Strategic Insights: Connection Games
by Shannon Appelcline
Early in 2003 I wrote a series of articles on strategic board & card games, with somewhat of an eye as to how those genres of games could be used in online play as well. Since then, every once in a while I return to the topic, as an interest strikes me.
Over the last couple of weeks, my wife & I printed out and put together the "Lyon Extension" supplement for the popular Ticket to Ride board game. The process (which involved lots of printing, cutting, and glueing) has gotten me thinking a lot about the general category of games, which I call "connection games", and so I've decided to provide a broad overview of the category of gameplay, as I have of others in the past (most recently TT&T #144, Strategic Insights: Auction Games).
Defining Connection Games
The basic concept behind a connection game is pretty simple. You have a number of "connection points" on a board which can represent cities, stations, or abstract grid points. Then you have some cost that must be paid to connect up multiple connection points.
If you remember back to my original discussion of game components back in TT&T #102, Designing Strategy: The Components you'll see that connection games are a great example of a very controlled "evolving environment". You don't get to change the whole board, but at least in games with some sort of connection blocking you can change the landscape of the board in limited but meaningful ways.
Common variants of connection games include train games, wherein you're connecting cities with railroad, and pipe games, wherein you're matching up the sides of tiles in order to make connection (though this latter type of game often goes beyond the traditional definition of connection games because it can include other types of gameplay).
The first question in a connection game is: How are the connections formed?
Freeform Connections: In this type of game players freely choose how to layout connections on some type of grid. Most railroad games (Empire Builder, TransAmerica, Age of Steam, Stephenson's Rocket) use this method.
Limited Connections: Instead, some games specifically limit where connections can be made. They tend to lay out specific routes connecting specific points on the board. This is the form used by almost all transient connection games (Power Grid, Medieval Merchant) and some railroad games which choose to make that limitation an important part of gameplay (Ticket to Ride, Union Pacific).
(Beyond this I'll also comment that every limited connection game I've ever played assumes that all connections are point-to-point; theoretically you could design a game based instead on three- or four-point connections, though how useful this would actually be in actual play is a good question.)
The next question in a connection game is: how much does a connection cost?
Turn Cost: Some games only have a cost in "turns", typically meaning that you can play one tile/connection/etc. per turn. Knots and Metro are good examples of this type of game, since you put down one tile a turn, trying to connect up your connection points (your sides of the board in Knots, your stations in Metro) in ways that are beneficial to you.
Linear Cost: Very similar are games that have a linear cost for playing connections. However, instead of the cost being turns, it's instead money or some other marker resource in the game. I'm actually not immediately aware of any games that use this method, as once they get down to this level of simplicity, they just go with turn cost instead.
Variable Cost: Variable cost connection games make connection costs dependent upon some factor of the landscape; like the earlier categories, that cost is them paid by some limited resource (usually money, but sometimes a fraction of a turn). TransAmerica is the simplest example that I'm aware of for this category of gameplay: mountains, rivers and other tough regions cost a full turn to connect while regular regions cost half-a-turn to connect. Empire Builder is another railroad game with more complex variable connection costs: from 1 for plains to 2 for forests and mountains to 3 for swamps to 5 for alpines.
Games with limited connections just tend to mark a specific cost for a specific route rather than players having to figure it out. Ticket to Ride does it with spaces for trains, while Power Grid and Medieval Merchant do it with routes having listed values on them.
Formulaic Cost: Finally, some games with variable costs take an additional step and make their costs formulaic. For example in Ticket to Ride you must pay the cost in cards of a specific color, while in Union Pacific you must pay with tracks of a specific type. For some reason formulaic cost connection games almost always tend to be "set collection" centered (as Ticket to Ride is) rather than using more complex formulas (e.g., 2 of one thing and 3 of another). The one exception I'm aware of is Attika, where you play unique hexes, each of which has a specific formula for play from a number of different resources.
Once you've made a connection, the question is: how long does it last?
