I read a great article on Gamasutra that was both fascinating and infuriating, simultaneously totally correct and absolutely wrong. It interests me as a coder, a designer, and a gamer, and since I so rarely get this excited about something, I figured I'd share my percolating thoughts.
I feel like the author gets the big picture mainly right, and has some good ideas for a solution; however, he gets the details wrong. The article is long, so let me summarize: there is Too Much Clicking! As computers become more powerful, they allow players to control more aspects of the game - he focuses on the number of units in strategy games. However, the player becomes bogged down in the number of decisions to make, spends all of his or her time tediously clicking, and as a result the game is less fun to play than it would be with fewer units.
The biggest problem I have with this article is that his main example, the Civ series, is a horrible one to use. First off, he bases his entire argument around Civ III, and off-handedly dismisses Civ IV as just being a graphical update that doesn't solve the underlying problem. This is completely and utterly backwards. The single greatest thing about Civ IV, in my opinion, is the incredible progress it makes towards solving the very problems he talks about: minimizing the amount of repetitive work a player needs to do. This was a core goal of the Civ team, as has been discussed for well over a year in interviews with the designers, and I think they greatly succeeded. They accomplished this through gameplay changes (no longer needing to clean up pollution, less painful civil unrest), through interface changes (more tasks such as changing build orders and setting research goals now available from the main map), and through successful encapsulation (the option of automating settlers, using city governors, etc.).
Not to mention that the Civ franchise isn't a good example of his core contention, that more powerful PCs have caused us to use more units. My current PC is roughly 400 times faster than the one I played the original Civilization on, yet there aren't that many more units than were in that game. I think the author's point applies more to real-time strategy games than turn-based ones, so it's odd that he focuses so much on Civ. In my experience, turn-based games have a much better appreciation for how much a player can reasonably be expected to keep track of.
Leaving aside the fact that his examples bug me, though, what's his point? It's essentially that today, most strategy games are God games, where the player has control over all the minute details. He would like to see the character much more in the role of, say, a general or a president, who has a limited number of subordinates that take and follow orders, but who is not responsible for ordering the actions of each individual soldier or citizen. In this way there are less conceptual objects for the player to interact with, and he or she can better focus on long-term goals than minutiae.
He has a number of interesting digressions. One is that he dislikes the choice many games (including Civ IV) present of "We'll handle the details for you, but if you want to you can take over them and control it yourself." He dislikes this because, first, AI has not yet evolved to the point where it will make decisions as good as what the player would make. On a related note, in competitive multiplayer games, someone who controls the minutiae will have an advantage over, and beat, someone who delegates to the AI.
On a possibly related note, he seems to hate 14 year old boys. I wasn't sure whether to be amused or annoyed by his obvious prejudice. Let's face facts: if you're a guy in your 20s, odds are extremely good you were a 14 year old boy at some point in your past. His annoyance seems to come down to the fact that these people are playing games the "wrong" way - such players are much better at clicking, and so they have an unfair advantage in cerebral strategy games. Personally, I think that the main (not sole) purpose of a game is to be entertaining, and that if someone gets pleasure out of clicking a lot and feeling like God when sending virtual soldiers to their death, that's just as valid a way to play the game as someone who enjoys being an armchair general.
(That said, I think it does behoove game designers to push players in the right direction. I love a lot of the changes that went into Civ IV, but didn't realize until after I started playing it just how much tedium I had endured previously.)
I think that Civ IV could conceivably stand as a powerful counterargument to the author's thesis. Again, he would like to see a very limited number of actors who the player interacts with in order to accomplish their goals - in the case of Civ, I would imagine this would look a lot like a game where you deal mostly with your advisors rather than units. But I think Civ IV shows the enormous improvements that can be realized by streamlining an interface, as opposed to radically restructuring it. This makes sense - if it's easy and quick to interact with items, you can handle more of them before you start to lose track of what's going on. In my opinion, the classic problem with Civ is not that you have 30 cities or that you have 100 units, it's that each turn takes 15 minutes and when you start your next session you can't remember what that one settler was supposed to do. Civ IV has too many solutions to this problem to mention here, but to pick a single example, you can now set rally points: "Every unit built by these 3 cities from now on should be sent to this location." That way you can forget about the tedious task of remembering what city a unit came from and is going to, and instead save your brain cells for the more interesting problem of who your units should attack once they've all arrived.
