Paradox Development Studio's Crusader Kings II has been one of the surprise critical hits of the year, garnering attention and sales beyond what might be expected after a dozen Europa Universalis-style grand strategy games. I'm one of those critics, having reviewed it positively, and then only grown more impressed with its systems. What makes Crusader Kings II special and deserving of this praise is that it successfully models historical human behavior using a transparent system.

When Paradox offered me the chance to interview Crusader Kings II's project lead and designer, Henrik Fåhraeus, I took it in order find out just how this had been accomplished. Modeling human behavior within complex systems is one of the great dreams of video games (just ask Chris Crawford). Crusader Kings II, almost out of nowhere, impressed me by "...[building] a system that effectively simulated the power dynamics of medieval dynastic politics, and then [forcing] me to engage with and learn just how destructive those systems could be." How was this accomplished?

My expectation, built on years of developer interviews stressing the hard work of game creation, was that Crusader Kings II's success came after meticulous planning as well as constant testing and tweaking. But what came across in the discussion with Fåhraeus was something different: the power of serendipity and improvisation in creating a great game system.

Influence

I asked Fåhraeus about the influences and goals for the project, and he gave two main motivators. First, Crusader Kings II was another evolutionary step for the team.

"It was more a matter of taking all the similar systems we've developed at Paradox Development Studio over the course of multiple games (Europa Universalis, Crusader Kings, etcetera) and trying to find a better, unified way of realizing the same thing, with the added requirements of a character based game."

Paradox's commitment to pushing the Europa Universalis game engine to its limits is both slightly bemusing (it was clunky enough when it debuted over a decade ago) and impressive in how much the team seems to be succeeding.

He also mentioned a few games as influences, saying that Crusader Kings II "...(drew) inspiration from various spiritual ancestors; some relatively obscure, like The Lords of Midnight and Birthright: The Gorgon's Alliance and others not so obscure, like The Sims."

I think he may overestimate the degree to which people would see the breezy suburban modernity of The Sims as comparable to the epic medieval scope of his game; still, it's definitely there. You build relationships with your peers by giving them gifts and building the kingdom's infrastructure in a manner conducive to happiness. Instead of paintings and refrigerators, you ensure that dukes possess the counties that form the traditional bounds of their duchy, and if that's not enough, you can grant them ceremonial titles like "Keeper of the Cups." The goal is still the same: encourage the people that you interact with have more positive than negative relationships toward your character.

There were also era-specific considerations. "Of course, we wanted intrigue and vicious backstabbing in Crusader Kings II, since it was so sordidly common in real medieval history. Novels like the A Song of Ice and Fire and Dune books (both of which I love) only reflect reality."

That Game of Thrones connection, as mentioned by several critics, is quite clear in the game, and CKII quickly spawned a detailed, high-profile Game of Thrones mod. Fåhraeus's specific example: "Conan II, Duke of Brittany at game start in 1066, was murdered with a pair of poisoned hunting gloves, probably on the orders of William the Conqueror." With this sort of event as the model, the goals of the game become clear. "In order to facilitate this, AI characters especially needed to have clear opinions they could act upon."

Planning

"Very early on during the design phase of Crusader Kings II, I realized that the way we usually model relations between states in our games would not suffice." This is something of an understatement: in Europa Universalis, there are a few hundred nations at most, and most of them never interact with one another. In CKII, there could be hundreds of people who need to interact within a single kingdom, thousands in the game world at any given time, and thousands more people who have died or are yet to be born.

Fåhraeus describes a clear plan for making this work: "What we needed for this sequel was a simple, unified mechanic; easy to understand, yet deep and meaningful. Thus, we came up with the Opinions of Crusader Kings II; one single value summed up from a number of clear reasons why someone would like or dislike you (e.g. holding desired titles, refusing a request, or just plain old personality chemistry.) Opinions are unilateral, meaning the feeling is not necessarily mutual. The AI, of course, relies heavily on the opinion values, meaning it can act intelligently and in character. The end result turned out to be one of my personal favorite features in Crusader Kings II."

These Opinions work on a scale of -100 to +100, and are most commonly used for understanding when vassals might rebel. A powerful duke may hate his emperor, and you'll see why -- maybe the emperor has just acceded to the throne (an immediate -20) or that emperor has seized titles from other vassals and is viewed as a tyrant.

"All these numbers are visible and easy to access via tooltips. As a player, your goal would be to placate that duke, perhaps by granting him more land or sending money, or perhaps to isolate him from his allies by befriending them instead.

