Designing a game for a wide number of players (e.g. 2-6)

I too find it difficult to design a game for a wide range of players. I don't think you should force your design for a number of players. If your design works great for 3-5 players, then don't come up with some convoluted variant to make it work with 2 or 6 players.

When I start thinking about a new design I always think about how a different number of players affects the game. Often it affects game length and number of components. For example, when you have a card game with a deck of 110 cards, where the game ends when the draw pile is exhausted, and each player draws one card per turn, a player will have less turns in a 5 player game than in a 2-player game. So, perhaps you should remove cards when playing with less than 5 players. On the other hand, it may be important for the game that all cards are drawn, which means that each player has to have more turns with less than 5 players.

It also affects less tangible things such as amount of chaos and downtime. These are harder to predict before playtesting the game. Usually, the more players you have the shorter individual turns have to be to keep the downtime bareable. Preferably, players should be able to performs some of the actions simultaneously when playing with 6 or more players, or the turns have to be really, really short and punchy. To make sure that there isn't too much chaos you want the impact of each individual turn on the game state to be fairly minimal with more players.

Designing for two players is very different than designing a multiplayer game. Different rules apply. You don't have to worry about kingmaking and bash-the-leaders effects, in a 2 player game, for example. It's OK to have player elimination. It's easier to design a two player game with a "race" goal.

However, some mechanics just don't work with 2 players. The idea of majorities doesn't work well with two players. Auction mechanics are a lot less fun, and you can't count on a bit of player interaction to cover up small inbalances in your game system. Two player games are often far more unforgiving. The trickiest of all, however, is that actions that affect the opponent negatively are often far more lucrative in a 2 player game than in a multiplayer game. This may result in a very "negative" game which may not be much fun.

I've run into this issue a few times. The Marble Game can be played with 4-6 players, but it's likely that at least one or two of a group this large will feel "shut out" of the game early and will likely not have a good experience with the game. Plus the area that you can play in shrinks from the whole board to one sector. The answer I came up with was to have people play in teams with shared goals.

Warp 6 is played with 2-3 players. I thought that it'd be good to suggest that people play with 4 or more players with the addition of more dice. Then I tried it. The game completely broke when that happened. Planning became impossible, and the game devolved into whoever was lucky enough to have his turn at a good time.

I'm working on a dexterity game with this issue. It works well with 2, but it'd be good to have versions that included more people. My solutions for this involve having slightly different goals for a 3-4 player version, and for an even larger version.

I'm also working on a factory game that plays 3-6 people. I don't feel scalability is that big an issue here. The big issue is I need to distribute goods for money, and do I feel like ripping off the auction mechanics of New England, Amun Re, or Evo?

Great points in this thread, all around.

Nuclear War (and Escalation and also Profliferation) is a great game that scales well from 2 to lots of players. In fact, the more players, the more I enjoy Nuclear War.

So why does Nuclear War scale so well? It's multi-player and it's very easy to attack each other, which could easily lead to king-making and bash-the-leader, but Nuclear War nicely avoids these problems, because there is a lack of information about who is actually the leader.

The winner is the last person with any population left, and sometimes nobody wins, thanks to retaliatory strikes by ousted players. Since a given player's current population is known only to that player, people can only guess who is most likely to win, based on assertions, diplomacy, tells (poker players can sometimes guess their opponents current emotional state based on minor behavioral quirks, known as "tells"), or even by how much change they need when cashing out a big population card to pay for the damage they just took from a bomb dropped on their country.

I start with a set of rules and mechanics that cater to a 3 player game. I find 3 is essential as it provides the perpetual tension of 2 adversaries (making game strategy decisions not neccessarily black/white, but shades of gray: what may hurt this player may help the other). Typically from here it's not too hard to add a 4th. Adding a 5th/6th might takes sound value changing of various mechanics: you may make them permanent or just per-play based on the number of players playing.

Then I take a step back and see how you can 'dumbify' the gameplay enough to make it work with only 2. I think 50% of the time you'll find it's just not possible w/o destroying the core enjoyment of the game (depending on it's lean towards action/strategy). Games like Ticket to Ride, for example, all it took was an extra addendum to make bi-tri-ways only routed once. Carcassonne took nothing. But games like Settlers of Catan really can't accomodate it because of the economic basis of trading (my brother likes to play it 2-player but he wonders why the game becomes a colonial-era 'road race' everytime!)

A couple of games that scale well are Power grid (2-6) and Carcassone (2-5/6). I've played quite a bit of 2 player power grid, and we usually get some good auctions. They play quite differently with differnt numbers. In 2 player hunters and gathers it's not worth much to complete an opponants forest, however, with 5 players that extra turn becomes much more valuable.

One approach to the problem of scaling from two to five (or six) players is to increase the game space and decrease random effects for greater numbers of players. This can decrease the aid (or damage) that any one player can cause toward any one other player. I'm having trouble putting what I mean into words, so maybe that's not very clear. I experimented with the concept I'm suggesting in Short Changed by simultaneously increasing the number of bits in play and decreasing the number of unknown / undiscoverable bits with an increasing number of players. With two players, there are fewer chips in the game, but a larger proportion of them are out of play for each hand. The opposite is true with six players, with more chips, but none of them left out of play. Both of these factors are scaled linearly for numbers of players between two and six, and it seems to work out OK for that simple game.

Another approach to this problem, one with which I've been experimenting in a hex and counter / consim game (some of you may have played it at Protospiel 2006) is to start with a game that, by default, supports the maximum number of players you're designing for. Then, either automate the play of any unused players (for example, Vanished Planet) or combine them so that some individual players may controls two or more of the factions (for example, Here I Stand).

This is an enlightening discussion ;-) A lot of viewpoints expressed give me that "duh, why didn't I think of it that way?" vibe. It's funny how people approach logic problems (and I dare say board game creation is one of the most boggling of logic problems) in many different fashions.