Still, I think the author's core assertion is right: too many games do require too much micromanaging to succeed, and future design should help the player stay focused on the most interesting aspects of the game.
His technical thoughts are right on as well... I love OOP and design patterns, but it's always confusing when technical terms overload words from the real world, and the definition of "Object" is probably Exhibit A. Personally, I think that everyone who graduates with a CS degree should be required to implement the Visitor pattern, just to get that into their head. We need to think about systems, not just items.
I think his idea of a user interface profiler is wonderful as well. I've never particularly enjoyed doing user interface work, yet my two major jobs have largely consisted of it: first developing a web interface at Cerner, and more recently working on the BREW client at Rocket Mobile. In both cases, I can sort of intuit when something is a good or a bad interface, but at least from what I see, there's very little process that goes into developing a standard. There are a few good hard and fast rules (CLR should always take you back a screen, all events should be responded to within half a second), but so much of the time it seems clear that the requirements are written by someone drawing on a screen, and not actually using a device. I love the author's idea of transforming evaluation of UI from a purely qualitative experience into a largely quantifiable one. (The way he wants to do it, incidentally, is really quite simple - log what the user does by counting key presses, mouse movements, mouse clicks, etc., then generate an overall score expressing how much effort the user took to accomplish their tasks.) It will never become a sole standard, but this could be a very low-cost way to quickly identify bottlenecks that should be addressed.
Returning back to the core problem discussion: I think he is absolutely correct that current AI technology cannot be as effective as a human. However, I'm much less convinced that the proper response to this is to force all players to use that inferior AI so they have a level playing field. This is a recipe for frustration and disaster. The first time a player sends a general to wipe out some rebels, and that general fails, the player is likely to quit the game. It's one thing if you lost yourself - that's a case where you can try again, learn from your mistakes, and eventually figure out how to do it right. But when you fail because the game failed you, I think players will be a lot less forgiving.
So what's the proper solution? I can think of one short-term solution, one long-term one, and one radical one. The short-term solution is the Civ IV approach: make the AI as good as you can, minimize the harm that can be done by an ineffective AI, and allow the player to oversee or replace the AI. Even though I am a long-time Civ veteran, there were enough new features in Civ IV that I felt a little overwhelmed in my first game; rather than worry about all of them, I chose to focus on a few at a time. A big example is improvements. Back in the old days of Civ I, settlers could do exactly three things: irrigate, build roads, and build mines. Each type of terrain had a single type of improvement that made sense (irrigation on grassland, mines in hills), plus you could build roads everywhere, so it was always clear what I should be doing. In Civ IV, by contrast, there are a staggering number of improvements that you can build, both those particular to a resource (winery, pasture, quarry, oil platform, etc.), and those which apply to multiple terrains (windmills, lumbermills, workshops, cottages, etc.). I just wasn't prepared for it, so for that first game, I had the AI control all my workers. Fortunately, the AI is very good, and allows you a fair number of general instructions to give ("Improve the nearest city", "Build routes between cities", etc.), so I could be reasonably sure that they were doing what I wanted. Occasionally I would take control, if I wanted to make sure that a particular road was built ASAP or something similar, but for the most part I was highly impressed; the end results were far better than I was used to seeing in previous Civs, where automated settlers would irrigate deserts and otherwise be nonproductive. I continued to automate workers in the second game, but this time paid more attention to what they were doing, and came to get a good general idea of where it made sense to build windmills, the order in which to irrigate tiles, and other topics that had previously seemed arcane. These days, I manually control the settlers around a new city to achieve a certain objective - if I intend the city to be a commercial powerhouse I'll build a lot of cottages, while a future Great Person generator will contain many farms - but once it's secure, I feel fine automating my workers with instructions that they not mess with what I have already done. They do a good enough job, and it's not worth my time to micromanage them. Once again, I think that this is an excellent design for the current era of gaming: give the player the option to delegate decision-making, try to make those decisions as intelligently as possible, and allow the player to interfere with the automation whenever desired.
