Share this project

Done

Share this project

Done
Lead a barbarian kingdom to glory during the final days of the Roman Empire. An empire builder by Jon Shafer, designer of Civ 5.
Lead a barbarian kingdom to glory during the final days of the Roman Empire. An empire builder by Jon Shafer, designer of Civ 5.
3,009 backers pledged $106,283 to help bring this project to life.

March 2018 Update - Diplomacy

22 likes

Hey all, in this update we'll be taking a quick tour of the design for the new diplomacy system.

In last month's post I talked a lot about high-level design goals, but another important consideration is how things actually play out from the player's perspective, especially with an abstract system like diplomacy. What is the actual thing that you're doing? What is the player trying to achieve? You know, what's the point?

This being a game about Barbarians in the Dark Ages diplomacy in AtG is pretty blunt and straightforward. Your goal isn't building trade networks, spreading your religion, or anything particularly nuanced - the focus here is on war and relationships which lead to it (or not). Having someone declare war on you is the punishment for unsuccessfully navigating the turbulent waters of international politics, and remaining at peace (and thus able to follow through on whatever agenda you're pursuing) is the reward for success.

Success won't come easy though, as some leaders are naturally cranky while others might be in the midst of a feud and expect you to take sides. Who you support and who you snub is a big deal in AtG, because make someone mad enough and you can be pretty sure that they'll show up on your doorstep with lots of unhappy friends.

Furthermore, once you've made an investment in a relationship there should be a tangible reward for having done so. This is often where diplomacy systems in other strategy games fall short, as the fickle whims of the AI RNG can easily spoil a lot of hard work. If the decisions you make don't effectively keep you out of war then everything falls apart pretty quickly, so we also need to make sure relationships in AtG are fairly reliable.

With all of this in mind here are the elements we've included, and the very specific roles we have in mind for them.

.

Relationship Components

.

Leverage

Leverage is simply the sum of the player's Relationship Level, Influence, and Reputation (see below) that is or would be spent in a single transaction in order to get an AI Leader to do something, such as declare war on another Leader. After a transaction is finalized, the Influence and Reputation used in achieving it is spent and forever gone, while the Relationship Level remains unchanged.

Relationship Level (RL)

The most basic metric into diplomatic relationship is your Relationship Level. RL determines how favorably or unfavorably disposed an AI Leader is towards the player. It shapes that Leader's behavior, e.g. willingness to make Demands or go to war. In terms of specifics, it is a fundamental component of how a Leader decides which "Approach" to use (see below) and RL is a factor in determining Leverage when negotiating a transaction. The RL scale ranges from -4 to +4 and normally starts at 0, but can be modified based on a Leader's Personality Traits (e.g. -2 RL if 'Belligerent'). Having 9 total levels provides both enough room to do interesting things but also not so many that keeping track of where you are on the scale becomes a chore - each level matters while still providing enough wiggle room to move up and down with varying granularity based on the situation (refusing a request for aid is bad, but declaring war is really bad).

Influence (Inf)

Influence is, in essence, diplomatic currency with a Leader. It can be spent alongside RL and Reputation in order to increase your Leverage with a Leader when negotiating a transaction of some kind (e.g. asking the Leader to Denounce someone else, or publicly Declare Friendship). Inf is usually earned and lost alongside RL - doing a small favor for a Leader might increase both your RL and Inf by 1. This gives players the option of achieving some kind of bigger diplomatic option once, after which the Inf is spent. The player's options will be more limited after that, but they will still remain friends with the other Leader. The Player can proactively earn Inf by Offering a Gift or Holding a Feast in a Leader's honor (see below).

Reputation (Rep)

The fourth and final Diplomatic metric is Reputation, which is similar to Inf except for being a global currency that can be spent with any Leader (in addition to or instead of Inf). Rep is fairly hard to earn, and usually comes from facing down some kind of difficult situation (e.g. agreeing to help a Leader in need, or refusing a demand from Attila knowing he will go to war as a result). The effect and usage of Rep can differ based on the Leader and their Personality Traits (e.g. an 'Isolationist' Leader might ignore Rep entirely, while it might have double the effect on a 'Gullible' Leader).

.

