Game development is a bit of a catch-all term. It includes the work of refining a rules set, balancing all of the game elements, cleaning up issues of presentation, cutting the fat from game systems, and, in general, trying to find the best possible manifestation of the design. To do this, most developers I know just iterate their way to the finish line, preferring brute force testing to theory-crafting. I do this myself. Though I spend a lot of time thinking about design problems from first principles, when it comes to the work of development, I haven't found a good substitute for the exhausting work of making small alterations and then doing rapid play-testing.
The challenge a developer faces reminds me a little of long-distance open water swimming (something I did many years ago when I was more foolish). On the one hand, you've got to keep your head down and your form good. You need to keep pushing along and try not to get exhausted. But it's equally importantly to occasionally look up and make sure you're still pointed in roughly the same direction. You don't want to waste time veering into an unexpected shallows or risk everything by getting swept out by a tide.
Throughout the development process, I try to schedule little check-ins with myself where I take a hard look at the design and try to figure out if the game is actually getting better. Here I'm sad to say that playtesters, valuable in so many other ways, are no substitute for your own taste. It's equally important to remember that, like history itself, the time you spend working on something is no guarantee of improvement. You are often as likely to kill a game as save it.
During this part of the process, I often find myself going back to earlier design diaries and concept documents and making sure that I'm still being true to the core of the game. Sometimes, it turns out that I was wrong in those initial writings about what the game actually was all about. This is one area where design work reminds me of academic research. You can spend a long time arguing and writing and researching only to realize that you were mistaken in some foundational way.
In game design, this kind of idea drift often manifests itself in games that are either overdeveloped (too many systems too closely balanced against each other) or that apply a lot of their weight and complexity towards systems that don't really matter.
One of my favorite ways to orient myself when in the middle of a big development process is to ask myself a couple of questions. The first is simple: what is this game really about? The second is harder: what share of the design is useful for answering that first question? In other words, I try to figure out if the game is distributing the weight of its central question across its structure evenly. By “structure” here I mean every aspect of the game: the design of its combat system, its resource system, the way it handles victory, and the various turns of the game. If the first four turns of the game do very little or nothing at all, why have them? If a combat system doesn't lock cleanly into the victory system, why have it?
This is where Oath gets into a lot of trouble. By design, the game can be about a huge range of things. The game essentially has four victory axes, and, by design, they are engaged with unevenly in each match. In one game, managing reputation will be the central puzzle of the game. In another game, conquest and military campaigns will be critical. Though many of the systems will be engaged with in some minor fashion, it was important that the design by capacious enough to develop a very different character based on the actions of players across many games.
This multiplicity meant that I couldn't always rely on my usual best practices for development. The core of Oath didn't really look like any game I had ever worked on. It was looser, and that was a scary thing. Developers need constraints to do their work well. But, if I placed too many constraints on Oath's development, I would end up with something closer to a typical area majority game and something less like the ambitious history sandbox that I wanted to build in the first place. To get around this problem, I imagined Oath as basically two games. The first of those games would be built around the central victory condition. Because there are for victory conditions, this meant that the first game could be one of four games. The second game within Oath was best thought of as the “spoiler game.” This part of the game was essentially a soup made up of all of the auxiliary victory conditions and game systems that might become important depending on the organic growth of the game world over several games or the happenstance of the draw of a vision card.
Perhaps it's unsurprisingly then that I started listening to a lot of GDC talks about procedural generation and how designers cope with projects that are more game-generators than games. These problems will be familiar to anyone who has worked on a game that features randomized game setups, only because of the scope of the system, the degree of variation in starting setup in Oath is immense. The way around this problem was mostly to develop a few systems of categorization and a strong language for describing different kinds of game states and then making sure that the general game system could tolerate those stranger configurations. There was really nothing fancy about this approach—it was honestly pretty similar to the kind of work I already did with my games. However, Oath was bigger and that size made everything else more difficult.
I don't have time to chronicle every little change that the game has gone through over the past few months. That essay would be much longer than just reading the game's rules! And it would be silly anyway as most of the design has stayed roughly the same. Instead, I'll be writing about a few of the major changes in this iteration and work through some of the reasoning behind this change. Then, at the end, I'll say a few words about what remains to be done with the design and some of the challenges we're still facing.
Chancellor vs. Citizen Dynamic
Outside of the game's adapting meta-game, there is perhaps no game element as central to the idea of Oath as the Chancellor's offer of citizenship. When the game was in its hottest period of development last fall, it was this part of the game that snapped the whole project into focus. Though I knew a lot of the game would change through development, I knew I wanted to preserve this offer.