The long-term solution is to make the AI even better. When I want to take a city in Civ IV, I need to build up a force, select a target, assemble the force, and act to take that target. I know from playing against the computer that its AI is not very good; they require massive numbers to achieve objectives because they have an inadequate grasp of the strategy involved (though they are still better than in previous versions of Civ). However, if I had a Military Advisor who could fight as effectively as I can, I wouldn't mourn the loss of ability to control individual units. I'd love to be able to communicate something like, "All surplus production should be directed towards building and upgrading an offensive military force. Once it is ready, capture Sparta and then report back to me." To me, that's the fun part of Civ: the grand strategic maneuvers, not making sure that your catapult always has a good defender in the same square. Once AI advances to that level, I think that the author's article will be much more feasible to follow.
Now for the radical solution. I continue to be fascinated by the promise of Spore, Will Wright's upcoming everything-game. One little nugget of information is that as your creatures' DNA evolves, it will eventually be "sent back" to EA, which in turn will use it to seed other random "worlds", including those on other computers. This means that, when you start a new game of Spore, the creatures you will see under water and on land aren't just built-in to the game, or something randomly cooked up by the PC: they will be the result of (ahem) intelligent design on the part of thousands of other players who have worked hard to come up with creatures that can survive and thrive. I find this wonderful because, again, humans are much better at this stuff than computers are, so every player's game will be enhanced by the work done by thousands of other players, without any one of them needing to do anything.
(Tangent: here's a cool presentation of Spore at E3. Will Wright shows off some gameplay, and chats a bit about the game; towards the end, Robin Williams takes over the controls and creates a hilarious creature.)
I would love to see something similar done for games like Civ. What's awesome about Civ is that it's a game for everyone: whether you love the military strategy, or the diplomacy, or building the physical empire, you'll be able to do it. What's not so awesome is that you also need to do the things you don't enjoy so much, or your empire will fall. What if the game allowed you to focus on just the parts you enjoyed, while human-created intelligence took care of the rest? For example, you could play the role of Commander In Chief and play the entire game marshalling your units and taking over enemies, while trusting the brainiacs to keep discovering cool new weapons for your army, and the builders to ensure you have a steady supply of trained soldiers. The whole time you are playing, the game is paying attention to what you do - seeing how you react to particular tactical situations, what sort of reconnaissance you do prior to taking a city, when you upgrade your forces. At the end of the game, it takes note of your overall performance, and decides how formidable you are and how to characterize your strategy. This tactical information is uploaded to Firaxis's site; when I start my game, I'll decide that I don't want to focus on military, so I'll look and see what intelligences are available. I might see "Scorched Earth Sherman," "Multiple Theater Manfred," "Defensive Dennis." I'll install one, issue occasional instructions when appropriate, and fire him if I decide I'm displeased with his performance. (Conversely, if you're playing as the general and decide you don't like President "One Defender Per City Chris", you can orchestrate a coup and install "Billions of Bombers Brad" instead.) The goal is to let each player focus on what's most fun for them, and leverage the intelligence of other players to create a game with more personality and smarts than a few programmers could throw together.
In games other than Civ, the possibilities are even more exciting. Imagine an espionage game where some players focus on selecting missions and equipping teams, and other players actually carry out the missions. Or a real-time strategy game where one player has a view of the entire battlefield, and other players get to focus on particular objectives. In that example, the click-happy 14 year old boy can click like crazy to take over a gold mine, while the grumpy Gamasutra writer can see the larger context of that gold mine, send reinforcements when appropriate, and plan for the next move.
