Design Practicum: How scale affects dynamics in MCDM's Super Codenames

Here's something weird. The RPG designers over at MCDM like to relax by playing Codenames together. Recently, they experimented with a super-sized game of Codenames, four times the size of a regular game. You can watch MCDM play Codenames here and "Super Codenames" here. 

For "Super Codenames," MCDM bought four copies of the game, in order to have enough agent cards. Instead of a five by five array of cards, the 'board' is ten by ten. Each five by five quadrant coincides to a regular key card. Two of the quadrants are blue-bordered keys and two are red, resulting in an even number of red and blue words on the board.

Some rules had to change. Because of the even number of clues, the team going second picked one of their cards to cover before the game started. Due to the increased opportunity to pick an assassin, a team would not lose unless they hit two assassin spaces. Additionally, MCDM normally plays with house rules that don't allow for note-taking of any kind, however that is untenable for Super Codenames. On their first play-through, the spymasters took notes; on the second, everyone did. 

Playing Codenames at this scale brought about a number of changes in dynamics. Spymasters very quickly changed clue-giving strategies to worry less about avoiding bystanders and even the other teams cards in favor of covering as many of their cards as they could. This was a valid strategy because each team started with over thirty cards on the table. This unconcern with wrong answers resulted in a loss of tension that a regular-sized game has. Even the assassin, who now required two hits to be deadly, was no longer a significant source of tension. 

The sheer number of cards also made the game feel more fair. A regular game always has the chance that one team has a whole bunch of easily relatable items and the other team does not. Playing four sets of cards at once evens out the chances of relatable items, especially since clues could span across the boards. A tangible result of this was that there were multiple instances of clue words given for six or seven cards, something incredibly difficult in a normal game. For less impressive clues, the clues were more straightforward, with fewer logical leaps. As field agents became used to more straightforward clues, they began to miss clues they would have had no trouble connecting in a regular game of Codenames (such as connecting heart to attack). 

The drawback to the number of cards (and clues given over the course of the game) was information overload for both spymasters and field agents. This was reduced somewhat by note-taking. However, gameplay shifted considerably with the introduction of note-taking (especially in game 2) towards a logic puzzle rather than a competitive game. The loss of competitive feeling became most evident when players from both teams, at times, would cooperatively brainstorm about what a clue might mean. 

MCDM plays with a meta-dynamic they call "signal jamming," which is where members of the opposing team try to persuade the guessing team to guess an answer that they believe is wrong.  It adds a delightful level of subterfuge to the game, getting more complex as a player might suggest an answer they believe is correct in order to steer a team away from it. Signal jamming is completely lost in Super Codenames. The field agents are so overwhelmed with figuring out their own clues that they don't have the mental space to figure out opponents clues AND steer them in the wrong direction. 

Another meta-dynamic MCDM plays with is the "unconnectable clue" guess. In the rules of Codenames, teams can always guess the clue number plus one. However, if there isn't a leftover clue to guess, most players will pass to the other team. Some of the MCDM players have made an art trying to get inside the spymaster's head and guess which word on the table was so different from the words they guessed correctly on that turn that it must be one of their words. This move has worked at least once, but shows up more often as a strategy to consider before discarding for safer plays. In Super Codenames, the unconnectable clue, like signal jamming, is lost. 

In fact, Super Codenames has such different dynamics that it started to build its own meta after only two plays. In set up, the four sets of clues were divided by a cross of black ribbon on the table to help the spymasters more easily see which quadrants they were looking at. However, because the field agents knew how many of their cards might be in each quadrant (8 or 9), they were able to eliminate quadrants late in the game, significantly narrowing down the number of cards they had to consider for a clue. 

I was not expecting how the gameplay shifted while still maintaining the basic rules. Super Codenames is a harder game than Codenames because of the amount of information needed to be processed by all the players. Super Codenames is an easier game because the connections are simpler to make and the tension is lower around possible mistakes. There is more room for expert clues (clue:0 and clue:infinite) and coming back from behind. Super Codenames is a much, much longer game. The game length became overly long from a viable product standpoint. But the added length gave more room for meta-guesses within a single play-through based on table talk and previous guesses. However, if a team were to fall behind significantly, the game could drag on for over an hour before the chance of a reset. 

