad design decision.  It’s an easy trap to fall into, given the constant complaints about unlucky breaks and game-deciding die rolls that proliferate gaming communities.  These bad stories are the flip side of a much brighter side of randomness.  In the right setting, random outcomes are exciting instead of frustrating.

The Skill of Luck
Variance is one of the most skill-testing elements a game can employ.  This is quite counterintuitive.  When players think of luck, they imagine their opponent, having made a multitude of mistakes, receiving some arbitrary bonus to defeat them.  What they don’t think about is the degree to which variance expands the possibility space of the game.  For example, Tic-Tac-Toe is a game that exists at the extremely low end of the variance spectrum.  This makes the game’s possibility space so small that it has been mapped out in its entirety, and the optimal moves are obvious to experienced players.  It would be ridiculous to assert that Tic-Tac-Toe is a more skillful game than Poker, Blackjack, Magic: the Gathering, or almost any other high variance game; there’s simply nothing to it beyond memorizing the optimal moves.

Chess may be brought up as a counterexample, a low variance game with a large possibility space.  This is true, but Chess’s depth is not without a cost.  It’s a consequence of its truly monstrous complexity.  Consider that there are twenty possible opening moves, and that the number of possible moves can often increase from there as the more mobile pieces are deployed.  Variance replicates this possibility space without the complexity burden.

Elements of luck might not always raise the skill ceiling, but they often can, and they certainly aren’t guaranteed to lower it.

A New Age of Storytelling

You won’t hear it from me very often, but games are a great storytelling medium.  All over the internet, message boards and social networks discuss and break down the stories of their favorite games, talking about favorite characters and plot twists.  Mass Effect 3’s lackluster endings provoked a huge fan backlash, demonstrating how important these stories are to the players.

Videogame stories offer their creators unprecedented control over the way their content is presented.  Game makers can tell stories that are told in a mixture of text, video, and game rules.  Because these elements can be combined so seamlessly, it isn’t enough to think of these stories in terms of which physical methods are used to convey information to the receiver, as has traditionally been the distinction between media such as films and literature.  However, we can examine how the story experience is generated and presented to the player and use this information to distinguish between types of videogame storytelling that are truly distinct.

In these examinations, its vital to ask who is telling the story.  This is because unlike non-interactive media, videogame storytelling is not solely the duty of the media artifact’s creator.  It often falls on the players or even non-human agents to create content in modern videogames.

Videogame as Narrator
In many ways, modern blockbuster videogame storylines are much like those of other media.  Players are introduced to characters and they experience rising action peaking in a climax and ending with a denouement.  The stories are often somewhat interactive, with a few branching choices and alternate endings.  These games are about expressing a story that has already been authored, and player choice in them is mostly superficial.

Narrator stories are very popular.  Call of Duty, Mass Effect, The Final Fantasy Series (mainline, not necessarily true for other members of the brand), Metroid, Legend of Zelda, and most other games that you’ve heard of.  They can tell very deep stories and explore storytelling methods untouched by more traditional media through the power of their immersive nature.

Videogame as Coauthor
There is another model of videogame storytelling, although it isn’t hailed as such.  In this model, the game has no definitive, unchangeable, or linear plot.  Everything that happens is determined by the rules of the game and the interaction of the player or players.

The Coauthor model can have a setting, characters, and rising action.  In fact, a good game following that model almost always has all three.  The difference is that the plot is a living thing determined by the course of the game instead of a folder of cinematics stored on your hard drive.  The Coauthor model generates stories out of player actions.

In some senses of the model, almost any game could be considered to use the coauthor model.  Chess has characters (the pieces) whose interactions are determined entirely by internally consistent game rules and player actions.  Those are the fundamentals of a coauthor system.  However, the model isn’t really relevant unless you are using it to try to create interesting stories.  Game as coauthor is a storytelling method that’s only applicable when the game is supposed to tell interesting stories.  For the purposes of this article, I’m going to exclude games like Civilization, which could be construed as having an story that involves player input, but doesn’t really have anything resembling compelling characters and deals entirely in abstractions.  These games aren’t truly interested in storytelling, and instead find their purpose in providing a series of interesting strategic decisions for their players

Dungeon Crawl: Stone Soup is a prime example of the Coauthor model.  It has a loose setting, a miscellaneous dungeon built out of intentionally designed but randomly selected levels.  Its cast of characters is gigantic, but randomly selected each time you start another game.  You are the protagonist, but who you are depends on a choice of 24 species and then another of 28 classes (although some combinations, like Mummy Priest, are disallowed).  You will fight hordes of monsters, worship one or more gods, and build a great collection of magical items.  In the end, you will probably die, whether it be in a Lair of Beasts, deep in the Orc Mines, or at the hands of Sigmund, a murderously powerful wizard with a scythe.  If you succeed, it will be because you reached the bottom of the dungeon, retrieved the Orb of Zot, and made it all the way back to the surface.