Diplomatic Actions

The game contains a list of core actions that can be taken by either the human player or AI Leaders. The full list is as follows:

  • Denounce
  • Offer Gift (Wealth)
  • Hold Feast (Food Gift)
  • Request Aid (Food)
  • Request Alliance
  • Request Denouncement
  • Request War
  • Demand Tribute
  • Demand Denouncement
  • Declare War

Each Action has different requirements, costs, and effects associated with them that I won't bother to list out in full here, as it's a lot of numbers and I mainly want to focus on mechanics. For the most part Actions can usually only be performed once, so it won't be possible to butter someone up by holding feasts in their honor every few months or anything like that (though you can of course declare war more than once).

The effects of an Action are pretty much fully predictable. If Attila Demands Tribute of you, then you know that by accepting your RL will increase and he'll leave you alone. Refuse and it will drop, making you a target but also earning Rep that you can spend with other Leaders (should you survive long enough to take advantage of it!).

.

AI Leader Behavior

The final piece of the puzzle is how the AI leaders actually decide what to do on their end, and this is ultimately defined by their "Approach" towards each other Leader. Different leaders will have preferences for different Approaches, but the effects of the approaches themselves is fairly consistent.

The early game doesn't involve much diplo - players are simply exploring the map and run into each other. Not much happens on the diplomatic front during the first 48 turns - from this point on things start to get interesting though.

Early interaction mainly takes 3 forms:

  • Bullies trying to put other Leaders in their place, primarily through making Demands (and then following up with war should they refuse to give in).
  • Friendly Leaders (especially those with the 'Social' or 'Needy' Personality Traits) making requests.
  • Leaders who dislike each other trying to undercut their rival with the Player's help.

These behaviors are defined by which one of several Approaches an AI Leader has chosen to adopt for each relationship:

  • Ignore
  • Bully - Active
  • Bully - Inactive
  • Adversary
  • Enemy
  • Appease
  • Support
  • Friend

Similar to the Diplomatic Actions above there are a lot of numbers associated with each of these that I won't bother with here. Which Approach an AI Leader chooses for a relationship very much depends on the RL, with Personality Traits and to a lesser extent randomness also being factors.

A Leader's Approach will often change based on game events, such as being denounced or given a gift. Valid Approaches, along with the percent chance of transitioning between them after different triggers (e.g. receiving a demand) are defined by the Leader's Traits. These Approaches don't predefine what a Leader does, but do define the pool of options and the likelihood of each being chosen after how long (e.g. if demand refused 50% chance of denouncing 3-6 turns later).

I thought a bit about whether or not a Leader's Approach should be visible to the Player, and after going back and forth for a bit I decided to go ahead and make it public and announce all changes via notification. Transparency is very important in a system like this, and having something which fundamentally drives an AI Leader's behavior hidden under the hood felt like a mistake, especially given how straightforward the rest of the system is (RL, Leverage, etc.).

Making Approach visible means announcing changes via a notification is necessary, as the last thing we want to do is encourage players to check the diplo screen with every Leader every turn! This was a flaw in some early Civ diplo systems that I very much don't intent to repeat in AtG. On the plus side, this will help highlight the system for newer players, as although nothing is changing mechanically it reminds you that diplomacy is, you know, a thing, and if other Leaders are starting to get upset at you that you need to get up off your butt and be proactive.

As one last aside, I considered adopting some kind of "storyline" system where different Leaders would follow somewhat pre-structured plans for their diplomatic endeavors, but ultimately deemed it to be too much work for this game - perhaps a project for a future game!

.

That's it for this month's update. Given that diplomacy is the last big system to go in from here on out we'll probably be focused more on details and less on high-level design concepts.

Thanks as always for reading, and we'll be back in April!

- Jon

Comments

Only backers can post comments. Log In
    1. Missing avatar

      clayffo on March 7

      I’m honored to receive 3 paragraphs of criticism lol. However, I’m one of the very few active alpha testers. My job is providing feedback.

      You are the dev who is 4 years behind schedule with no firm release date. You asked for the money and broke promises not me.