In Principles of Design #7: Scale and Proportion, we talked about scale being the actual size of objects. For Super Codenames, the object that increases in size is the 'board'. The bigger an object, such as an oil painting, the more room for additional information to be presented to the viewer. (The Garden of Earthly Delights is twenty times larger than the Mona Lisa.) Making the 'board' larger increased play time, cognitive load, and drastically changed dynamics. By changing the scale, we see how the addition of more information in the form of components can both simplify and complicate a game, make a game both easier and harder, and make a game feel both more balanced (greater fairness in clue giving) and less balanced (an overall messier game with greater possibility of an uncatchable leader). The shift in dynamics was solely because of the increase of information presented to players as a result of a change in scale. 

Design Practicum: Dissecting Quacks of Quedlinburg

Unlike our look at Roam, I am not going to focus on design principles for Quacks of Quedlinburg. Instead, I am going to 'beat out' the game to examine the emotional impact of each action. 

When actors beat out a script, they are looking for goals and obstacles. The interaction between goals and obstacles creates emotion. So, instead of trying to describe the emotion of the players, we first must be aware of the goals and obstacles. Let's look at an overview of the set up and first round of the game.

Set up: Recipes are mini goals that if reached might achieve the main goal of winning the game. Variable recipes mean that the game is harder to develop a universal strategy for.

Fortune teller deck: Almost fifty percent of the deck contains effects that are not immediate, but become goals for the next round. The obstacle is that players must meet a certain condition to receive the reward.

Potion phase: The goal is to pull chits out of a bag in a way that scores the most points. One obstacle is that the only way to effect this is by buying chits in the evaluation phase to manage the probability of pulls. Another obstacle is the busting mechanic, making when a player stops pulling from their bag part of the central tension of the game.

Evaluation phase: Players see how well their goals were met and plan for future rounds. The goal is to manage probability and the obstacle is that players are limited to purchasing two chits, of different colors, per round. 

Here we see where player actions and decisions come up against the restrictions of the rules. This builds tension and creates a pay off for effort at the end of the round. However, Quacks relies heavily on mechanics that are luck based to the point that player goals don't even come into play. These mechanics also influence the player experience, even though the players have little control over them.

Uncertainty and surprise are built into the fabric of the game from set up. My husband and I have been playing thru the recipe sets, but even though we know we will be playing set 4 next, we don't know what that means strategically and won't until we play it. That variability gives a sense of excitement from before the game starts, as the players learn the version of the game they get to play that play-through.  Quacks chooses to spread out when certain recipes are available. The purple and yellow recipes show up just in time to be pleasant surprises of new possible mechanics. A large portion of the mechanics are luck based with the central mechanism being push-your-luck. There are three randomizers in the game: a deck, a die, and bag-building. This high number of randomizers fuels the feeling of uncertainty and lack of control. 

The central emotion of Quacks is one of constant expectation of reward. Just reading the recipes allows players to imagine getting the rewards in the near future, which kicks off the game's relentless reward mechanic. The first action that happens after set up is a random bonus to one or more players from the fortune teller deck. This sets the tone of randomized rewards for the rest of the game. The next thing that happens is the potion phase: a series of possible rewards in the form of chits pulled from your bag. Not every chit is a reward, and the reward value varies depending on the order the chits are pulled out. As players come to the end of the potion phase, they begin to calculate the risk to reward chance of pushing further in order to come out ahead of the other players or reach a certain milestone. Importantly, busting does not remove all rewards from the potion phase; it merely limits which rewards you can have. After the potion phase, the lead player gets an additional reward in the form of a die roll. In a more strategic game, it would be ludicrous to give the player doing the best such a strong bonus. Here, the die roll is just one more reward players are competing for. The rest of the round is spent tallying the rewards received and upgrading in the hope of future rewards. Rounds 2 and 3 add diversity to possible rewards with the addition of the yellow and purple recipes. 

A simple overview of the gameplay might look like this:

Set up: First, players are uncertain what recipes they will play. Then, players get an overview of future rewards as they learn the recipes they will be using.

Fortune teller deck: Before drawing a card, players anticipate what rewards will be available that round. After drawing, players are either rewarded or have a new goal for the round. 

Potion phase: Anticipation builds with each chit pull. Some chit combos create exciting moments of reward. As players come closer to busting, they also push for end of round rewards. 

Evaluation phase: Lead player receives a random reward. Players tally rewards and upgrade for the next round. 