Implementation

With the idea of multiple unilateral opinions set, Fåhraeus started adding them to the game. "The original Crusader Kings... (is) one of progenitor systems of the new Opinion mechanic, though it was far more limited. For Crusader Kings II, we took many of the traits from the original game, but with some modifications. For example, I wanted to have the Seven Deadly Sins and the Seven Virtues as traits. There are other changes too, of course. For example, some traits are now congenital and can be inherited by your offspring."

What I found fascinating about this is that Fåhraeus simply "wanted" to have those particular traits, so he had them added to the game, and they worked. In fact, they work extremely well at multiple levels: they add a medieval, religious flair to the game, but they also effectively give the characters more personality and set them at odds with one another -- a Lustful duke will automatically be wary of his Chaste queen, even if they get along at every other level.

This is what made me particularly interested in the testing process. My theory was that this was an intense process, with each trait examined and tweaked for optimal performance. I was wrong.

According to Fåhraeus, "...Actually the opinion system never really got much criticism and did not undergo any major revisions from the first implementation in Crusader Kings II, although the testers helped the development and came up with many new opinion modifiers and helped balance their magnitude. In my experience, quite often you know a feature works really well when people who test the game do not talk about it much. Beta testers usually do not bother with praising well-functioning mechanics, being far too busy to find faults in the game."

In an attempt, perhaps, to make video games and the act of creating them seem more serious, Western developers often seem to emphasize just how much work, how much planning, how much "iterating," and how intentional their game creation process is. Japanese designers, on the other hand, often seem to emphasize feeling, speaking about their work almost as if it's poetry. Fåhraeus certainly isn't being poetic here -- he's quite analytic in all his responses to my queries -- but I was delighted to hear that Crusader Kings II was improvised. Its developers had a good enough idea of what they wanted to see in the game that they were able to play it by ear.

I asked Fåhraeus just how much of it was unintentional and unexpected. "Thinking back to how I envisioned the system, I don't think I anticipated the sheer number of opinion reasons we ended up with. It is surprising how it can still 'click' in the end, with characters doing things for mostly logical and predictable reasons in Crusader Kings II."

Creating the unexpected for the player, on the other hand, was ideal. "The goal is to give players access to everything they need to know to keep their realm in order, but still make it vulnerable to accidents and tiny mistakes that can quickly escalate out of control. For example, if you give a title to an ambitious, deceitful courtier with no land, he will be grateful for a while, but then he'll want more; it's just in his nature."

But Fåhraeus stressed that while there were surprises, they were surprises within the team's control: "To be honest, I generally don't like to be surprised by the game mechanics; that means our design was too sketchy, that we didn't think it through.

There is, however, a huge difference between unexpected (weird) and expected (reasonable) random outcomes. The former is normally undesired (with rare exceptions), whereas the latter is exactly what we are aiming for. In fact, the main reason I love to play the games I've worked on, even with my detailed understanding of how they work under the hood, is exactly that intrinsic element of chance, yielding virtually infinite replayability."

Reception

Fåhraeus is understandably happy with how his game turned out, and the role of the opinion system in that experience. "In the end, I really feel that Crusader Kings II turned out quite unique in the grand strategy genre as a game that is all about characters, their ambitions, gambits, personalities and ultimate fates. Just the fact that players care less about which country they are playing than the epic story of their ruler and dynasty as it unfolds through the centuries is fascinating. Because no two playthroughs are ever the same -- or even very similar -- and the opinion system in Crusader Kings II really is at the very heart of whole gameplay experience."

He's pinpointed perhaps the most impressive aspect of the Crusader Kings II experience: it forces the player to role-play. By making the player control the person at the head of a dynasty, and not the dynasty as a whole, or a full nation, the game sets up an intrinsic group of constraints and goals, which are constantly changing thanks to deaths, births, and the Opinion system. Strategy gamers often give themselves these arbitrary goals in order to tell stories or get more out of a game they've mastered, but Crusader Kings II gives that experience to everyone.

This is, I think, why it's been so well-received even from people who otherwise wouldn't given yet-another-Paradox game significant critical attention. Fåhraeus expressed pleasure and a bit of confusion at this. "Well, I always hoped that players would grow attached to the characters and engage in their unfolding, open ended, stories. For me, that is really what Crusader Kings is all about, and what separates it from our other games.

"Of course, it is really gratifying to see that we've succeeded in that respect! I am surprised by just how different people perceive Crusader Kings II to be compared to other games they've played... I am not complaining, though -- no one could be happier than I about the level of success and the amount of attention that Crusader Kings II is getting!"