Permanent Connection: The connection lasts for the rest of the game, though in same cases might be removed by some disaster. Almost all rail games follow this model.
Long-Lived Connection: The connection is not transient, but it slowly decays until it has to be rebuilt. An easy method would be for any connection to last X turns, where X was a flat number [e.g., 10] or a derived number [e.g., based on the length of the route]. I'm not aware of any games that use this method.
Transient Connection: A connection is transient. You pay a one-time cost to use it (which you often won't have to repeat because you now have a presence at the new connection point), and afterward you have no ownership of the connection. The non-railroad connection games, such as Medieval Merchant and Power Grid tend to use this method.
(Parenthetically, the latter method of connection usage shows how connection games can start to shade over into more traditional games where you're moving tokens around the board, as a tranient connection cost isn't that different from a movement cost from one location to another.)
How coherent your connection is answers the question: how much does your connection block that of other players?
Total Blocking: In these connection games players are usually taking control of squares or hexes rather than simple connection lines. In any case, there's no way to get through a connection than going all the way around it. Examples of this include: Twixt, which is a rare game where you built thin connections, but other people can't cut through at your vertices, and Attika, where you're building connections of adjoining hexes.
Route Blocking: In games with limited connections there's usually only space for one, sometimes two, players on one of those connections; once they're there that specific route is blocked. Ticket to Ride and Union Pacific both use this method.
Edge Blocking: In these games, connections tend to be built on a grid; you can block other players from taking a specific edge on that grid, but other players can build right through you at vertices. Empire Builder is the classic game on a triangular grid, while La Strada is an example of a game that's otherwise total blocking, but players can build through each other at cities (which are essentially special vertices).
Point Blocking: Instead of blocking connections, some games instead allow blocking of the connection points. This is usually the case with transient connection games such as Medieval Merchant and Power Grid, each of which allows a limited amount of building in each city (from as low as 1 player to as high as 8 structures).
No Blocking: Finally, it's possible that you could have a connection game where players don't block each other at all, though I'm not aware of any.
So, you've got a great network of connections, the last question is: what are your connections used for?
In general I think there are five major answers:
Networking: The connections are used to form a network, which is itself valuable in some way. Usually, players are trying to get specific connection points connected, either because that's their explicit victory condition (TransAmerica or Ticket to Ride) or because specific connection points give them specific things they want or need (more on that in second).
Connection Point Usage: The last element suggests our next answer: the connections give access to special powers at certain places. In Power Grid you're trying to get to cities because you want to supply them with electricity and in Medieval Merchant those cities can give you victory points through market dominance. In Empire Builder connection points provides you with specific goods.
Limitation: An omnipresent use of a connection grid is simply limitation, as suggested by Connection Coherence, above. Most games make this a part of the game. Medieval Merchant and Power Grid are both examples of games where taking control of connection points can cut players off from parts of the board, while Ticket to Ride and Union Pacific do the same with control of connections.
Valuation: In some games the connection itself provides the valuation/scoring in the game. For example in Metro you get points according to the length of the connection leaving each of your subway stations and in Knots you win by managing to connect your two halves of the board. Sometimes the valuation is more obscure as in Union Pacific, where you add connections to increase the value of train companies, then play stock to try and take over those train companies. (This method of connection usage is more frequent in pipe games, which tend to be more "pure" connection games, without the overhead of stocks, delivery, and whatever else that you find in train games.)
Transport: Lastly, connections are used for delivery.
In railroad games this is often marker delivery: you pick up a marker worth some specific value (it's usually a goods marker) and then you use your network to move those markers to other places on the board. (To a certain extent this is just a combination of Networking and Connection Point Usage.)
It could also be a token delivery as is suggested in this very clever La Strada variant where you move majority-control markers around in limited ways using your network of tracks.
Making a Good Connection Game
With all that said, what makes a good connection game? Unlike auction games, and some of the other sorts of gameplay that I've explored in the past, I don't think there are easy ways to break or make connection games. Instead, there's a lot of variety within the genre, and thus a lot of different ways to go.