The rewards come in many different flavors. Victory points are the most straight forward reward and determine the winner of the game. Rubies and coins are rewards that can be turned into upgrades in the form of refilling the flask (an item that allows a do-over in the potion phase), moving the droplet (a direct upgrade to the player board), buying chits (an upgrade to the player's bag), and conversion to victory points. The fortune teller deck and the bonus die both contain a variety of rewards. The powers of the chits determined by the recipe books reward players with additional progress on their player board and occasionally in other ways. Each chit pulled also impacts the chance of busting, so every pull that does not increase that chance feels like a reward. 

Rewards are exciting and surprising in Quacks. The uncertainty of the rewards builds excitement for the expectation that players will eventually be rewarded. The frequency of rewards fuels emotional investment in the game mechanics. The unevenness of reward distribution builds tension. The lack of player control and high luck ratio keeps players in the desired emotional state where more strategic choices would actually undermine the emotional content of the game. 

The rhythm of the game adds to the overall excitement level and reinforces the the feeling of randomness. For instance, the simultaneous play during the potion phase introduces a level of chaos on top of the push-your-luck mechanism that moves the mechanism away from a feeling of high stakes tension. Instead, players feel swept along in the chaos, only returning to a feeling of tension at the end of the phase as players begin to stop pulling. The complexity and rewards ramp up nicely from round to round as players have more and more possible chits to pull from their bags. My first impression of the game was probably most strongly influenced by the rhythm. To me, Quacks feels like a pebble kicked down a hill that grows into an avalanche, and the players are merely trying to stay out in front of it or get buried. That feeling comes from the uncertainty, lack of total control, and the chaotic, building rhythm of the gameplay. 

I should mention that none of this emotional content comes from the theme. The theming is somewhat creative, the art is nice, the layout of the boards is nicer, but nothing about the actions of the game feels like the story paragraph found in the rules. In fact, if the bags were the cauldron and players were randomly ladling potions out that would feel more thematic. In my ideal world, the chits would be 3D bubble-shaped and total potion composition would have an end of round effect. There are implied customers, but we never see them. My point is not that the lack of theme makes the game weaker (that's up to your personal taste), but that this game's emotional content exists in spite of the lack of strong theming. I wouldn't call Quacks an abstract game, although it sort of is. The luck and variable powers in Quacks seem to put it in a different category, even allowing for a broad definition of abstract games. (I don't recognize "pasted-on theme" as a game category, either.) I would describe Quacks as an experience-based game, specifically an emotion-driven game, meaning that the central experience of the game is the emotion(s) players feel while playing. Specifically, the feeling of uncertain reward that steadily but chaotically ramps up throughout the game. 

The Seven Categories of Actions in Board Games

I have workshopped this idea on Facebook, so I know exactly how controversial it is. I believe all game actions can be distilled down to seven categories, which are largely component and mechanic agnostic. That said, this may be the most mechanic-heavy post I ever write. 

What are actions? 
Actions are player-driven changes to a play-state that occur once a game has begun and before a game ends. Thus, I am not considering set-up or final scoring when thinking about action categories. Actions are governed by rules. Anything that occurs outside of the rules is not an action. Actions are usually organized into turns, rounds, and/or phases. Turns are not synonymous with actions. Examples of actions are move up to 2 spaces, attack, collect resources, etc. Actions may be a blending of multiple action categories. Actions can be organized by a number of different methods. Most commonly, actions are organized by mechanisms or components that are used. My action categories are abstracted beyond mechanisms and components. Action categories look at the goals of a (successful) action. What is this action trying to accomplish? 

Action Categories

1. Acquisition- Actions that result in games elements being claimed exclusively by a single player are ‘acquisition’ actions. Acquired elements may be either secret information, private information, or public information. Example: drawing cards face-up or face-down.

2. Deployment- Actions that move game elements from a hidden or inactive status to a visible, active status are ‘deployment’ actions. Deployed elements maybe played onto a personal play-space or a shared play-space. Example: playing cards into a tableau.

3. Spatial Adjustment- Actions that change the physical location of already deployed game elements are ‘spatial adjustment’ actions. Physical location includes height, such as in a stack. Adjusted pieces may ‘belong’ to the player taking the action, another player, or be a part of the shared play-state. Example: moving tokens around a track.

4. Value Adjustment- Actions that change the assigned value of a game element are ‘value adjustment’ actions. Value adjustments are most often numeric. Value adjustments can occur to a player's elements, their opponent's elements, or shared elements. Adjustments to a player's elements tend to increase value while adjustments to opponents elements tend to decrease value. Commonly used to mitigate luck. Example: attack damage. 