There were a number of reasons for this. For one, I saw it as a way to mix up the bash-the-leader dynamic of a lot of area control games, including Root. The system was also built as a kind of safety value for some of the misfortune a player might experience in the early game. If someone made some bad choices or had a few exquisitely bad draws then they might be a good candidate for citizenship. Essentially the commonwealth allowed the weak players to form a coalition against the stronger players.
With players eager to explore the system, offers of citizenship flew left-and-right. However, as players gained experience with the system, the offers became rarer. Eventually, they were almost non-existent.
The problem, as summarized expertly by one of my experienced play testers, was that there was little to no room for a win-win deal. All voluntary deals in games (as in life) need to engage with the problem of comparative advantage. Basically, player A needs to have something to give that player B doesn't have easy access to. In my desire to build lots of tactical flexibility into the system, I had created a framework where you never had a reason to deal with another player in this way.
So the first step here was to make it so players could differentiate their player positions more. I did this two ways. First, I did a pretty aggressive audit of the game cards and made them all stronger and tried to build in new liabilities to them as well. Second, I built a new system of offices. There are five different offices that each offer a very good permanent power. Now, during the citizenship offer, the Chancellor can offer a power that both they and the newly-minted citizen will get. Critically these powers are fixed. Even if the citizen gets exiled, both the chancellor and citizen will still keep their power. At first we played with symmetrical powers (that is, both Chancellor and Citizen got the same version of the power) but we found that the negotiations were far more interesting if the powers were randomly coupled. That is, the 5 Chancellor Offices were randomly paired with the 5 Citizen Offices. So, perhaps you want the military advantage power, but if I grant you that power, I will gain a popularity boost. The asymmetry went a long way in creating room for deals.
But it still wasn't quite enough. It took a long time to realize that the basic logic for the citizenship offer was just broken. Basically, any potential citizen was gaining some kind of power but was also losing flexibility while gaining a victory liability. That is, an Exile just needs to do one thing to win. A citizen needs to make sure the commonwealth wins and has secured their secondary prestige condition.
Here I found myself sorting through my notes on John Company and thinking about the importance of risk in negotiation games. A little noise goes a long way. If two players are unsure of the basic value of something when they enter a negotiation, then the range of possible “good deals” can be quite wide. To this, you can add calculations of leverage and, of course, the differences in strategic position that I described above.
To this end I spent a solid month tweaking different prestige systems. Sometimes there were stable prestige points where citizens and the Chancellor could spend various things to get a point here or there. Other times there were little “prestige goal” cards that would get revealed during play by various means. These cards provided players with a huge range of little conversions whereby they could gain prestige points and ensure their line of succession.
Nothing worked, and the basic problem remained. Either citizenship made sense or didn't make sense, and the only good deals occurred when one player failed to realize that he or she was being duped. Not only that, but these systems were increasing the number of rules, the number of things to track, and the number of components that the game would require. Oath was already a massive game. Why in the world did it need to get bigger?!
It was that thinking that led me to the present solution. Oath already had 4 paths to victory, why not just build one of those paths into the citizenship drama. Every Oath got paired with a kind of shadow-Oath called a succession condition. Basically if the Chancellor managed to end the game on top, then you'd refer to the succession condition to determine if a citizen might have unseated the chancellor. This allowed me to dovetail the core uncertainties of the game with the uncertainty around the citizenship offer.
Forcing Players to Play and Sorting through Some Problems with the End
Figuring out the citizenship problem forced me to confront a few related questions that had to do with the more general shape of the game.
For a long time, games of Oath followed a pretty similar pattern. Players would gather their resources, secure a vision, and crouch on their haunches until round 5. Then everyone would spring into action and attempt to secure victory and hope they got lucky on the die roll. I hated this. Basically, players would spend most of their time figuring out how to pull resources out of the game. In other words, good play was basically figuring out how to avoid playing the game. This gave me awful flashbacks to the passing problem in TI:3e (Yssaril Tribes, I see you!).
This dynamic could be seen in just about every game system. For instance, because it was expensive to hold and carry warbands in your cohort, players would just avoid paying the effort penalty by carefully picking up and dropping off their warbands at the proper moment. This mean that only inexperienced players ever had to deal with the problems of supply and initiative that I was hoping would be critical to the game. And, because armies could be mustered instantly, it was often just a lot cheaper to carry around gigantic stacks of favor and cash them in the moment you need them.