    2. Jon Shafer Creator on March 7

      @ Pesticide

      I definitely agree that other strategy games have generally done a pretty poor job with diplomacy (including the last major title I designed, Civ 5). Diplo in AtG will be much more based on the discrete decisions you make and the personality of the leaders involved. Attila can't wait to get into a fight, while Genseric avoids them at all cost. Give in to Attila's demands though and he won't touch you, but refuse and you should expect war, and soon. It's a lot more direct than more traditional RNG-based systems, but the gameplay works out a lot better in most cases (though it can also be occasionally quite frustrating - not necessarily a bad thing though, from my game designer perspective!).

      @ Cerberus

      Good call - I'll definitely be discussing the Romans in more detail in a future update. They have occasional requests for you that you can use to earn rewards and keep them on your good side, and the late game will be focused on taking them down, but I've decided to push them more into the background given that often times you're just not that close to them for most of the game. AtG is a fair bit more asymmetrical than a game like Civ, and I found after playtesting that making the Romans absolutely fundamental wasn't something that worked super-well.

      @ clayffo

      I sincerely appreciate your feedback and how invested you are in AtG, and recognize that you strongly disagree with my decision to incorporate diplomacy into it. Given how other diplo systems have turned out and past delays in delivering this game I don't blame you for being skeptical. That said, I'm not going to change my mind, and it would be best if you accept that. This wasn't something I decided on a whim, but a topic I've been thinking about for many years now. As the designer I have to be the final arbiter in terms of what the game is and needs and I know that not everyone will agree with my decisions.

      In anticipation of that (as well as supporting something I think is both important and cool), AtG already has an extensive modding system available so if you're not a fan of the official vision it's very much possible to start reshaping things in whatever manner that you'd prefer via XML and code. Should neither of these options appeal to you I can provide a reimbursement of your Kickstarter contribution.

      I love your passion though and hope that's not the direction you choose to go. Your feedback on the forums is helping to shape a lot of the pacing and balance changes I'm now making to the game. If you care about AtG your best road isn't to fight me, but continue helping me make the game better. I want to see the 4X genre pushed in new directions and I know that you want the same thing. We aren't always going to agree, and that's okay. All I ask is that you accept these final few big decisions I'm making for the game that you clearly already see a lot of potential in. Will adding a complex new system take more time? Absolutely. But given how long we've already taken I'm certain that both from a fan and a business perspective it's better to provide a great game 10% later than an okay game a bit sooner. Conifer isn't Firaxis, Paradox, or Stardock - we obviously don't have the fanbase or finances of larger studios, and thus our ability to support this game long-term will be primarily based on how its received at launch as this will determine the shape of the game's lifetime sales curve. Anyone who wants to see the best AtG possible and more games like it should be rooting for the 1.0 version to be as good as possible.

    3. Missing avatar

      clayffo on March 6

      @pesticide Jon's own words from the update "Alas, this isn't really something that game AI is up for at this point in time, in large part because it's AI, and there's nothing you can do as a developer to convince players otherwise."

      do you really want to wait another year just to find out that a one man shop cannot do something that large companies like Fireaxis, Bethesda, Ubisoft and Take-Two can't do very well?

      Finish the game and start moving this property in the right direction

    4. Missing avatar

      Pesticide on March 5

      please do not look at games like total war or civilization for diplomacy, those games diplomacy systems have been broken and exploit prone since their inception.

      if anything i want to know why the ai decides to declare war or piece, not a random number generator, but real ingame indicaters, if its similar to the previous two mentions games u might aswel toss it out the window.

    5. Cerberus™ on March 2

      Thanks for the update, this transparent system sounds nice. I share your idea that intransparent diplomacy is not much fun.

      One thing I would also be interested in is contact with the Romans. Some information could be found in earlier updates, but it would be nice at some point to hear more about what their goals are, what they will do, and what I am supposed to achieve with them.

      I'm glad the game seems to be progressing, good luck with the rest!

    6. Missing avatar

      clayffo on March 2

      so this is not an update. version 23.1 was the update, but you didn't talk much about it.

      This is what we call a "white paper" which is defined as " is an authoritative report or guide that informs readers concisely about a complex issue and presents the issuing body's philosophy on the matter. It is meant to help readers understand an issue, solve a problem, or make a decision."

      You promised at the end of last year to provide monthly updates. This system you are proposing is really nothing special. it's a more complex and convoluted version of Civ 5 diplo.
      you should focus on actually improving the game and releasing version changes so we can see actual progress