The types of stories Dungeon Crawl tells are not traditional stories.  They aren’t economically designed or deeply profound, and they won’t win any literary awards.  What they are, however, is practically inexhaustible.  You could die a thousand deaths in Dungeon Crawl and still be willing to play again.  The array of possibilities Dungeon Crawl gives the player affords them the freedom to carve their own path through the designers’ content.

Advantages of the Coauthor System

It makes players feel free.  Oregon Trail is a game that people my age are widely familiar with due to its frequent presence on school computers.  The player takes on the role of a family heading west in a oxen-pulled wagon to start a new life on the frontier.  It’s mostly a game of random chance, with a sprinkling of player choice, but it shows its coherence to the coauthor model by giving the player the possibility of failure, often via deadly dysentery.

You’ll notice with the examples I’ve used to describe coauthor games, random chance is often a significant game element.  This is because nondeterminism is very important in the coauthor model.  Narrator games often have possibility “bottlenecks”, in which plot branches are consolidated and pushed into the main line plot. The coauthor model never entirely collapses its possibility structure.  At the end of every game of Diablo II, the player will have completed all the quests, watched all the same cinematics, and killed all the end of act boss monsters.  At the end of a game of Oregon Trail, they could be dying of dysentery on the Green River or living it up in San Francisco with their 99 Grandfather Clocks.

It gives games replay value. Sid Meier’s Alpha Centauri is a sequel to Civilization and reuses most of its mechanics.  Despite this, it’s beloved by critics and remembered with great fondness for an engrossing story that accompanies its mechanical complexity.  Rather than the loosely characterized historical figures of Civilization, Alpha Centauri allows players to play with and against seven ideologues, whose personalities and backstories are explored through snippets of text distributed throughout the game.  The player is shown one of these snippets whenever they do anything significant: constructing a building for the first time, researching a technology, completing a secret project.  Alongside this there is a narrated story that players are shown which is triggered by events that happen in the game.  This story provides more characterization and ties in with the quotes, but takes a back seat to the coauthored story of the player’s exploration and conquest of Chiron, the planet where SMAC takes place.

Alpha Centauri tells the story of a clash of timeless ideas.  The environmentalist Gaians, led by Lady Deirdre Skye, inevitably clash with the capitalist Morganites, led by CEO Nwabudike Morgan, over their disagreements of economic policy (which is an important choice the player makes in the game: green economy, or free market?).  This clash of archetypes is typical of Alpha Centauri.  Other factions include the fundamentalist Lord’s Believers, the militaristic Spartans, the high-minded Peacekeepers, the amorally scientific University, and the repressive, communist Hive.  Each faction is deeper than a mere stereotype, while still being emblematic of its given theme.

A lot of what Alpha Centauri does could be done in a traditional Narrator model, but ultimately the coauthor model is crucial to its appeal.  The model lets the clash play out differently every time, which allows the game to be impartial where a more narrated model couldn’t be.  While the setting and characters are mostly static, the game is a fractal possibility tree of how they might interact and who might ultimately prevail on their new world.

It’s more efficient.  In a game like Mass Effect, every story branch and piece of dialogue has to be authored individually.  As a narrated game’s story gets longer, the amount of work it takes to author increases exponentially unless it remains mostly linear.  Every branch and sidequest adds content that many players might not even reach which has the same development cost as content all players can reach.  This places a large economic incentive on linear storylines in a medium where audiences have come to expect nonlinear and branching storylines.

For large stories, the coauthor model can produce far more content with far less authoring.  It doesn’t rely strictly on creator labor, instead harnessing player activity and algorithmic randomness to make interesting things happen.

Radical Storytelling
It’s not unexpected that the games that are most well-known both within the community and in the mainstream media use traditional storytelling methods.  These methods produce stories that are accessible, enjoyable, and unifying, with every player told the same story, and thus able to discuss it with each other on common ground.  As safe as these stories are, it can be difficult for less traditional storytelling methods to get their moment in the spotlight.

