A very early Oath the Board Game prototype, laid out on a table with players engaged

Designer/Developer Diary, Oath -

Oath | Designer Diary 5: Cards and Continuity

In the first diary I posted, I mentioned the centrality of cards to the design of Oath. Most of my games feature cares prominently, but none are anything like Oath.

This is a big subject, and I will be talking more about the cards for probably the next half dozen pieces I write about this game. Part of the problem with talking about the cards in Oath is that they do a lot of different things. They are timers and modifiers and abilities and assets and events and even full fledged conspiracies. The card roster in this game is over twice as large as the one I used in Pax Pamir and, by certain metrics, three times as complex. At the same time, the individual cards are simpler.

The trouble with talking about the cards is that they modify a pretty dense knot of interconnected systems. One thing tends to lead to another. Normally I'd just start by talking through those systems. But, those systems won't make any sense until I explain the basic movement of cards in the game and their general use. Both of those elements are, in turn, responses to the demands Oath presented me. I didn't come to this design with any particular mechanism in mind. Instead, I wanted the game to be able to tell certain kinds of stories, and every game system was built from the ground up to help make those things possible. In short, those broad intentions created a steep set of demands for the game's mechanisms. I can't think of an element of the game that illustrates this better than how the game uses cards.

Let's start at the beginning.

As with much of my thinking about design, things first got serious back in 2012, the year both Pax Porfiriana and Great Zimbabwe crashed into my table. Both games shocked me in a few ways. First, the designs had a tremendous amount of interaction and parasitism in their play. They were also elegant distillation of previous systems (the Lords games in Porfiriana and the classic Splotter logistics games for TGZ). Each game also challenged what I thought about what a card could do. In both games the cards that were played transformed the space of the game. Only a small portion of them would be seen each game, and the character of the game could change dramatically depending on which cards happened to be played.

The key, I think, is that the cards in both games sat on top of very expressive systems. I use the word “expressive” hear because I think it's sometimes useful to think about a game's design as a kind of language, complete with a vocabulary and rules of syntax. Some languages lend themselves to the expression of certain ideas. Whereas in other languages, it might be difficult to describe a particular concept. Games are like this too. Some games have no trouble portraying a mob of disaffected strikers. Other games might require a half page of errata to render the same idea within their gameplay.

When I was building the cards for Oath, I knew early on that I needed the core system of the game to be incredibly expressive. To this end, I built a list of 20 or so “things” that I wanted to be made possible with the play or use of a card or one of the game's core systems: a palace coup, a surprisingly discovery, an economic depression, an uprising led by a charismatic messiah. As I cycled through the many iterations of Oath, there would come a time when I would attempt to speak the game's language and see how the idea of those imaginary cards might sound. If I found myself stumbling through a wordy effect with lots of little rules and exceptions, I knew the system I was building was not providing me with the right vocabulary. Even of the interaction was otherwise interesting (and many were), I threw it in the trash bin. The expressiveness of the system was far more important any any cute mechanism I might have built.

It should surprise no one that the design I settled on featured multi-use cards. The basic design is somewhat similar to the cards in Root, which each card having, essentially, an effect and a suit. The suits represented different cultural aspects of a society and its neighbors. Here I should credit Kyle Ferrin with coming up with the first list of these aspects and helping me workshop what would become the game's list of six suits: Order, Hearth, Discord, Arcane, Beast, and Nomad (those may not be final names but they are good stand-ins for now). Each of these suits have a general mood which maps loosely on to one or two of the games different systems. I'm not a Magic the Gathering player but I think my logic here is probably quite similar to how they think about the different deck archetypes.

