Well, it’s that time of year again; GDC is almost upon us! I will be giving my third official GDC talk, entitled Playing to Lose: AI & “Civilization”. Unfortunately, it’s in the dreaded 9:00 AM slot, and since this is the first GDC where I will be sleeping at home instead of a nearby hotel, I better make sure I master the BART schedule and get there on time! Here’s the summary:

Playing to Lose: AI and “CIVILIZATION”
Speaker: Soren Johnson (Designer & Programmer, EA Maxis)
Date/Time: Thursday (February 21, 2008) 9:00am — 10:00am
Location (room): Room 2018, West Hall
Track: Game Design
Secondary Track: Programming
Experience Level: Intermediate

Session Description
Artificial intelligence performs a crucial role for any strategy game, providing a compelling opponent for solo play. While many of the challenges of AI development are technical, there are also significant design challenges as well. Can the AI behave like a human? Should it? Should the game design be adjusted to accommodate the limitations of the AI? Should the AI be exposed to modders? How do we make the AI fun? Should the AI cheat? If so, how much? Do we even want the AI to win? This session suggests some possible answers to these questions using the “CIVILIZATION” series as a case study. Ultimately, games are many things to many different people; fantasy, competition, narrative, even construction set, and the best AI will support as many different approaches to the game world as possible.

Idea Takeaway
This lecture is intended primarily for game designers and AI programmers who would like a deeper understanding of the consequences of high-level AI development decisions on the final product. Further, important lessons will also be shared for all developers interested in crafting a compelling single-player experience.

Intended Audience
Attendees will leave with a better understanding of the difference between a “good” AI and a “fun” one. Furthermore, they will learn the trade-offs inherent in deciding between the two options.

Essentially, I will be talking about the difference between thinking of the AI as the player’s opponent and thinking of it as simply an extension of the core game design (what one might call the difference between “good” AI and “fun” AI). There will also be a long section on AI cheating – the bane of my existence for many years – concerning which type of cheats are acceptable to players and which type are not, using Civ as an extensive case study. Further, I hope to prove that, for Civ at least, there is no such thing as – and never could be – a “fair” difficulty level where the AI is playing the same game as the human. Your mileage , of course, might vary.

This talk will be a bit of milestone for me as, presumably, this will be the last time I’ll be giving such an extensive talk on Civ. In fact, I feel a little sheepish about giving it as a non-Firaxis employee. I’m so used to getting feedback from my old colleagues on my presentations that I can’t seem to shake the feeing that now I’m just some dude spouting off about Civ, and the world already has plenty of those!

At any rate, hope to see some of you there…

When I was growing up, the first thing I would do when opening a new computer game (or a new wargame, for that matter) was flip to the back of the manual to see if there were Designer Notes. It was always a thrill to see them, giving me a little insight into the decisions, compromises, and challenges faced by working game designers. I regret that the industry has moved away from this tradition over the years (although the blogosphere is on its way towards replacing it), and so I was very happy to get a chance to write a lengthy Designer Notes section at the back of the Civ 4 manual. Since then, people who don’t have access to the paper manual itself have often asked me if this piece was available somewhere on the Web. Well, thanks to Steam, now it is!

Here’s the link.

The answer to the question from my last post – why was the Unit Workshop from SMAC not seen as a success within Firaxis? – doesn’t actually have anything to do with the game mechanics themselves. The problem is the graphics.

The Unit Workshop allowed the player to create new unit types. Of course, in order to make such a system work, you need certain limitations. In this case, the player creates a new unit by choosing parts from a list of Chassis, Weapons, Shields, and Reactors. The unit’s graphics were then dynamically generated based on the choices made. The problem is that all the units ended up looking very similar, even if they had quite different game values. The game had to cover all possible combinations, which led to generic-looking units because the graphics came from generalized algorithms instead of the imagination of the artists.

For Civ 4, we didn’t want to have one basic warrior model that could carry either a club or axe or spear or sword. Instead, we wanted to emphasize the difference between the units; a spearman would look a lot more shiny and metallic than the rougher, more barbaric axeman, for example. Being able to distinguish units is a key graphical issue (perhaps the key graphical issue) for strategy games, and the Unit Workshop tied the hands of the artists trying to make the game’s sci-fi units look distinct.