Coauthorship models are radical, and games that use them can be easily confused with games that lack a storytelling model altogether.  Coauthored stories often resemble stories about games, rather than stories told by games.  There is, however, a fundamental difference between just playing a game and taking part in the creation of a story.  Playing a game is about the players: their feelings, their struggles, and perhaps their victory or defeat.  Storytelling is about the inhabitants of the story world of the game.  If the players appear in the story, it is only through avatars, characters that represent them within the bounds of the story.  This is a confusing distinction to make, but critical in fairly examining storytelling methods.

Being a mechanically minded person, I tend to prefer storyless games.  The story gets in the way of the raw gameplay system, and leads other players to demand unreasonable things like “realism”.  Coauthor model games, however, have a special appeal to me.  They remind me of the type of games I played as a kid.  The stories I came up with when playing with Legos were never solely the sum of the bricks, or pulled as whole cloth from my brain.  It was always a contest, the limits of what I could build pitted against the limits of my imagination.  I’m glad for coauthored games because they give me an outlet for this type of behavior while also offering me new systems and mechanics to play with.

I think as expectations as to how many possibilities a game should offer you rises, we’ll see a natural increase in the rise of coauthor model games.  The economic incentive is fairly significant, as is the emotional attachment these games can generate.  The model works especially well alongside a game that is continuously updated like Dungeon Crawl because of its ability to attract long term players who are always looking for the next update.  I look forward to the day that blockbuster games will excite me with highly nonlinear stories and be hailed as unique storytelling experiences.


Sid Meier, creator of the acclaimed Civilization series of history-spanning strategy games, once described a game as a “series of interesting decisions”.  His vision of a game is an intellectual challenge, focused on the player’s strategy and its consequences. 

While intellect and strategy are key aspects of games, the visual, aural, and textual trappings that communicate them to the player deserve attention in their own right.  Memorable and entertaining moments in games appeal to both the intellect and the senses. I call methods that create these moments “visceral game elements”, and their aim is to impact the player and imprint the game into their memory.


Visuals in games are one of the easiest ways to affect players and create memorable moments.  While modern video games are filled with flashy visuals, some of the strongest examples of this come from the least graphically sophisticated games.  The following screenshots are from Dungeon Crawl: Stone Soup, a recent member of the lineage of Rogue, a game designed at UC Santa Cruz in 1980 (Go Fighting Banana Slugs!).

Rogue and its descendants, called “roguelikes”, give the player control of an intrepid adventurer (represented by a heroic @) searching for a powerful artifact at the bottom of a dungeon full of monsters.  The game is turn based and displayed entirely in the 128 letters, numbers, and typographic symbols of the ASCII character set.  The right side of the screen is full of information about the player character, while the left displays the surroundings of your character.

Olgreb, the Naga Venom Mage, discovers a room full of hostile orcs.

Olgreb emits a brilliant green light, sickening all those it touches.

Many readers will look at these screenshots and not understand the drama or impact of the change in the display.  This demonstrates a principle of visceral design: only players who are are otherwise invested will be viscerally affected by game elements.  Visuals can be universal to an extent, but ultimately players need to understand the meaning of the images.  The more abstract an element or visual is, the less viscerally affecting it is.


Sound design is essential in giving players visceral feedback.  A great example of this is the character Dr. Mundo from League of Legends.  Dr. Mundo is a crazed, roid-raging madman who throws meat cleavers.  These cleavers travel as unguided projectiles in a straight line for a fixed distance and damage the first thing they hit.  The satisfying “THUNK” of your cleaver hitting an opponent is immensely satisfying, and makes the effort of aiming worthwhile and memorable as a testament to your skill.

On-hit reward sounds are a fairly common and effective practice.  Team Fortress 2, a class-based first person shooter, has a great option where every time you deal damage to an enemy, a little bell rings.  In a game that can often be hectic and visually confusing, the audio conveys little messages of success that convey positive reinforcement and make the player feel satisfied..


Text in games can sometimes look like this:

Player A slashes Goblin for 5 damage!

Goblin attacks player A and misses.

Player A slashes Goblin for 10 damage! Critical hit!

Gobin dies.

Player receives 5 gold and 5 exp.