Oath preliminary suit icons for Order, Hearth, Discord, Arcane, Beast, and Nomad

This expressive quality was necessary because the game's deck of cards is its primary and most textured memory system. Here's how it works. Depending on the course of a game, new cards will be drafted into the deck, taken from a vast library of around 200 cards. Basically, if the player that won the game had relied chiefly on cards belonging to the Arcane suit, the next set of Arcane cards would be added to the deck. At the same time, other cards from the game's discard piles would be taken out of rotation, joining the ranks of the dispossessed. After even just a couple games, each copy of the game's deck will diverge in strange ways, creating new opportunities that adapted to each history the players were creating together. There's a lot going on with this card list and the dynamics of each suit which I'll talk about in a later designer diary. For now, just know that basically each game players collectively draft the deck for the next game.

Okay, let's do a little arithmetic here. I wanted each suit to have four packets, representing the entrenchment (and/or perversion) of that cultural aspect. With four packets of cards for each suit, there was enough drift that the decks could get truly strange. Players could also cycle dispossession cards back into circulation so there's no real end-point here. That's 6 (suits) multiplied by 8 (cards in each packet) multiplied by 4 packets for 192 cards total. Add the 50 or so starting cards and we get to around 250 unique cards in the card list. Yikes.

If the size of that card list was one problem, I quickly realized that there was, in fact, a much bigger problem. I wanted the history to be a little lossy in Oath. What I mean by that is that I wanted the ramifications of one particular game to sometimes be subtle. Perhaps you introduced a card into play that wouldn't be seen for several games. Then, many games later, perhaps a literal relic of that game might resurface.

Pax games have this because they only use half of the cards in play in the course of the game. But, if the game deck in Oath was only around 50 cards, in most game systems every card would get seen and played. Heck, in Root it's not uncommon for players to cycle through the deck twice. That is a fine dynamic but it wouldn't allow me to have the game-to-game variance I wanted to achieve with Oath.

Cards from Pax Porfiriana; Note from Cole: The only card I had left at the end of yesterday's really mean Pax session.  For a brief overlapping point in time both Sonora and Chihuahua had declared sovereignty.
You might play Pax Porfiriana dozens of times before Sonora ever has a chance to declare independence!

I needed the entire game of Oath to take place using only 20 cards. That is a very small number for a game that also wanted to advertise an exceptionally large card base. At the same time, when I thought about the scope of the game's storytelling, this was the proper number. With only 20 cards in play and an essentially an 12% deck turn over from game to game, players could develop meaningful relationships with the many of the game's characters and dynamics. Things could still change, but they would change at a pace that felt fair to the players.

This design requirement created many headaches. Neither Root or Pax Pamir would work with so small a card base. With Root, the cards are essentially a resource. In Pamir, market velocity and game pacing demands a pretty stable (and large) card base of about 50 cards per game—and each of those cards is going to get seen and considered by every player.

Around this time I was going through some very old drafts of Pamir and stumbled upon an idea that I had discarded way back in 2013 or 2014 while working on Pamir: the shared tableau. Originally in Pamir, there were no player tableaus. Instead there was just a tableau for each coalition that players could add cards to. The cards would then be essentially controlled by two players: the coalition at large and the player who happened to own the card. It was a cool idea but was also goofy and didn't really work within the context of Pamir.

This system was particularly well-appointed to Oath because a shared tableau simply required fewer cards than having large individual tableaus as in something like Race for the Galaxy. It also resonated well with my hope to provide players lots of room to share infrastructure and space ride on each other's coattails.

The basic idea was quite straightforward. Most cards would get played to the world of the game, where they would sit on locations. Whoever ruled that location would essentially have those cards in their tableau. But, critically, players could also visit locations and use those cards as if they ruled them. In addition, players would have a very small personal tableau that would supplement the shared tableau. I'll go into the deeper dynamics of this system in a later entry, but hopefully you get the general idea. Basically, the cards were part of the game's geography. And, it so happened that having about ~10 cards publicly held nicely corresponded to the comfort level of players in terms of how much information they had to hold in their heads.