The Unit Workshop was undoubtedly a cool feature (in fact, it has certain parallel with Spore). However, game design is a series of trade-offs, and it’s not clear if the plus of creating your own units outweighed the minus of the units all looking the same.

In the comments section of my last post, José asked a question about why we didn’t incorporated either the commodity-based economy from Colonization or the Unit Workshop from Alpha Centauri into the core Civ franchise. It’s a very valid question as a number of ideas from these spin-offs have made their way back into the original series; for example, the civics system in Civ 4 is quite obviously an adaption of Social Engineering from SMAC. In the case of Colonization, its commodity system is simply too complex to match the simplicity of the other sub-systems in Civ. A detailed commodity system fits Colonization because that game streamlined many other aspects of the standard Civ model, such as technology or even military. As for the Unit Workshop – well, that is a very interesting question indeed. To many fans, this system was one of the highlights of SMAC, a unique feature that put Alpha Centauri in a class by itself for turn-based strategy games. However, it may surprise people to know that – by and large – the Unit Workshop was not seen as a success inside of Firaxis. I’m curious if anyone can guess what was the fatal flaw of this feature?

So, the second expansion for Civ4 came out last week. It has been received very well; in fact, it’s the top-rated “recent” PC game, according to Metacritic. I’d like to heartily congratulate my old mates at Firaxis for a job well done – I am especially looking forward to trying out Jon’s and Alex’s new game concepts as well as seeing the job Sirian did with adding events to the core game. The variety of mods and scenarios included in the expansion (ranging from space to fantasy to WW2 to world history) is truly impressive and demonstrates that our efforts making Civ4 so moddable have paid off. I am especially proud of the mods which were contributed from the fan community; I am always surprised how much enjoyment I get watching people become game designers within the Civ universe.

I have to admit, it’s a little odd being on the outside looking in on a major Civilization release. It’s a game that has gone through many different shepherds over the years and will probably continue to do so as long as people want to rule the world. Well, it was a fun ride while it lasted. Good luck to the new generation!

ApolyCon was last weekend here in the Hunt Valley region north of Baltimore. The convention was organized by one of the major Civ fan-sites, Apolyton. It was interesting to find out how far many people had travelled for the event; I believe that we had at least four from Europe (two from the UK, one from the Netherlands, and one from Greece). A number of Firaxians (including Sid, Barry Caudill, Dorian Newcomb, Alex Mantzaris, and Jon Shafer) dropped by to talk with the community. I really enjoyed the event – it is very interesting to meet people who know all about the issues that have been floating around my head for the last five or six years.

Dorian and I gave an extended version of our GDC presentation on prototyping Civ4 – “extended” meaning that we were no longer constrained to fit it into a 50 minute time-frame. I believe it was recorded on video, so I suppose that will probably surface on Apolyton at some point. I’ll post a link when it does… until then, here is a link to the original slides from the GDC site.

Our Civ4 interface programmer, Pat Dawson, was a big fan of World of WarCraft. One of the most impressive things about that game is the flexibility it gives users to create their own custom interfaces. The interesting thing about that decision is that while it taps into the incredible resources of the user modding community, it is also a tacit admission that a game’s interface is best developed in concert with the players.

I first started playing WoW over a year after the initial launch. Thus, I assumed that a number of modifications had been made to the official interface since then. I noticed that the system for updating the progess of your quests (SHIFT-clicking on them so they appear permanently on the right side of the screen) seemed a little hacky. The text, for example, didn’t have a background and sometimes overran other interface elements. Also, the limit of only showing five quests seemed quite arbitrary… but it sure was useful! I asked my hard-core WoW buddy about this feature, and he said – sure enough – it was added in a post-release patch. Now, I have no way of knowing, but I strongly suspect that a user-created interface mod inspired them to make that change. The on-screen quest display seems like a classic case of showing what the user cares about as opposed to what the designer thinks the user should care about.

At any rate, getting back to Pat… he pushed hard late in the project for us to move all of our interface into Python. This decision really paid off in the long run as the amount of interesting and useful Civ4 interface mods is growing rapidly. In fact, a couple of these mods were rolled into our last two patches, such as ulfn’s Proper Score Graph and the health bars from 12monkeys’s Plot List Enhancement. Quite simply, they fixed things we could have done better – no one knows how best to make an interface than someone who uses it day after day after day. We play our games a lot, but we can never play them as much as our fans do.