It’s generic, boring, and we can do better.  Anyone who’s read a novel knows that text can be a visceral experience.  A game called Dwarf Fortress shows that visceral text isn’t unique to literature.  Dwarf Fortress is technically a roguelike, but gives the player loose control over a band of dwarves who must strike the earth and build a home for themselves that is self-sufficient and able to withstand attacks by goblins, demons, and forgotten beasts.  Combat between creatures isn’t shown in animations, but rather described in combat logs.  Dwarf Fortress’s combat simulation is much more complex than most, including representations of fat, skin, bone, muscle, and brains, all of which can be torn, bruised, broken, or sometimes melted.  Here’s a sample of a Dwarf Fortress combat log:

Callum hacks Urist in the upper body with his steel great axe, tearing apart the muscle and tearing apart the middle spine’s nervous tissue through the large cow leather armor!  An artery has been opened by the attack!  A tendon in the middle spine has been torn!  The steel great axe has lodged firmly in the wound!

Urist loses hold of the steel spear.  Urist falls over.  Urist misses Callum!

Callum twists the embedded steel great axe around in Urist’s upper body!  Callum hacks Urist in the right front leg with his steel great axe, tearing apart the muscle through the large bronze high boot!  An artery has been opened by the attack and a motor nerve has been severed!

Callum hacks Urist in the upper body with his steel great axe, tearing apart the muscle, shattering the left false rib and tearing apart the liver through the large cow leather armor!  An artery has been opened by the attack!  A tendon in the left false rib has been torn!  The steel great axe has lodged firmly in the wound!

Callum twists the embedded steel great axe around in Urist’s upper body!

Urist has bled to death.

(Combat log is from, names changed to protect the innocent)

There are no abstracted numeric measurements of “damage” in this log.  It manages to be objectively gory and descriptive, communicating real game data while still being entertaining to read.  This is truly a great accomplishment.


This is probably the most unusual category, but it definitely deserves mention.  Games made of cardboard and plastic can’t rely on the same sorts of visual and audio designs of their digital brethren, nor do they have the textual generation capacity to create detailed and gory combat logs.  However, they have a powerful tool that is uniquely their own: tactile manipulation.

Even old board games demonstrate an understanding of this tactic.  The little plastic Risk dudes aren’t really necessary, and can be quite a headache to keep track of, but moving them en-masse feels powerful and dramatic.  Even dice rolling, frustrating as it can be, has a powerful tactile component that makes it more exciting than, say, using a dice simulator on a website.  While the excitement of hidden information can be easily translated into a digital form, the feeling of holding your fate in your hands cannot.

I notice this very strongly in Magic: The Gathering’s design as well.  When Magic designers want a card to feel dramatic and powerful, they make it move around a bunch of other cards.  A famous example of this is Wrath of God, a card which destroys all creatures.  What this means is that every creature card that players have played so far (that hasn’t been otherwise destroyed) is scooped up and dumped into that player’s graveyard, which serves as a discard pile*.  The drama of a newly clean play area, and of having to toss aside your erstwhile minions, gives Wrath of God and other effects like it an impact that many other spells, even arguably more powerful spells, cannot match.

Immersive Joy

Visceral elements are more than entertaining illusions.  They have a realness to them that allows people to buy into them emotionally and get excited even by an activity as abstract and unreal as playing a game.  When players feel they have a license to play pretend and get involved in their games, their accomplishments and failures in the games become more authentic, and their time investments in these pretend worlds pay off a little bit.  The realness of game worlds is what allows them to compete with the real world for our time and attention and win.

*Magic breaks all of its own rules.  There are lots of cards that interact with the graveyard.  The REAL discard pile is a zone called Exile, although even there cards aren’t safe.


As a companion piece to my long term design essay, here’s some annotated patch notes from league of legends.

Reblogging from myself because the stupid came in the window.

Long Term Design and League of Legends

Long term design has become a particular interest of mine recently.  My interest was sparked by these two articles by Mark Rosewater, Head Designer of Magic: the Gathering (this one and this one).  I was struck by the shear longevity of Magic as a game, and the measures needed to preserve it.  It never occurred to me that producing too much content could be dangerous to the staying power of a game.  This kind of design is particularly relevant in the modern age of game design, with its constant patches and dedicated gaming communities.  A game it’s particularly relevant to is League of Legends.

League of Legends is a successor of the vaunted Defense of the Ancients mod for Warcraft III.  Warcraft III was an innovative real time strategy game that focused on small unit skirmishes and high single unit complexity, with the most salient feature being “Heroes”, special units that gained experience over the course of the game and had four abilities each.  Defense of the Ancients took this premise and ran with it, focusing entirely on the heroes, which were made vastly more complex, and removing all aspects of base-building.  A basic gameplay summary is that there are 10 players divided into two teams, each of which chooses a single hero that gets stronger over the course of the game.  DotA, as it’s often called, was itself an experiment in long term design, with many iterations over the course of its active development (with the most recent update being this January).  League of Legends is best thought of as a more accessible version of DotA, with some of the more frustrating gameplay elements removed or modified.

