Hey all,

The focus over the summer was implementing the final mechanical changes on the economic side of the game. Well, prior to that task I’d spent time laying down the basic infrastructure for the diplomacy system, but it was still a ways from being functional. I knew I was onto something, but it would take more time to hook all the different pieces together, especially since there was basically no actual game content yet (e.g. the various interactions that can take place). This month we’ll be picking that feature back one up last time and talking about the now fully implemented – though still in need of polish – diplomacy system.

And phew, has it been a big task. I’ve spoken about it a bit before on Twitter and elsewhere, but this has been easily the toughest game development challenge I’ve tackled in my career. First, on the design end, there’s the question of what “good diplomacy” even looks like. After all, even representing how things work in the real world might not be any fun. And we haven’t seen many good examples of strategy game diplomacy being universally praised, so there’s not even a particularly good frame of artificial reference to start from. Furthermore, the job gets even tougher when you then need to build a system which is both powerful and flexible enough to accommodate a design which will require a huge amount of real-time iteration. This is the kind of feature you need to be able to tinker with a lot, since the initial numbers you plug definitely won’t add up to a fun experience. You can get there, but you’re gonna have to do some grinding just to see how things are playing out, let alone identify what to adjust. There’s a reason why “great” diplomacy is basically unheard of – in a sense I like to think of it as the ‘final boss’ of game design.

 

The Basics

I’ll start with a quick review of the basics: there are few different kinds of diplomatic stats (Relationship Level, Influence, and global Reputation) which add up to a Leverage score that determines whether another Leader is willing to do something or not. RL is simply a measurement of how much the other leader likes you. Influence is a “currency” usually earned alongside RL in equal amounts that can be cashed in once in order to temporarily increase your Leverage. Reputation is the same concept as Influence, except it can be spent with any leader, and earning it requires some kind of great or bold act (such as, say, insulting Attila!).

…

In terms of design, the idea is to use these basic building blocks within a larger system where the focus is on building and managing relationships with other leaders. Leaders will naturally compete with one another due to personality and a bit of randomness, and in so doing will often put you in the middle of their dispute. The choices you make will have long-term, visible impacts. You’ll also be able to make interesting choices, such as insulting one leader in order to gain Reputation points to spend with another.

Rather than trying to make AI leaders behave like humans the idea is to lean on the system as a strategy game feature and also inject an interesting personality/storytelling element into our unpredictable, asymmetric world. We’re never going to get players to see these computer opponents as human, but we can definitely present a compelling argument that they’re interesting characters. Speaking of which…

 

Leader Traits

Leader Traits are things that mostly work under the hood, shaping the decisions that leaders make (although also occasionally changing how much RL or Reputation or whatnot is earned/lost after specific interactions). This makes it very easy to hook fun, predictable behaviors into the system right as I’m adding interactions to the XML. Being able to do both simultaneously both saves time and makes the implementation a lot more cohesive. Another cool side-effect is that this will also make it super-easy for modders to not only customize leaders but even make their own crazy creations above and beyond what’s included in the core game, and with very little effort. Whether those combinations will be fun to play against is a job for them though!

As I’ve been adding new types of interactions I’ve been thinking about things like, “okay, so how would a ‘competitive’ leader behave in this situation?” Because of the cool infrastructure I established a while ago it’s easy to add direct modifiers which push leader behavior in very clear directions. One of the biggest issues with diplomacy in other games is that the leaders tend to all seem generic, random, or (gulp) both. The system I’ve put together for AtG should solve this. That said, everything isn’t perfect yet and it’s going to take some iteration to get things right, but there’s a strong foundation to work from, in large part thanks to the traits system.

 

First Meetings

The biggest new “bullet point” AtG offers on the diplomacy front is that when you first meet a Leader there’s now a special sequence which helps define what the future relationship will look like.

Upon first contact a Leader will now give you a gift, make a demand, or simply say hello. You then get an opportunity to respond, and the choice you make will affect your relationship going forward. If the other leader gives you a gift, you will be expected to reciprocate, otherwise it will actually hurt your relationship (although you will, of course, get to keep the goods!).

Even if a leader simply says hello you still have a choice to make because it’s possible to give a fairly cheap gift and earn +1 RL and +1 Influence. This is a one-time opportunity though, so if you don’t take advantage of it the turn you meet someone that’s it – you’ll have to find some other way to befriend them. A particularly fun example of this mechanic in action is when you not only refuse a demand from a ‘Haughty’ leader but then escalate things even further and then tell them to shove it. This basically makes a permanent enemy out of the other leader, but in return you earn +1 Reputation which you can spend as Leverage with other leaders. Very cool!

The rough percentage breakdown of leaders choosing each of the three FM options is 20/20/60, gift/demand/hello. I like these numbers because usually (60% of the time) the interaction is simple, but also almost half the time something interesting will happen, good or bad. I can’t say for sure yet until I’ve playtested things thoroughly but this feels about the right frequency, as we don’t want it to swing too far to either extreme. If 75% of your neighbors all demand your lunch money the game is going to start getting frustrating pretty quickly!

So, why add this system at all?

First, it means there’s always something important for you to decide upon meeting another leader, as even if the they simply say hello you yourself have an opportunity to make a small sacrifice and kick the relationship off on a good foot. When you’re a small fish in a big pond this can be incredibly useful and satisfying. Second, it helps provide a sense for who this new leader is and what they’re all about, rather than doing the usual thing and making a big deal out of meeting a big fancy 3D leader… who you probably won’t talk to again for another few hours. Sometimes you even need to play multiple games before you start picking up on what differentiates the leaders, and starting off each relationship with a FM sequence completely avoids that problem – when you spot Attila next door you’re going to quickly recognize that means something important! Few previous strategy games have offered significant and early opportunities for leader behavior to shape the game, and I think we can do better.

 

Diplomacy & Asymmetry

The diplomacy system really starts to shine when combined with the AtG’s asymmetrical approach to faction design. Basically, the world will start pre-populated with factions in various states of development. You’re not going to be competing against a bunch of equals all trying to achieve the same goal, as in nearly all other 4X games – this is a world your tribe was just born into, and your situation and goals will differ from everyone else’s.

As for what that means diplomacy-wise, in each game some factions will be extremely strong, and so you really have to pay attention to whether or not they like you. This helps ensure AtG won’t be a game where you can simply ignore diplomacy completely and expect everything to turn out alright. I’ve focused more on quality than quantity, so the game doesn’t have a trade table or lots of little trade options, but if the approach works out well I see it being a real springboard to doing a lot more interesting things down the road. I’m confident the focus on personality, relationships, and meaningful consequences is the right way to go, but we’re breaking some new ground here so the jury is still out.

Diplomacy is a feature that takes some time to fully digest, so I don’t expect this to be one of the bullet points that really stands out to more casual players, but I do expect and hope that as people dig deeper and play more they’ll really see the advantages that this new approach brings to the table. The system does still need work though, and as I’ve noted already I’m going to need to spend some time playtesting and tweaking the numbers. Thankfully, that’s a whole lot easier than designing and building the system in the first place, so I can safely say this side of the game is now in a really good spot.

…

That’s it for the October update. Thanks again for following along, and I’ll be back in early November with an update on the last few tasks left before we’re finally ready to ship this thing!

– Jon