4a. Value Generation (Geller corollary)- Actions that set the value of game elements that were previously null are ‘value generation’ actions. Value generation is most commonly luck based, with the frequent use of dice. This is largely a sub-set of value adjustment. Example: dealing cards randomly to auction board spaces that have set values.

5. Social Leveraging- Actions that change the perceived value of game elements are ‘social leveraging’ actions. Changes in perceived value do not change the absolute value of the game element. This category is the subjective version of Value Adjustment. Social leveraging nearly always includes discussing the play-state with other players, with the goal of getting other players to act the way you want them to. In order to fall into the 'Social Leveraging' action category, the actions must be a part of the game as described in the rules. Examples: negotiating, bluffing/lying, alliances. 

6. Resource Generation- Actions that create game elements as a part of game play are ‘resource generation’ actions. Resource generation actions are most common in party games. Examples: drawing, acting/miming, giving clues, story-telling. 

7. Null Actions (Hanson corollary)- Actions that leave the play-space unchanged are ‘null actions’. The most common example is the option to pass on a player’s turn. 

What about failure?

Not all actions succeed. Actions do not have to succeed to qualify for their category. Attempting an ‘acquisition’ action and failing still qualifies the action as an acquisition action. In fact, many mechanisms are built around the high potential for failure. Auction and dexterity games have high levels of acquisition failure for highly desired pieces and lower levels of failure for less desired pieces. All social leveraging actions have a high expected failure rate, as those actions rely on the subjective ability to persuade other players. Spatial adjustment has a middling failure rate. Players can be out-maneuvered and blocked from moving into a space, but will stop having fun if this happens too often. Deployment has a low failure rate, because not being able to activate pieces is not fun and slows game tempo. Value adjustment and value generation also have low failures due to frustration factors. Instead, players are usually offered mitigation and retaliation options rather than the ability to cancel a player’s turn. Resource generation, by its nature, has zero failure rate. Provided all players are participating, no matter what quality of element they create, elements created will be used in play. Whether or not those elements are effective generally falls into other categories. Null actions that occur frequently would be considered a failure by the designer, because not acting is not fun. The actions themselves, by definition, always succeed. 

Using Action Categories

As stated, action categories look at the goals of an action. I use action categories as a diagnostic tool.  When a design needs a mechanism, or needs a replacement mechanism, I look at what the goal of the mechanism should be then examine the various mechanisms that best reach that goal while fitting the constraints of the overall game. I imagine most designers think this way to a point, but I will literally think, "This way of acquiring cards doesn't work. What is another acquisition action I can put here instead?" The reason why the categories don't necessarily feel balanced against each other (spatial adjustment may seem like a subset of deployment to some, for example) and why many actual game actions are blends of the categories (dice combat includes value generation via dice and value adjustment to a player's damage track) is because these are tools of thinking about actions that help me determine a path forward in design. Other designers use other categories. Isaac Shalev lists game subsystems as resource acquisition, information, conversion, conflict initiation, conflict resolution, scoring, end-game triggers, and markets. We can see there is a certain amount of crossover between our lists, but clearly we have different goals when using our respective diagnostic tools. Where my list is more concerned with how components shift throughout gameplay, his is more concerned with the underlying information those components represent. 

I'm not sure how useful action categories are for other designers. But if you do want to use them, here are some things to remember. Mechanisms can fall into a number of categories or more than one category at once depending on how they are implemented. (Components also fall into various categories depending on how they are used.) Worker placement involves a player deploying a worker to acquire a resource. Dice worker placement involves value generation before placing and frequently value adjustment to mitigate luck. Some worker placement games have rule that you can move a worker that has already been placed, a form of spatial adjustment. So there isn't much point to listing every mechanism that can apply to each category, so much as simply understanding what goals the categories describe. If I am designing a hand management game, I need a mechanism for players to acquire cards. By understanding that the goal is acquisition, I can evaluate mechanisms- existing and those I invent- by how well they meet the goal of acquisition of cards in a game that is more about managing the cards after I have them. So, simply drawing cards may not give enough choice, but pick-and-pass drafting may pull too much focus from primary goal of hand management. In which case, I would continue to evaluate other acquisition actions until I found one that works in the context of the game.

Most often, I use the first four categories when designing. The remainder came out of a brainstorming session that existed partially on Facebook (thanks to Eric Geller and Cody Hanson for their suggestions). Whatever system you use, find a way to think critically about your design at a granular level and you will become a better designer.