A very early Oath prototype, laid out on a table with players engaged

The cards themselves did a lot of different things. There are essentially six types of cards (not to be confused with the six suits). A card might modify a core rule of the game or provide players with a particular combat advantage or a new action they could take. Heck, some even provide players new ways of scoring points. The power of the cards is deeply contextual. Because most of the cards are going to be on a shared tableau which all of the players will be able to fight over, I privileged card powers that were expressive and interesting over a card list that was strictly balanced. This was a hard won lesson from my time working on Vast: the Mysterious Manor and Root: the Exiles and Partisans Deck. In the first Root deck, I did my best to make sure the cards themselves wouldn't unbalance the game. I think this sometimes made the deck a little less interesting than it might have otherwise been. The Exiles and Partisans deck is far goofier and, I think, far more interesting strategically. I'll be going into all six types over the next few weeks as I begin to describe some of the game's core systems and how they interact with the card base.

With the basic framework set and a sense of the scale and power of the cards in mind, I was nearly ready to start building a proof of concept game. However, first I needed a card flow. I knew early on that I didn't want to rely on a Pax style markets. These markets present players with 12 or so cards and are priced by the column with the cheaper cards on the left and the more expensive cards on the right. As cards are bought, the market shifts down so the neglected cards get cheaper and cheaper. This market is a wondrous thing. But, it's also a nightmare for new players because they have to confront 12 cards right at the start of the game. It can be dizzying, even for experienced players. Most of my games of Pamir start with everyone starring at the initial market for a solid five minutes before we dig into the game.

The Market, an array of cards as an example from Pax Pamir

Pax markets also require a market velocity that is powered in part by a steady stream of discards and one-offs. For this reason they demand pretty big decks. I didn't have that option.

At first, I thought I would make an extensive use of a discard piles. Players would draw a number of cards from the deck, pick one to use, and then throw the rest into a discard pile. I always liked the painful choices players confronted on a draw in Race for the Galaxy and I wanted to do my best to recreate that tension. Critically, the discard pile could also be drawn from. If players draw from that discard pile, any cards they discarded from that draw would then be discarded into a new, second discard pile and so on. This essentially kept the cards in circulation. But the system lacked a geographic anchor and pretty soon there were half a dozen discard piles. What's more, players could discard cards and then immediately fish through their discards to get more cards. This undercut the opportunity cost of picking a card. If you wanted two of the four cards you drew, you could just fish in the discard card with another action.

Around this time the map for the game was clicking into shape and I realized that the game was probably going to have three regions, each composed of some number of sites that the players would be moving through. With three regions, I now had a designation I could use. I created a basic discard flow that worked like this. If you drew cards in the central region of the map, you would toss your discards into the discard pile of the next region out. If you were all the way out in the Hinterland, any of your discards would get tossed to the central region's discard pile, forming a zero-sum loop. These discard piles could be drawn from, like any other deck, but all discarded cards would follow the general discard flow.

A diagram of movement from Cradle discard to Provinces discard to Hinterland discard, back to Cradle discard

This brought me to another card flow question. If the game deck was 50 cards, why would someone opt to go digging in a discard pile when they could dig into the game deck. I needed to design some kind of friction that would discourage players from digging more than half way through the deck. I did this by mixing in five special cards into the top half of the game deck. Depending on how many of these cards had been drawn, players would have to pay an action penalty for digging into the deck. The regional discard piles however would always be exempt. In practice this means that in the early game players will put 10 to 20 cards into circulation and the cost of adding more cards will be sufficiently prohibitive that most players will be content to hunt through discard piles.

This friction essentially gave the game something like a zero-sum card economy. Cards don't really exit the system. Instead they get kicked around from one place to the next. Whereas Pax style games confront their players with a cascade of new cards, Oath offers essentially a swirling tidal pools of reoccurring opportunities and old acquaintances that shifts from one game to the next depending on the choices players make.

The five special cards that enabled this circulation also allowed me to hook the card play of the game into the game's victory system and how that victory system adapts to the way a game is won or lost. I'll be talking about that more next time.


- Cole Wehrle


See the full discussion on the original post on Board Game Geek.

Back to Blog