Anyways. That's it for the article. Now that I'm thinking about it, I thought I'd write a bit about why I'm responding so strongly to it.
I've recently thought a fair amount about how my attitude towards games has changed. It was only a few years ago that I could easily spend hours and hours each day on games of all kind: I'd be playing an RPG for hundreds of hours on my PC, plus Super Smash Brothers Melee and Virtua Fighter 4 against my dorm buddies, plus stupid Flash games, plus the games I was actually WRITING, etc. These days, I can easily go for a week without playing any games at all. I just have so little time left to play games, and when I do get free time there's a lot of competition for what I'd like to be doing with that time. Games often get the shaft.
Which is pretty sad, because I enjoy gaming so much. The best games I have ever played, ones like Ultima VI and Ico and Civilization and many more, have touched me as deeply as any novel or movie. However, when it's hard for me to find time to watch a 2 hour movie, how much harder will it be to find time to beat Elder Scrolls IV: Oblivion?
Because of that, I've been forced to be a lot more selective in my gaming. Even some ones I've mentioned on this blog I have had to step away from. Jagged Alliance 2 has been a lot of fun, and in the old days I'd be addicted, but each individual battle takes over an hour, and there will be more than a hundred battles to get through it, so I'm just not able to steel myself to continue it. I started a platformer called Cave Story which is really retro and, again, a really fun game; but it's taking a long time to play through, it's too easy to die, and once again, I don't think I'll be able to stick through it to completion.
So what games do I play? It depends. RPGs are still my first love, so it's inevitable that I will play Oblivion some day, possibly when I'm unemployed. However, the way I play these games is different from before. I play purely for the story, for the excitement of discovery, and for the pleasure of losing myself in a virtual world. There was a time when I would have been interested in building up super-powerful characters, or spending hours to find the ultimate weapons and spells. No more. That sort of effort is time that I could be spending on an entirely different game, and the amount of pleasure it brings me just isn't enough to justify the opportunity cost.
That's also why I got such a kick out of Grim Fandango: when you play a good adventure game, almost nothing is rote or repetitive. Every thing you do in the entire game is either story or puzzle, so you are either being entertained or actively solving a problem. At every step of that game, I felt like I was doing what I wanted to be doing, not doing grunt work so I could do something fun later on.
On a sidenote, I think this is a big reason why I love GTA: It fits so nicely into my new situation. The mission-based structure lets me feel like I've accomplished something even if I can only play for half an hour before bed. At the same time, the unbelievable scope and sweep of the game world make me feel like I'm not being limited, and if I have time, I love exploring the nooks and crannies. I'll never spend the time to, say, find all 100 snapshots in San Fierro, but I will love cruising past the ferry building at midnight and spotting a car for the import/export docks. The sheer number of things you can stumble across in those games make it a very fun-dense game.
And really, it is all about fun. There was a point where I realized, "You know what? Getting a 100% completion score just isn't as much fun as messing around. Just let it go." It feels weird to think about fun so analytically, like it can be measured and quantified, but the truth of the matter is, I know when I'm having fun and when I'm not, and I do prefer the former.
Bringing it back around to the original topic: this article resonates with me so much because I've come to realize that I've put up with an awful lot of non-fun parts of games: leveling up to take on the boss, cleaning up pollution, trying to find 4 pieces of bloodmoss in the swamp. In the past those have been acceptable game design choices because they increased the complexity and length of games, and people in their target audience demanded games which lasted longer. As the games market continues to expand, though, it will be critical for developers to offer games which appeal to busier people who want the fun, challenge, and escapism of a good game, and to have it delivered in doses as concentrated as possible. This may lead to a bifurcated market, with "real" games on one side of the aisle and "mini" or "light" games on the other; what I would prefer to see, though, is an increase in games like Civ IV or the GTA franchise that flex themselves according to the whims of the player. Such games will reward the player no matter how they prefer to play, which will expand the market of consumers for each game and expand the universe of games for each consumer. I have hope that this will happen.