Mid-Game Structures – Player Interaction

Written by Matt Pavlovich

The next stop on our tour of mid-game structures involves player interaction. Of course, player interaction is not strictly limited to the middle part of a game–in some games, every single turn involves some measure of interaction with the other players. But we thought our series about the pivot points and decision crossroads was the perfect place to discuss interaction: the outcome of so many games can hinge on whether you choose to attack your opponent or leave him alone, trade or embargo, call or raise, claim the action that’s best for you or impede on what is best for him.

In the next two sections, I’ll propose a two-dimensional model of interaction that considers both how interactive a game is and how antagonistic it is.

First axis: degree of interaction

thedegreeofinteractionThe first axis of interaction describes the degree (or intensity) of the interaction that exists in the game. I’ll discuss this axis as if it encompasses three distinct categories, but like every other topic in game design, it really exists as more of a continuum.

This dimension spans from games that have literally no interaction with the other players to games where every action depends on someone else’s and in turn affects someone else’s. And like most continuums in game design, most games fall somewhere in the middle.

No interaction “You do your thing, I’ll do mine.”

nointeractionIn games with no interaction, a player can’t affect either the game state or the strategic options/available actions of the other players. Fans of games with extensive player interaction might derisively refer to a game as “multiplayer solitaire” if they perceive it as having too little interaction. But true multiplayer solitaire would literally confine each player to his own corner, doing his thing without paying any attention to what anyone else was doing.

It’s difficult to think of real examples of games with no interaction, so I’ll use a few nontraditional or incomplete examples to get the point across. Something close to a true no-interaction game is a crossword competition, a contest of skill with an objective scoring system that you can definitely win or lose, but where your success and strategy have no bearing on what anyone else is doing.

An episode toward the end of season 4 of the Korean game show The Genius had the final four players simultaneously solve a series of visual/spatial and numeric/logic puzzles. The first player out of the four to solve the puzzle correctly (and grab a token from the middle of the table) would win an advantage for a partner who was simultaneously playing a hand of a poker variant. The poker phase of the game was of course interactive, but the puzzle phase only had the possibility of interaction in the final “grab the token” step.

ubongoPublished strategy games are rarely if ever completely non-interactive. Designers are not exactly incentivized to develop scaling rules for multiple player counts or complex scoring structures if the entire game could boil down to “whoever solves the most sudokus is the winner.” The published game that closest approaches non-interaction might be Ubongo, a strange quasi-African themed puzzle game. Like that episode of The Genius, there is a non-interactive puzzle-solving phase of the game (where the players solve tangram-like puzzles) coupled to a scoring phase (which involves moving along a track and a set-collection mechanic).

Implicit interaction “Different puzzles with the same box of pieces.”

implicitinteractionA mode of interaction particularly favored in European-style games is something I’ll call implicit interaction, where the players can’t affect each other’s game states but can affect their strategies and the options available open to them. In general, the interaction in implicit interaction games originates from denying the other players options and manipulating the opportunity cost of the options that are available. Implicit interactions frequently occur in what we might call the “drafting superfamily” of games, including worker placement, role selection, and hand/deck building with card drafts.

In worker placement games–which we’ve just established as another way of saying action drafting games–a player cannot, for example, force his opponents to place his workers in certain spaces. But it’s certainly possible to claim an action to prevent your opponent from taking it; this “action-blocking” is a central element of the strategy in many worker placement games.

Last WillMore subtly, it’s also possible for one player to make certain actions more or less attractive for his opponents to take. In Last Will, one important method of scoring is manipulating the real estate market; the aim of the game is to spend as much money as possible, so naturally you want to buy property when it’s the most expensive and sell it when it’s worth the least. If I buy the same type of property that you already own, you now have less incentive to devalue it (because that would help me as well).

San JuanSimilarly, in role selection games like San Juan, a player can’t force his opponent to select a certain role, nor can a player directly affect the cards in his opponent’s tableau. A large part of the interaction in San Juan is instead understanding which actions your opponents are most incentivized to take and being able to “piggyback” off of those actions while also accomplishing your own agenda.

7 WondersFinally, in draft-heavy tableau builders, you can obviously prevent your opponents from pursuing a certain strategy by taking a card ahead of them. Or, you can disincentivize players from making certain strategic decisions by monopolizing a particular strategy. In 7 Wonders, if that last gear card is going to score you sixteen points, I might take it instead even though it won’t net me more than three coins. Alternatively, if I already have five military points late in a particular draft round, and you only have one, I’m comfortable passing you a two-military card. Either you’ll take it (which is a suboptimal play for you), someone who can’t interact with me will take it, or it will get discarded.

Direct interaction “Let’s get ready to rumble.”

directinteractionIn direct interaction, the players are able to directly manipulate their opponents’ game states, whether that’s through attacking their territories or stealing their citizens or trading resources. Maybe the most common sort of direct interaction is hostile, as when you rob your opponent in
Settlers of Catan or sweep through a declining empire in Small World. Shooting your opponent in either Coup or Bang! is of course hostile direct interaction.

toledoBut direct interaction can be beneficial to one party or the other. Patronizing an opponent’s business in Toledo might help you if you really need that gemstone, but it definitely helps her to get your cards. Trades in Settlers directly modify both players’ game states, hopefully to both players’ benefit. And a timely donation of the Medic’s abilities to the Dispatcher in Pandemic can mean the difference between a group win and group defeat.

Generally, direct interaction is most common in territory control games (where you’re often attacking or stealing that territory), resource trading games (where all of the players involved in the trade are trying to get the best possible deal), and deck-builder games (though these can span a large range of the first axis from rather non-interactive strategies to incredibly aggressive ones).

Second axis: friendliness of interaction

The second axis describes the nature or “agreeability” of the interaction. In other words, how mutually beneficial–or spiteful–is the interaction? As I’ve just alluded to, not all interaction in strategy games is antagonistic, and some of it is even mutually agreeable. The nature of interaction is a strong function of the type of game, and as we’ll see later in the article, also a function of the “degree of interaction” axis.


Hostility “This town ain’t big enough for the both of us.”

Twilight StruggleHostile interactions are all about sticking it to your opponent, openly and assertively. This is the big missile exchange in Twilight Struggle, the Witch that curses all of the other players in Dominion, the Cyclops wrecking everyone else’s Domain in Elysium. Hostility can be a useful and even expected mechanic, especially in games with war or conflict-related themes, and it is usually (but perhaps not strictly) features in direct-interaction games.

Whether hostility drives satisfaction or dissatisfaction depends mightily on which side of the attack you find yourself on. The best implementations of hostility mechanics increase net satisfaction by making the attacking player feel very powerful and accomplished while making the attacked player feel only slightly penalized. The worst increase net dissatisfaction by making the victim feel demoralized while making the aggressor feel only slightly rewarded.
Hostility might be the most divisive mode of interaction. Some players accept–or even expect–it, while others want nothing to do with hostility in their strategy gaming.

Outmaneuvering “Now you see the genius of my plan.”

Interactions based on outmaneuvering are generally more subtle (but sometimes no less antagonistic) than outright hostility. If hostility is the antagonistic mode of interaction that corresponds to direct interaction games, then outmaneuvering is antagonism for the implicit-interaction set.

Power GridIf I’m playing Power Grid, and I seize on the opportunity to corner the trash market so that I’m the only person who can fire a trash plant, that’s an example of me trying to outmaneuver you. Like hostility, this category contains interactions where one player attempts to advance his strategy at the expense of the other players. However, in the Power Grid example, I’m not laying waste to your power plants or stealing your resources.

orleansAnother good example from a recent favorite is in Orléans: if I know I have the first action selection in a given round, and I see that there is only one monk left in the common supply, it’s a great idea for me to prioritize activating the Monastery. Again, I’m not stealing your monks or invading your monastery, but an important part of my strategy is making sure I, not you, get to add the monk to my bag.

Simple Competition “Bingo!”

In simple competition scenarios, I’m trying to complete a task faster (or more effectively or more efficiently, etc.) than you. Simple competition is associated with low- to no-interaction games where there isn’t much opportunity to attack (or help) another player’s strategy. Think of simple competition like a bingo game, or like the crossword puzzle tournament I mentioned earlier: the players have basically no way to affect each other’s game state, but it’s still competitive in that all of the players are trying to accomplish their goal and win.

BetrayalYou could even envision an actual game of multiplayer solitaire as an expression of simple competition. And some phases of some games constitute simple competition, like building up your own character in Betrayal at House on the Hill before the Haunt starts. But as with games in the “no interaction” category, there are few examples of published, commercial games available that rely strictly on simple competition.

Marriage of Convenience “The enemy of my enemy is my friend.”

One of the first articles that I wrote for Games Precipice covered the topic of positional balance, and a concept that I described in detail was something I called the “tripod effect.” The tripod effect is a player-enforced (rather than rules-enforced) means of balance where players who are about to lose naturally gang up on the player who is about to win.

RiskIn brief, it’s why games of Risk take so long when there are three players left: the players in the second- and third-strongest positions are incentivized to work together to take down the player who has the most territory. But eventually, one of those two is going to take over as the strongest player, and the other is going to team up with the former leader. (In Risk in particular, this cycle repeats itself until one of the three breaks through with an opportune card draw.)

mall of horrorThe important thing about the marriage of convenience is that it’s a temporary-at-best coalition motivated not out of making an agreement with my de facto partner but out of desperation to stay alive in the game. If Alex and I are playing Mall of Horror [The backstabbing precursor to City of Horror] with our friend Jeff, and he’s about to win at the end of the turn, then the optimal move for Alex and I is to team up to throw Jeff to the zombies. I’m not doing this because I want to improve Alex’s position in the game but because I want to at least give myself a chance to come back and win.

A marriage of convenience has elements of both implicit interaction and direct interaction. In the Mall of Horror example, I’m not modifying Alex’s characters or altering his game state, simply entering into a brief agreement with him. If implicit interaction is all about manipulating opportunity costs, then I’m buying myself some more time in the game at the cost of also buying one of my opponents some more time. But the part where we agree to sacrifice one of Jeff’s characters in order to buy that time is clearly a direct modification of Jeff’s game state–one that he’s not likely to forget (even years later).

Mutualism “I’ll scratch your back if you scratch mine.”

Compared with the marriage of convenience, mutualism represents a friendlier take on cooperative interaction. Mutualism is an agreement between two (or more) players where each player thinks he’s getting something good out of the deal.

BohnanzaThe most obvious expression of mutualism is in trade-heavy games like Bohnanza. Let’s say I flip a Red Bean that I want no part of because I would have to rip up my lovely Soy Bean field. My opponent kindly offers to take it off my hands with the promise that she’ll send her next Soy Bean my way. Of course, that’s a great deal for her: she gets a relatively valuable asset for free (one that’s even more valuable to her because she already owns one!). But I’m willing to make the deal because it removes a lot of headache for me.

Essentially, it’s another opportunity cost calculation: I’m saving myself from damaging my position in the short term (and getting the tentative promise of extra resources in the future) at the cost of substantially boosting my opponent’s position now. But in the end, deals like this improve both players’ positions, which is why both are willing to make the deal.

TongiakiEven more fascinating examples of mutualism occur with multiplayer agreements. Tongiaki, which has a criminally underrated social component, requires players to work together to chart courses on the high seas. The variance, and the excitement, comes in that nobody knows ahead of time how many different players must team up to survive the voyage. On some level, all of the players have to scratch each others’ backs if anyone wants to score any points.

SpeculationStock holding games like Acquire, Imperial and Speculation have a unique tendency of developing shared interests. Speculation is a great example of multiplayer mutualism. If three of us own shares in Eketorp (one of eight stocks to trade in the game), the three of us are suddenly aligned to a common objective. I might choose to shepherd Eketorp’s rise through the stock market, even knowing that it also helps the other two players’ positions, because I know they’re likely to do the same. Speculation does a great job of letting players feel like they’re making big strategic moves–thereby increasing net satisfaction–even though they’re not necessarily doing much to gain a competitive advantage over each other.

Alliance “We’re all in this together.”