Long term design is about managing content production, so I’d like to establish some baselines for how that works for these games.  Magic has a quarterly content production schedule, with card set releases at the end of September, January, April, and one in mid-July.  Each of these content releases is somewhat self-contained in that they are designed to be played on their own.  They are also playable alongside any other Magic product in existence.  Old cards are rotated out of regular play and become harder to find, so while there’s a pressure to produce new cards, reprints of old cards are accepted and often rejoiced by the community.  Reprints and functional reprints gain a sense of novelty in the new context provided by the self-contained sets.  Essentially, Magic’s designers get to use the design space of their game methodically, picking off little pieces and exploring all the nooks and crannies.

This forms a stark contrast to the hectic development update and dangerous future prospects of League of Legends’ long term design strategy.  League of Legends updates once every two weeks.  These updates are not self contained and only exist in the context of the wider game.  Much of the content is balance changes and reworks, but each update also contains a new champion.  A champion consists of a block of statistics (health, mana*, armor, magic resist**, attack damage, attack speed, movement speed) a set of four spells, and an innate passive which tends to be a defining trait of the champion.  In other words, the champion design space is very large.  There are tons of possible champions to make.  The danger is that all of these champions have to be similar in power level, as they all exist in the same environment.  League of Legends doesn’t have the luxury of old, powerful game elements phasing out and letting the new things shine, and that is perhaps its biggest problem.

The item system further compounds this problem.  The item system exists as a mechanism for champions to scale throughout the game, and also to allow players a little more customization space to play around in.  Ultimately, the item system is what truly defines champion archetypes and roles within the game.  They determine whether you deal tons of damage to the opponent, or are impossible to kill, or impart valuable bonuses on your allies, or can cast spells constantly.  Your item choices are ultimately more important than your champion’s base stat block.  The item system, as of now, is unbalanced and underdeveloped, with multiple items that are considered unequivocally the best at what they do.  Every item is available to all champions, so the effect of new items can be huge in terms of the unlocking new playstyles.

Whichever direction the designers of League of Legends take it, I’ll be interested in following its changes.  It’s not in any sense too late.  Magic made plenty of mistakes in its early years, some of which it still suffers under (see: the reserve list).  I look forward to an era of game design that is cognizant of these problems and that masters the art of iterative game design.

*Mana is spelljuice.  Champions use mana to cast spells.

**Magic resist is like armor, but for spells.  It reduces damage from spells.

Optimal Strategies in Games

We’re going to remove ourselves from design a bit this week and talk about how players navigate game systems in order to win.  Before we start, lets establish some definitions.

Resources are game elements that players can accrue and/or lose.  They can be spent or leveraged to affect the game state.  Money in Monopoly, chess pieces, and minerals in Starcraft are all resources.

Objectives are game elements that directly lead to winning the game.  Money in poker, points in sports, and the destruction of your opponent’s structures in League of Legends are all objectives.

The above categories are not mutually exclusive.  Regions in Risk, for example, can be considered resources in that having control of them helps you win the game, but are also objectives in that controlling all of them is how you win.  Money in poker could also be considered to be a resource, as it controls how long you can play.

Game strategies are often built around a particular game objective or resource.  Attrition strategies, for instance, are built around resources that can be stockpiled.  In an Attrition strategy, players attempt to use resources more efficiently than their opponents and thus run them out of resources before they themselves run out.  A particularly pure example of this type of strategy is the “control” archetype in the trading card game Magic: the Gathering.  Control strategies take advantage of a particular resource chokepoint in MTG: players only draw one card per turn.  Thus, control strategies attempt to trade each of their cards for one or more of their opponents cards, in combination with methods that draw more cards.  The end game is an opponent whose resources are exhausted facing the control player with a hand full of cards.

Control/Attrition strategies are often opposed by Aggressive or Aggro strategies.  Aggro strategies race to the objectives, hoping to catch the opponent off guard.  They sacrifice late game resources for early game power.  League of Legends was dominated by a particularly powerful version of this strategy at one point.  League of Legends is a member of the MOBA genre, a strange offshoot of the Real Time Strategy genre in which each player controls a single character from a non-fixed 3rd person perspective.  The objectives in League of Legends are destructible towers with very strong attacks that ward off all aggressors and, behind all the towers, a building called the nexus.  To combat this, the aggro strategy used all five characters on one team, all wielding a spell called “Promote” which created a powerful minion, to quickly push down all the towers in their way and destroy the enemy nexus.  The strategy was strong enough that the Promote spell was removed and later reworked.