I have no idea if this is happening in the mod communities of other games, but I also enjoy the “compendium” mods that seem to be popping up, which merge together all of the useful interface mods out there. Guarav’s Yet Another Unaltered Gameplay Mod is a good example. There are lots of good changes here – a Foreign and Military Advisor, a Civilopedia with a persistent menu “pane”, Great People quotes, triggering diplomacy reminders and messages, showing turns left for Culture and Great People Points, a customizable Domestic Advisor, and so on. These changes are very interesting to see as a designer because they meet the informational and aestethic demands that the community has for the interface.

This is pretty cool. I’m not sure how it works exactly, but it’s probably just grabbing whatever info it can (turns, score, years) from the PitBoss app and then spitting it out to the Web. I love to see these types of “secondary” utilities appear – they provide a strong argument for using non-proprietary data and scripting (such as XML and Python for Civ4). By choosing standard formats, it becomes much, much easier for modders to create tools that extend the functionality of the original game. The CivStats site provides a great service for PitBoss games – allowing all users to get a quick overview of the game’s management and pace. It also provides a neat voyeuristic feel – I like lurking to see if all those 18-player game can acutally work.

Everytime I see an 18-player game, I am also reminded that, for better or worse, the limits we set as developers truly matter. If we made the limit 32 players, those would all be 32-player games… and would probably be progressing four times as slowly. These decisions are always tough calls.

Awesome.

Actually, it is a little spooky to read, especially the parts he gets right. For example, when deciding that Liberalism + Scientific Method = Communism, I meant to suggest that the latter is a scientific response to the former. The line between science and philosophy was pretty vague in the nineteenth-century. Of course, studying and analyzing history to predict its eventual outcome – as Marx did – is now no longer really seen as an achievable goal. I’ve always felt that his inspiration for doing so was the great leaps being made by comtemporary science in understanding the natural and physical world. Perhaps he felt the “socio-economic world” could be dissected just as well…

Well, it is done! Actually, it was done a couple months ago, but it has taken me awhile to wind down. I suppose it’s something like the bends – you can’t just go from 80-hour weeks to “normal life” in a weekend. Or, at least, I haven’t learned how to do so.

We’re not quite done yet, either – we’ve got at least one more important patch coming, with some memory optimizations and connectivity fixes. However, it’s probably about time to get a little reflective. I’ve been working on Civ games for five of the last six years of my life, and when I came on board at Firaxis, I don’t think I had any idea what was possible for the series. I loved, loved, loved Civ I – it was basically the only game I played in college – and I had a ton of ideas for how to improve the game when I arrived in Maryland in early 2000.

Most of the ideas, I’m afraid to say, were terrible. It took a lot of time to learn how to design in the Civ universe – a love of history and enthusiasm for games was definitely not enough. The parts of Civ which have always worked are the parts that are most transparent – that the player can keep in his or her head the easiest. It goes without saying that simplicity is important, but so is consistency.

For example, the appearance of Great People in Civ4 was originally going to be a random event based on certain factors, somewhat akin to how Military Leaders were created in Civ3. The problem is that the user is not in control – there’s a secondary layer separating them from the actual game mechanics. For some titles, that type of indirect gameplay is ok, even desirable. Civ, however, has a tradition of “boxes filling up with stuff” and following that tradition was one of the reasons culture worked well in Civ3. Thus, we wised up and follow that model in Civ4 – once a bar fills up with Great People points, Shakespeare (or Michelangelo or Einstein) is born. Players can now strategize how to get the most Great People and from which cities and how soon and of what type and so on.

At any rate, it’s been a long process, not just understanding what ideas work but also what type of ideas work for Civ. I’ve been wrong more often than I’ve been right, but fortunately there has always been an abundance of good feedback available, from our internal play session to closed beta testing to the general forums. I could have never designed Civ4 without it.

Here’s a few links from the last few months that might be of interest:
Gamespy Interview looking back on the game’s successes and failures
Planet Civilization Interview on the testing process (conducted by Thamis, one of our beta testers)
Eurogamer feature from my press tour through Europe with Sid

And I hope you’ll forgive me for this, but I just have to share. Civ4 came from so many people, from the team to all our fans, from Sid and Jeff to my family – and I am so thankful.