However, I would offer one rule for making good connections:
Some Variety is Required: A connection game could be really simple: you try and connect point A to point B. It could also be really boring. Instead, the best connection games offer a few different game systems to interrelate with the core connectivity, increasing possibilities for strategy and tactics.
Ticket to Ride is a good example. It introduces elements of hard decisions (as you must either take the cards you need to build track or else use your cards to build track each turn) and brinkmanship (as you could lose either cards or track if you don't take them). It also makes the game a bit more interesting by having a couple of methods for scoring (by laying long tracks; by successfully connecting certain cities; and by having the longest contiguous track).
Knots on the other hand, doesn't offer a lot of variety, and I think that hurts the gameplay. You have two choices each turn: to lay a piece to extend your own connection network (which you're trying to build from one side of the board to the other) or to lay a piece to limit your opponent's, and each turn is also sharply constrained by what you drew and what spaces there are left on the small grid.
A Few Examples
Here's a few examples of how these game elements all work in actual play.
Ticket to Ride: In this railroad game by Alan R. Moon you're trying to connect together American cities in order to earn points for long routes and also to complete your "destination tickets", each of which lists two cities.
The game is built around limited, formulaic, permanent connections, wherein you play a set of like cards to take a route which you'll control for the rest of the game. As you'd expect with limited connections, there is route blocking.
The main uses of your connections are connectivity, since you're trying to create cross-country routes and blocking, as you can sometimes block other players from completing their tickets. In addition there is some element of valuation, as all connections you build are worth points, and the longest are worth a lot.
Power Grid: This is partially an auction game, but you're also paying connection costs to get from city to city, essentially building a "power grid" which you'll use to supply those cities with electricity.
This game is built around limited, variable, transient connections, wherein you pay costs (which are laid out on the board) to get from one city to another. As with most transient connection games, there's point blocking: only 1, 2, or 3 people can get into each city, depending on the stage of the game.
The main point of the connectivity is connection point use as you need cities to supply with power, though there's limited use of networking, wherein you move toward other cities that you want to supply, and hopefully are cheap to build toward, and limitation, wherein you can block a city to prevent an opponent from getting to the connection points on the other side.
(And for those who remember my auction article I'd call the auction half of the game: an unconstrained English auction with open, turn-based continuous bids; it's single payer, single winner with no particular limitations. In other words, it's a pretty vanilla auction system that mostly supplements the connection-based gameplay. There's also a logistic/resource system in Power Gride, creating a triad of gameplay systems, of which connectivity is just one part, which of course refers back to my advice that "variety is required".)
Empire Builder: The Mayfair "crayon rail" games are pretty well known because they allow players to lay out their own track on a grid (by drawing them with crayons). Using our language of connectivity it uses freeform, variable, permanent connections with edge blocking.
When looking at the connections' use, we find networking, connection point usage, and delivery as you need to pick up goods at certain cities, then travel along your network and deliver them to other cities. Finally, there's extremely limited limitation, as you can sometimes force a player to take a slightly less desirable route because you've already taken the best edges.
There are many railroad games, most of them of American design, and I haven't tried to play them all. Notable games which I haven't played that could have offered their own examples to this article are the 18xx series of games, Age of Steam, and Stephenson's Rocket.
I think I've played everything else that I've mentioned herein. I've also written reviews for some, including: Attika [4/4], Empire Builder [3/4], La Strada [3/4], Lunar Rails [2/3], Steam Tunnel [3/3], Ticket to Ride [5/5], the Ticket to Ride Lyon Extension [5/5], TransAmerica [5/3], and Union Pacific [4/4].
Connection games are a well-established type of strategy games that allow for a lot of different types of play. Most games to date have been railroad games, but there's a lot more room for expansion in the genre. Beyond that connection-based gameplay can be used to supplement lots of other types of gameplay, as shown by Power Grid (and many other games).