As frustrating as it may be to have the game ended prematurely by an Aggro strategy or drawn out into a long resource game by an Attrition strategy, Lock Down strategies have a frustrating side that is uniquely their own.  Lock Down strategies may work in any game where players have the capability to restrict their opponent’s actions or stop them entirely.  Essentially, lock down strategies treat action as a resource they can deny their opponent.  Chess is, in some ways, dominated by this strategy in that the game is only won when the opponent is forced to make a move that will end in their defeat.

Area Control strategies are a subcategory of Lock Down strategies.  In games that take place in two dimensional spaces (in other words, most games that aren’t card games), area control strategies are almost always important.  The entirety of defense in football is area control.

Information Control strategies are an overarching concern in all games with hidden information.  Bluffing, wards in League of Legends, scouting in strategy games, and calling your opponents names in chat are all information control strategies.  These strategies are often deceptively powerful.  Stealth, as it exists in strategy games, exemplifies the power of these strategies.  Strategy game stealth gives you vision of an area without your opponent knowing you have vision of that area.  It also lets you have unexpected area control.  The combination of these factors creates situations in which the only player who really had any agency over the situation was the stealth user, leading to incredibly one-sided fights.

When a strategy becomes dominant in a game community, players can analyze that strategy and discover weakness.  When this happens, players will often develop Metagame strategies which rely on the assumption that the opponent will use the dominant strategy.  The term “metagame” refers to the game beyond the game, where players try to outplan each other before the game itself begins.  These strategies are ultimately a sign of a healthy game community.

I’m sorry for the delay on this post and the lack of images.  I’m still getting a hang on this “blog” thing.

A Brief History of Control Schemes in Games

When I refer to a game’s control scheme, I’m referring to how players interact with the game world.  In the case of traditional games, this has been fairly simple and not a matter of much discussion.  In sports, board games, and card games the control scheme has been simple: players manipulate the pieces of the game as the rules say they are allowed to do.  In sports, this manipulation is the entire game, and is restricted by relatively few rules.  In contrast, card and board games use physical manipulation to update the game state, with the actual action of playing the game being abstract.  The control schemes of sports and board/card games serve as opposite ends of a spectrum of control schemes exemplified in modern games.

Control Spectrum

With the advent of video games, this spectrum has been populated with strange and interesting new hybrid control schemes.  To understand these, it’s important to define the extremes of our spectrum, and the sorts of features common to each end.

Abstract Controls

Abstract controls are used by traditional board and card games.  These control schemes often use turn systems to determine which player gets to play and in what order.  Agricola, mentioned in my very first post, fits solidly on this end of the spectrum and is designed to overcome the issues raised by turn systems.  Turn based RPGs and strategy games also sit squarely on this end of the spectrum.

Direct Controls

Direct controls are used in the purest form in sports.  Soccer, baseball, basketball, and almost anything where players have to run fits squarely on this end of the spectrum.  At the extreme end, players use their entire bodies to alter the game.  Video games never quite reach the purest form of this control scheme, but first person shooters probably come the closest.  Dance Dance Revolution could be here also, when played with the footpad.

These extremes bookend the control spectrum, but they leave out a lot of the middle ground.  Here are some of the more important hybrid control schemes.

Organic Controls

Organic controls try to mimic direct controls without actually giving the player the full use of their own body.  They often involve control of an avatar body along with an interface that induces a feeling of being “in” that body.  There are some strange examples here: racing games in arcades that have a “cockpit”, the webgame QWOP (, and some, but not all, action rpgs.  Shadow of the Colossus probably fits here, especially its grip mechanic, which has the player gripping the control with all their might to stay on the back of their mighty adversaries.  Arguably, shooters fit here rather than in the more extreme category of direct controls.

"Real Time" Controls

After doing all the other categories, I discovered that I’m not really sure where real time strategy games fit.  They certainly aren’t turn-based, and often involve some element of “twitch” or reaction time based gameplay.  An RTS player has to think fast and move fast.  The controls, however, are certainly abstract.  The player doesn’t have an avatar, and all of their attempts to manipulate the gamestate are throttled to a thin stream of orders given to in game “units”.  I’ve decided to put these types of controls somewhere around the middle of the spectrum.

Tags: gaming

Anonymous said: Didn't you say that you were going to update every Friday? Wo ist meine neue post? Vixen

I think I said Sunday.  Speaking of which, it’s going to be somewhat late-ish today.