In Characteristics of Games by Richard Garfield et al., a dynamic which emerges within a large number of multisided games is explicated...

Imagine a game, which we’ll call the “chip-taking game,” where each player starts with a pile of ten chips. Players take turns going around the table. On her turn, a player may take one chip from any player and discard it. The winner is the last person with any chips left. Most people would not enjoy playing this game for long. There is no real skill involved, other than the skill of convincing other people not to take your chips. And even if you possess that skill, once the other players notice you have it, they will probably react by trying to eliminate you first.

Unfortunately, many multiplayer games reduce to the chip-taking game, in the sense that most of their game features are irrelevant for determining the winner, who is instead chosen ultimately in chip-taking fashion. All that’s necessary is that the game be highly interactive, in the sense that players can affect the positions of other players, and also that players can target whoever they affect. Players can simply choose to hurt (“take chips from”) the leader using whatever means the game offers. Even if the leader is highly skilled, he is unlikely to be able to withstand the onslaught of all the other players. Once the leader is eliminated, or at least knocked back from his leading position, the players can attack some new player.

As a simple albeit artificial example, suppose we modify the chip-taking game so that on a player’s turn, she chooses another player and plays a game of chess against him; if she wins he discards two chips, and if she loses he discards only one. This game has all the complexity and skill of chess, but it doesn’t matter. Kasparov is no more likely to win than anyone else at this game, and probably less; the other players are likely to choose him consistently until he’s eliminated.


When players can target other players in an arbitrary way that differentially affects their game states, we refer to this as politics.

Although the above language has a definitive tone, it seems that in some cases, even meeting the aforesaid criteria and even when the ranking of player skill is common knowledge, it could be possible for the most skilled player to avoid suffering politics by manipulating how the probabilities of each opponent winning evolve as players are removed from contention. For example, in a three player game of Magic in which it is common knowledge that Player A is the most skilled followed by Player B and finally Player C, there could in principle be some strategy for Player A such that either or both of Player B and Player C would be punished by aggressing against Player A before one of the other players gets eliminated.

My first thought was some kind of defensive "barrier" strategy, an extreme example being gaining quasi-infinite life. Another was a sort of "landmine" strategy, such as combat tricks which make attacking Player A very dangerous. However, these should only be effective (against game theoretically rational opponents) if Player B and Player C care about who gets second place. If their only goal is to maximize their individual probabilities of getting first place, then even though the barrier or landmine would make Player A the hardest or most dangerous player to aggress against, there is not necessarily any incentive to aggress against easier or less dangerous players first; the barrier or landmine will have to be conquered to get first place either way. (The dynamics appear to get qualitatively more complex by adding a fourth player, i.e., if Player C and Player D aggress against Player A, it may be rational for Player B to aid Player A, preferring to face Player A rather than both Player C and Player D together, since Player B would in the latter case become the next victim of politics.)

Are there any real deck/gameplay strategies in games such as multiplayer Magic which overpower politics?


4 Answers 4


In Magic there are several ways a powerful player can overcome politics, in large part because a player can choose their deck. This makes most sense to analyze in the case of repeat play among a stable playgroup. Here are three ways to do it:

  1. Play a deck that is opponent-count agnostic such as combo. You need to play combos that go "infinite" (such as Exquisite Blood + Vito, Thorn of Dusk Rose) rather than combos that get you to 20 damage very efficiently (such as Channel + Fireball). Here, the skill is in piloting your deck, dodging two player's worth of removal, and setting off your combo before the aggro of two players can kill you.

  2. Play a defensive flavor of control. This involves cards that that make it obnoxious for people to attack you, like Ghostly Prison, Lightmine Field, and Hold the Line. So long as you are not overtly offensively threatening, you can change the calculus for other players attacking you such that it isn't worth it. Your goal here is likely to get the game to go to a standoff that your deck has some way to win (such as a combo or milling opponents).

  3. Play a deck that is worse. Reducing your threat, so long as you can publicly signal that you are doing so, will make two other rational players fight each other first. This has solid game theory foundations. See the Ted Ed Wizard Standoff for a game theoretical example.

  4. Play a deck that leans into politics. There are many cards like Bazaar Trader, Phelddagrif, Zedruu the Greathearted, or even Howling Mine that can let you manipulate the state of the table to change people's priorities. As much as you may be the biggest threat in the abstract, if you can change it so someone else is the biggest threat in the concrete, that can change the focus of player's ire. Credit to @TimC for this point.

  • 1
    I'd add one more (but I don't feel like it should be its own answer) which is to increase the threat posed by other players. A card that causes other players to draw cards durably increases the threat of aggro and combo decks. If I play a Howling Mine, eliminating me won't make the combo player un-draw their cards, so unless they can kill me before anyone else draws too many extra cards, savvy players would target the combo player first instead of me.
    – Tim C
    Commented May 21 at 23:36

I'll focus on the explicit definition of politics in that quotation:

When players can target other players in an arbitrary way that differentially affects their game states, we refer to this as politics.

One option for countering this is to is to constrain players' abilities to target other players. For example, the card Pramikon, Sky Rampart has the abilities:

As Pramikon, Sky Rampart enters the battlefield, choose left or right.

Each player may attack only the nearest opponent in the chosen direction and planeswalkers controlled by that opponent.