PandemicFinally, a full-blown alliance is the friendliest sort of interaction in strategy games. Here, you’re literally on the same team as the other players, and you win or lose as a team. Maybe the most obvious example is in cooperative games like Pandemic, where the players all have independent game states but are working together to eradicate diseases. An example of a non-collaborative game that relies on alliances is Lord of the Rings Risk, where the players controlling Gondor and Rohan are allied against Sauron and Saruman. Gondor and Rohan work independently, and the players are free to pursue their own strategies, though they need to coordinate their moves to prevail as a team.

DiplomacyAlliance is not always encoded in the rules of a game. In games with “last man standing” player elimination mechanics, it’s completely possible for two or more players to enter into a verbal alliance and agree on a joint victory (maybe once one of the legs in the Risk tripod has snapped, or if you and one other oligarch have established an unbreakable Franco-Russian alliance in Diplomacy).

One danger in designing wholly collaborative or alliance-driven games is that particularly dominant players can impose their strategy on the entire game. Although alliance is one of the most intricate forms of interaction, a single loud or forceful player can easily turn a strategy game into a completely non-interactive spectator sport.

A case study: Cosmic Encounter

Cosmic EncounterThe reason for Cosmic Encounter’s perennial popularity is its intricate interaction mechanics. Most of Cosmic Encounter’s interaction is direct interaction, which occurs when you invade another planet or join in an attack or defense or agree to a mutually beneficial deal to share a planet. But its interaction can also be more implicit, when you fortify your planets enough to dissuade your opponents from invading them, or when you don’t join in an attack in an attempt to extend an olive branch to another player.

Invading an opponent’s planet is clearly a hostile interaction, while timing those invasions correctly so that you aren’t perceived as a threat until it’s too late is the more subtle outmaneuvering. At some points in the game, all you want to do is beef up your own hand of cards, which might be a form of simple competition. Once one player has her ships on other planets, it’s almost certain that the rest of the players will enter a marriage of convenience, at least for long enough to knock her down a peg or two. If two players have had generally peaceful and forthright negotiations, they may decide that both playing Negotiate cards is a reasonable mutualistic decision. And if that happens when both players already have a presence on four planets, they may decide to close out the joint win in an alliance.

The case of Cosmic Encounter nicely illustrates that games are not limited to one degree or one style of interaction. The best games can offer their players wide varieties of interaction types to choose from and have mechanics that support all of them.

Mapping the degree/friendliness space

Click to see our Patreon CampaignYou might have noticed as I was describing the axes that they’re not entirely independent. Hostility as I’ve defined it is necessarily a direct interaction, because you’re affecting your opponent’s game state when you conquer his territory or steal his resources. Most mutualistic interactions are direct interactions, too, especially those that involve trades.

Similarly, outmaneuvering is almost always associated with implicit, rather than direct, interaction. Often, if I’m bothering to outmaneuver you rather than attack you outright, it’s because the mechanics of the game lack the means for me to make that attack. Marriages of convenience can be either direct or implicit interactions, depending on whether they’re based on attacking the guy who’s about to win or simply denying him access to strategic opportunities.

And what I’ve described as “simple competition” is the controlling mode of “interaction” in a game only when the game is essentially non-interactive anyway.

As tempting–and possibly informative–as it would be to try to plot games, or even genres, in the degree of interaction/friendliness of interaction space, the example of Cosmic Encounter illustrates how difficult that would be. Single games can span large swaths of the space by themselves, and there is enough variation within any given genre, that trying to assign games to points in this space would be basically impossible. The take-home message here is that greater interactivity usually leads to more possible modes of interaction.

Bringing it all together: the moment of first contact

I’d like to end the article by bringing things full circle and going back to why we think player interaction is such a natural fit for mid-game structures: the idea of a change in the nature of a game’s interaction as a mid-game turning point. The first example I’ll use is Blokus, a simple territory-control abstract with some of the most fascinating player interaction dynamics of any abstract out there.
BlokusIn Blokus, all of the players start in their own corners and are forced to expand outward; an important decision early-on is which direction to build toward, as encroaching on someone else’s territory is the closest that the game gets to hostility. Actually encountering another player is a two-edged sword that limits the amount of space you have to build but also gives you intriguing new foundations for your expanding territory. A surprising amount of mutualism can arise during these encounters, whether intentional or not, if you and your opponent both leave voids perfectly suited to the other person’s tile.

La CittaLa Citta is another territory game whose dynamics change considerably when the players begin to encounter each other. To begin, each player’s territory comprises a few small cities, but as each city adds buildings, its physical footprint grows, which can bring cities into conflict. The primary interaction mechanic in La Citta is the sphere of influence that each city exerts: citizens in my city may defect to yours if you have better educational facilities or houses of worship. As I expand my city, I have an important choice to build towards yours (which brings us into conflict) or away from yours (to avoid conflict).

In this way, the mid-game of La Citta marks a transition from implicit interaction based on outmaneuvering (for example, by strategically selecting cards from the common pool) to direct, hostile interaction (when we can steal each other’s citizens). Of course, the possibility of marriages of convenience is always open; when one player is about to build a commanding lead in terms of the number of citizens, the other players can purposely build toward that player in an attempt to lure some citizens away.


Although players often interact throughout an entire game, the mid-game often uniquely couples interaction with strategy: a player’s decision to be friendly or antagonistic–or to stay out of the fray entirely–has repercussions throughout the rest of the game. Two not-quite-independent axes describe most of the variance in player interaction: the amount of interaction, and the friendliness of that interaction. While a true “map” of games or game genres on these axes is nearly impossible to construct, the moral of the story is that more interactive games can give rise to more different modes of interaction. And the best games not only offer many different modes of interaction, they allow the players to interact in different ways as the game progresses.


8 comments on “Mid-Game Structures – Player Interaction

  1. E

    I just found your website. You have given me lots to think about.

  2. Chuck Durfee

    A game that immediately came to mind in the little to no interaction category is Firefly: the Board Game. In the base game, players can do little to help or hinder their opponents. They can theoretically try to meet up and trade, though it’s usually easier to get resources in other ways. Or, they can move obstacles that travel much more slowly than their ships.

    1. Alex Harkey

      Great example Chuck, I haven’t played it yet but your description is certainly indicative of the no interaction category. Very interesting!

  3. geoffreygreer

    Nice analysis, as usual, friends. I would like to simply remark that sometimes non-hostile interaction, such as outmaneuvering, can end up becoming hostile. I’m thinking particularly of our experiences with Ticket to Ride. As scarcity of routes becomes as issue, the race to claim certain routes becomes very competitive, and after you’ve been blocked two, three, or four times by your opponent (sometimes accidentally, and sometimes not), it gets pretty frustrating and not a little tense around the table. ;)

    1. Alex Harkey

      That’s so very true, Geoff. There can occasionally be a separation between the intentions of an aggressor and how they might be perceived. Even if it’s not intentional to block you or I, say, one player lays down track in TTR which sparks reactionary placements by other players – being blocked out collectively by the remainder of the table can make us feel targeted. So we might feel hostility, yet everyone else was just engaging in simple competition for their own best interests.

      Being on the frustrating side of those inadvertent player interactions is one of the subjects we dove into in player control, a subset of player interaction.

  4. Nate Scheidler

    My friends and I often lump Ascension: Chronicle of the Godslayer into the multiplayer solitaire category. There are limited interactions in the base game, mostly restricted to slaying monsters that come out randomly with different effects. You can influence the player to your left to a limited degree by virtue of what you remove from the center of the board, but you have substantially decreasing influence after that point. Its not true “multiplayer solitaire”, of course, but it fits most of the criteria for the discerning player and as a result we rarely played in any format other than 2 player, where you have the greatest influence over each other.

  5. Shao-Ying Chen

    Great article! I’m a boardgame gamer and designer in Taiwan, and I’d like to translate this article into Chinese to facilitate the understanding of boardgame designing principles for boardgame designers in Taiwan. Of course, the original author and the location of this article will be subscribed in the beginning of the translation article. May I have the pleasure to translate this article and publish on the Facebook fan page about boardgame designing and testing? Thanks!

Contribute to the conversation, leave us your thoughts and feedback