These are often the hardest kinds of development problems to solve. Instead of a single problem, the game has a irritating character that can be seen in half a dozen systems at once. This particular quirk likely came from my own tenancy to give players a lot of tactical flexibility. To fix that I needed to rebuild some key systems so that they generated more strategic inertia. I also needed to take a hard look at the victory point system and see if I could spread the drama of the game more evenly across the 5 to 8 turns of the game. In short, I had lots of problems and no single solution was going to cut it.
To fix the first problem, I rebuilt the game's cash and action economy. This actually solved a bunch of related problems. Instead of having the 2/3 action system, I just gave players a stash of action points (called focus) and asked them to spend them however they wanted. The focus scaled based on the number of warbands in play so crafty players couldn't easily dodge the penalty for having a bunch of warbands.
Because players could have 5-6 action points instead of 3 actions, it allowed me to slowdown each players engagement with cards and the game's economy. This meant less draws of 8 cards and less favor dumps. The various resource conversion interactions were grouped under a single action, called Encounter. Encounters basically all work the same way. You put some resources on a card and then you get something in return. If you want to raise warbands, you place a favor on an empty card and then you get two warbands. Critically, this card is now “used” and so players raising big armies are going ot have to go to high population parts of the map and will be more likely to divide their favor among several banks. This in turn broke up some of the common virtuous cycles where players could just build their whole game around a single suit.
I also limited all card draws to 3. This was as much of a downtime fix as it was a game-pacing fix. The limited card draw focused players and created less situations where the right strategy was to dig through the deck as fast as you could.
These changes happened in concert with some critical adjustments to the game's victory system. While I loved the game-ending die, I recognized that it wasn't always producing the tension I wanted in the game. It might have even been the foundational pressure that was pushing Exile's to avoid playing the game and then jump in at the last moment to score a quick victory.
Thankfully, an easy tweak neatly fixed the problem. Instead of having the game-ending roll instantly end the game, it would only end the game IF the chancellor (or a citizen) held the Oath. This stopped the cheesy one-turn wins that the Exiles kept pulling off and gave the Chancellor a nice little buff.
But I still wanted Exiles to be able to win early on the Oath. To this end I pitch the victory point system in favor of a victory system where aspiring exiles simply had to hold the Oath for a couple turns. This gave the game a nice Inis feeling and also eliminated yet another track. All good things!
However, we soon discovered that it was very difficult for an Exile to pull off victory. Holding an Oath for two full turns was a hell-of-a trick. Previously Exiles had managed to pull this off just fine, but such a thing was only possible if the other players didn't feel like they were any threat. By coupling that condition with an insta-win, I had pretty well made the game impossible for Exiles.
In general, I was okay with this difficulty increase. An exile winning on the Oath needs to demonstrate absolute superiority in that victory condition. To me it made sense that a game with a few Exiles passing around the Oath condition would end with a weakened Chancellor continuing their reign. Even if this didn't seem fair from a design perspective, it made perfect sense historically.
But, in the shift away from victory points to holding the Oath token, I had failed to keep the game's victory point scaling system whereby later turns were worth more victory points than earlier turns and folks could catch up a little in the late game. To this end, we built bonuses into holding the Oath and flipping it so that players might gain a bit of inertia for their trouble. I also created a victory priority system for the last turn that would some risky exile strategies time to pay off.
Revamping Popular Support and Rebuilding the Royal Blessing
After retooling some of the game's core systems, I did an audit of the game's four victory paths. Here I tried to pay attention to a few things at once. First, I wanted to make sure the system was generally expressive and had a few different strategic and tactical approaches to each path. Second, I wanted to make sure the game state was readily readable to all players. The less counting the better. Finally, I wanted to make sure the different paths interacted with each other in meaningful ways.
Under such an audit, two victory paths showed clear problems. First was popular support. I liked making this condition a function of the favor in the banks. This was a very expressive and interesting metric that interfaced with a lot of other game systems. But the math around figuring out who had what popularity was a nightmare. I went through half a dozen counting systems, but nothing worked all that well.
Eventually, I went with a solution that seemed almost too simple at first. I pulled the favor banks off the board and made them floating supplies that could be passed around. Now, whenever a player revealed a face up adviser, they would check to see if they had more advisers of that suit than the current bank holder. If they did, then they would take that supply. They wouldn't get to spend freely from the supply, instead they would just keep it in their play area and the favor on it would count towards their popularity.
This created some really interesting dynamics. Players were now pressured to play cards face up so that could get the tiebreaker bonus and take supplies. It also made counting popularity a snap. You hardly ever had to do it as it was immediately obvious who had more or less popularity by just glancing around the table. It also forced me to get rid of any rules for cascading payments since that system was now hardly ever used.
One of the last places where that system was used was in the Royal Blessing auction. I knew that I needed to revisit it anyway. The old system was always a bit of a stopgap, and, as the game developed, it became obvious that it wasn't going to survive. Simply put, it had all of the problems of the old burst muster system but without any payoff.
The first step in drafting a new system was to slow it down. Rather than having a single bid to take the Blessing, I needed something that players could invest in over the course of the game. This would allow players to interrupt one another and would provide players a way to easily see which players were developing their position on that victory axis.
Part of the original problem here was that hording favor was just too good. Players would just accumulate favor and know they could either easily turn it into armies or a big bid on the Royal Blessing.
The new system had to do a few things. While I was comfortable keeping the favor auction as central, the bids needs to be smaller and spread out over the course of the game. I also wanted it to directly interface with at least a couple game systems. To this end I built the current Court system. Basically, players could spend money to send courtiers off to influence the game's culture. They did this by using the Encounter action on site cards (like mustering and the other conversion actions). They would spend an amount of favor and then place one of their war-bands in one of the size court boxes. The more crowded the court box, the more expensive the courtier was. After they did this, if they had the most courtiers among all court boxes, they would get the Blessing (now called the Court's Esteem).
These changes gave this axis of victory a nice geographic texture, whereby players might need to travel to pay off certain cards, which nicely mirrored some of the cultural systems I wanted to model. But I also wanted it to be more than just a favor auction. To this end, I hooked the system to the Darkest Secret. Now, whenever a player took the Darkest Secret, they could place Fashion Token that would either multiply a single court box or cause one to ignored. These markers couldn't be moved by any normal game system so they created a nice variety of situations with some general stability.
But I also wanted the courtiers to count for more than their victory condition. I tried a few different options here but have instead opted to just create a number of card powers that interact with the systems. All of the six edifices attach special meanings to courtiers belonging to that color and there are several card effects in the deck that reward players for investing in the court.
As with most systems in the game, in certain situations this system could be critical or it could be a sideshow. By tying to this cards (especially edifices) my hope is to let the importance and consequences of the court be perhaps more dialed into the game's evolving game state than any other system.
Odds and Ends
I hope that gives you all a pretty nice sense of how the game has shifted. There are a bunch of things I don't have space or time to go into detail about. The game remains in development and there's still room to alter core systems if things are firing correctly. But the game is quickly firming up as well. For the first time, I can clearly see this present iteration making it more-or-less unscathed to the finish line, but we'll see.
The release of the Kit B to the public marks the start of the next phase of the games development. Instead of releasing an abridged kit we have decided to release the full print-and-play kit for the game, including all 198 denizen cards. Some of these cards have been played with a dozen times and others are newer and more experimental. Taken together, they represent a sense of the range of what Oath's design can do. Even after nearly 200 cards, I can say with confidence that there's still a lot of territory to explore.
In the coming weeks, my own time is going to be split three ways. First, I'm going to continue pushing on the development of the card powers along with our editor Josh Yearsley and Nick Brachmann. Second, I'm going to prepare for the next stage of solo design. Right now the solo system is working but it needs some formalization and certain systems are a little out-of-date. I imagine overhauling it will take about a month and I'm hoping to have something to share with you all as a print-and-play by the end of April. Finally, with the help of Josh and our in-house graphic designers (Nick Brachmann and Pati Hyun) we're going to begin to take everything into final layout so we can begin to build our usability studies for once the shelter in place orders are lifted.
For the most part, Oath's timeline has thus far been unaffected by COVID19 and the global pandemic it has caused. This will not remain true much longer. Much of the playtesting has moved to online, thanks in large part to the amazing work of our TTS builder, AgentElrond (who also built the exceptional Pax Pamir mod). However, as we move into final testing and usability, we will need about six weeks of face-to-face testing to make sure we can do the necessary usability work. Some of this can likely be adapted to a digital platform, but I'm holding off serious plans until the end of this month. For now we've got to wait and see.
Thank you again for your support of this project. I'm heartened by your many kind messages and comments here on BGG and elsewhere. Oath is precisely the kind of game that could not exist without your faith, and I'm so excited for you to all explore what we've been building.
- Cole Wehrle
You can find the original post and associated discussion on Board Game Geek.