Attacking is one of the primary methods players have for targeting each other to try to make them lose. This card and cards like it let you control that aspect of the flow of the entire game.

More generally, you can play cards that take away other players' options for targeting you specifically. This can include cards like Blazing Archon to prevent anyone from attacking you, Leyline of Sanctity to prevent them from targeting you in the sense that the MTG rules mean it, Avacyn, Angel of Hope to prevent them from destroying any of your permanents, etc.

Even more generally, any card that lets you disrupt other players' targeting of you or each other can be a counter to politics. This can include counterspells, removal, and other things.

  • This is a great approach I didn't consider; instead of incentivizing players not to target you, just make it impossible. I still wonder if there's an approach in terms of incentives. Maybe I'll try to work out the necessary and sufficient conditions in more formal terms and ask a more specific question.
    – user10478
    Commented May 21 at 19:01

Diplomacy, I'm told, is the ultimate game of politics that isn't the Great Game (which I don't think is on-topic here). Even "politicsless" (can't talk to other players) variants, the goal is to get your ideas across to your potential allies through your plays such that you have "agreed" on a strategy. Diplomacy is also a long-standing, very healthy game for half a century or more. I'm not very good at it for some reason...

All this to say that politics in games, Garfield's analogy aside, is not necessarily a bad thing, or a thing to try to minimize. This is despite all the questions on this site on "how to stop kingmaking" (aka I've taken you out of contention for winning, how dare you use your remaining power in the game for revenge?), "how to create games where 'gang up on the leader' is not a go-to strategy" (because we know we can't beat the table, but we don't know how to beat the table, just the game), and all the rest.

The winner of a game should be the person who played the game best under the restrictions and freedoms provided by the game. In a game that "suffers from" politics, "politics" is part of the skills required to play the game best and win. As far as I can see, multiplayer M:tG "suffers from" politics in almost all of its incarnations, to the frustration of Spikes who don't consider "politics" a skill.

As others are saying, there are ways to mitigate, but not remove, politics from any multi-contestant game (removing from consideration games where there are >2 players, but only 2 "sides", like Bridge or Princess Ryan's Space Marines). And certainly, that is one of the reasons that "pure skill" players that want the winner to be "the person that showed the most skill that game/day" don't like multiplayer. Especially if they aren't the most skillful in politics.

But, there's even politics in 2-player games, at least at the non-professional level. Consider constructed FNM, or the local club duplicate. If skill is too heavily favoured, it becomes a game of "show up, pay our entry fee, K wins again (or maybe Y, once a month); we're Dead Money; why don't we spend our money at the movies tonight instead?" (or "why don't we invite a few of our fellow dead money over for Commander, or play Fluxx?") Now, maybe that "game" is more for the LGS/club management than the players at the table, but even K and Y end up losing if the game dies. So even there, it behooves the "skill should prevail" people to find some way for others to win sometimes, even if it's not showing up one week, or playing a tier 2 deck and trying to win anyway, or agreeing to mentoring games or bracketed events being held.

  • I think the last paragraph constructs an unrealistically nihilistic model of a new player's capacity for human thought. Many are not purely myopic reflections of their past win rate, but also process factors like how fun the game play seems it will be when they have more experience, whether it feels like their heuristics are improving over time, how satisfying each win and each loss in a game are, etc.
    – user10478
    Commented May 21 at 19:06
  • Folks in a fixed mindset will be more focused on the immediate gratification, while those who adopt a growth mindset will tend to look past. Surely a world in which players self-select among games well-suited to a wide variety of time preferences is better than a world in which all development is constrained to fight over the most myopic possible caricature of a person?
    – user10478
    Commented May 21 at 19:07
  • The choice of letters K and Y for anonymous players comes across as somehow personal.... Commented May 21 at 19:43
  • It was in fact actually random...But yes, I might have come up with a better combination. I feel Alice and Bob (or Angel y Bari) get a bad deal.
    – Mycroft
    Commented May 21 at 19:53
  • @user10478 I have, to my despair, 30 years of experience with this. I'm not talking the "new players"; they're fine. It's the "life novices" or "life Cs" that are the issue. They frankly are the bread and butter of the world - they're as good as they're going to get, but they still enjoy the game, so they show up every week to play. They don't enjoy knowing that if they do well, they'll eventually have to play "the every week winner" again (or even any of the strong players) and get slaughtered, again. But eventually even they get tired of it and go online, or go back into the kitchen.
    – Mycroft
    Commented May 21 at 19:58

If "strategies" for playing a game that ameliorate the "politics" problem Garfield is talking about exist, then the game only partially suffers from the "politics" problem. And indeed, there are multiple game-design techniques to mitigate against "politics", one of which is allowing for strategies that make politics less important. It is not a binary thing -- any given game is not either no politics or all politics.

Garfield basically claims that limiting the ability to target specific players is the easist way to achieve this in any game, but it is not the only way.

Allowing defensive strategies can help, but mostly only for non winner-take-all games (there needs to be a benefit to coming in second). Otherwise it can maybe obscure the politics, but they still crop up.

Another possibility is to give a benefit from eliminating a player, in which case the game often becomes a contest to identify and eliminate the weak(est) players first. Politics can still come into play here, but it is much less stable.

Even if you just think about limiting attack possibilities, the trivial "only attack the player on your left/right" is not the only way -- you can have windows that open and close for attacking specific players.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .