Archive for the ‘Replayability’ Category


A trap I’m tempted to fall into when I design games for something other than my own enjoyment is to start designing levels for the game. It’s not that I’m bad at designing levels, or that levels are inherently bad, but it’s just too easy to let your games degenerate into static, linear puzzles. I mean this very broadly. It may be a blast to play through Half-Life 2, but eventually every fight has been puzzled through, consumed, solved, and exhausted. After I finish a game like that, I rarely come back.

I want to make games that will stay on people’s hard drives long past its expiration date, and still make sales to nostalgic gamers after they’re discounted to a dollar on Steam a decade later. These types of games inspire me as a designer, and drive me to analyze them and ask what makes them tick. Most seem to have at least three things in common — I’ll discuss them, then give examples of how Civilization V, Starcraft 2, and Team Fortress 2 meet them. These three examples are chosen because they are all games with the potential to consume hundreds of hours of play time.

The challenges change continually, both within the game and between sessions.

Multiplayer, procedural content, and artificial intelligence are all ways of continuously transforming the challenges so that no two games are exactly the same. There is some element of chaos and unpredictability in the system. This isn’t just the flavor of the challenges either, or the color of the textures — it’s a recombination of the fundamental elements of the game in ways that totally change the game. Most games are about solving some kind of problem. In these games, the problems are infinite.

In Civilization V, the randomly generated terrain and layout for the world, alongside the AI opponents, means that every game tells a different story. In Starcraft 2, each online opponent will play a different strategy against you, and the real time nature of the game means even the same two strategies will play out differently every time. In Team Fortress 2, every recombination of players, classes, positioning, and timing makes each game unique.

The player is rewarded for applying new, creative solutions to the challenges.

The games that grab me are the ones that let me use my head. That doesn’t mean figure out how to solve a see-saw physics puzzle the designer put in (another loving Half-Life 2 reference), that means there were many ways to solve the problem, and you are free to find the best one you can. The rewards for a better solution may be efficiency, aesthetics, preparation for future challenges, or radically changing the direction of the story.

In Civilization V, I can try out different build orders, diplomacy strategies, and military tactics, finding degrees of success along different axes. In Starcraft 2, every situation has multiple viable solutions with different outcomes, and even at the highest level of play, players have distinct and recognizable playing styles, and often work to discover new tactics. In Team Fortress 2, every skirmish and shootout is a puzzle to outwit and outplay your opponents, and winning faster or taking less damage requires quick, creative play.

The game is almost impossible to completely master.

As soon as a player has developed a mental algorithm for successfully solving every problem the game can throw at them, the game tends to be less interesting. This may not be true for all players and all games, but the ability of a game to never lose its challenge will help it to be spared the shelf for a longer period of time. Golden paths, unstoppable strategies, and permanent “god mode” abilities kill games.

In Civilization V, even if a game-breaking strategy exists with one civilization on one map type, it may be totally non-viable when playing with a slightly different setup. In Starcraft 2, there’s no way to anticipate every possible response and counter to any strategy, and even complete mastery of the mechanics leaves a rich depth of gameplay in strategy and thinking on your feet. In Team Fortress 2, the dynamism of multiplayer is bolstered by new maps and metagame-changing patches on a regular basis, continually altering and adding challenges whenever older ones seem mastered.


Read Full Post »

A couple weeks ago, I posted about how I’m not a big fan of linear predictability in level design, and the Groundhog Day effect described by Jonathan Morin (Lead Game Designer at Ubisoft Montreal) in his talk about predictability in games and how to avoid pre-conceived experiences in game design. Put simply, I think it ruins replayability, and thus the effective length of the game — and that’s something I value highly in games.

But I just got done reading another argument comparing Groundhog Day to repetitive gameplay, this time favorably. Robert Buerkle argued in Digital Déjà Vu, an article in The Escapist, that the opportunity to replay a section of the game and fix your mistakes can be a tremendous pleasure in playing. After all, isn’t that a chance that people can’t get, but often want, in the real world? He makes a persuasive argument that there’s pleasure to be had there, and even though I argued contrary to his view on those kinds of sequences, I have to agree on a basic level that I can enjoy the same thing, at least for awhile. And when he uses Groundhog Day as well, this time he’s pointing out that while the time loop is a curse for Phil, it serves to allow him to grow and learn and adapt until his old miserable life has been turned around and he has the perfect day. It’s a curse for him, but it’s also wish fulfillment for the audience — the fantasy that we could have the chance to do the same thing.

I think he’s right, and that’s a lot of the appeal of the movie; indeed, it’s one of my favorites. I’ve even read the screenplay. But while on a small scale, fixing your mistakes is a blessing, when extended over a long time, many people find it to be a pain when they’re forced to do so. Phil starts out bewildered, then becomes decadent and enjoys his predicament, but eventually he becomes bored and depressed, even resorting to suicide. For me, it’s the same thing — as fun as it can be to redo a section to get things right, if you’re forced to do it again and again and again without letting your mistakes be forgiven, it becomes a serious problem.

I think that’s why a lot of modern first-person shooters implement constantly regenerating health. The Call of Duty games started out with health packs, and I can say from personal opinion that my enjoyment increased greatly once the series switched to constant health regeneration. While a bit of Groundhog Day is not bad, over the long run it’s more fun for me to be a James Bond or John Rambo, enduring endless action without dying, than a Phil Connors, repeating the same scene over and over again.

But doesn’t Phil Connors get the girl in the end? Doesn’t he grow and become a better person? And haven’t I seen that movie more than First Blood, the only Rambo movie I actually liked?

These are weak and pointless objections, but what I’m really trying to do is segway into a defense of the movie that draws a division between what Groundhog Day does, and what games that restart you at checkpoints do. First, it’s worth noting that Groundhog Day represents the entire day as one large loop; he doesn’t get to do things piecemeal and revert if he makes small mistakes, he has to do it all in one go. This eliminates some of the nasty five second life grind I get on occasion in Call of Duty games. Of course, restarting at the start of a level is even worse — now the player has to repeat all the things they did right. That, I think, is the other difference with Groundhog Day. The movie represents a time loop trapped in a small town, but within that town are infinite possibilities. He can replay the same sequence of events, or he can do something entirely different. I haven’t played The Legend of Zelda: Majora’s Mask, but I know it’s based on the same principle, only on a three day time cycle. Like Groundhog Day, the Zelda games have never been linear affairs, the original was the closest thing to an open world that existed when it was released. Shigeru Miyamoto drew inspiration for the game from the woods and caves outside Kyoto — the possibilities and sense of exploration and adventure. If I’m going to repeat a sequence or scene, I want to be able to play with it, explore it, immerse myself in a broad possibility space. I want it to have replayability.

I believe that a big part of the pleasure of repeating the same sequence of events over and over again is the pleasure of learning and mastery through practice. I think the exploration of this through repeating a precise sequence of events is interesting for awhile, but too puzzle-like and frustrating to give a game longevity. Roguelikes expand this from a micro-scale “learn the level” cycle to a macro-scale “learn the game” cycle. Because they use procedural content, you aren’t replaying the same sequence of events, avoiding much of the drudgery. Because the monsters, spells, and mechanics stay the same though, you can still learn how to fight different creatures, what tactics work, and how to stay alive in an unforgiving world.

After all — the chance to live life over again and use your wisdom from past lives is just the macro version of Groundhog Day.

I’m not saying that roguelikes are for everyone, or that they’re a superior form of game, or that Robert Buerkle is wrong for taking presumably greater pleasure in learning a specific level than I do. But I do think that on a fundamental, psychological level, roguelikes (which are more my style) grant a very similar gratification to players that he’s describing. They appeal to the same kind of perfectionist instinct, just on a very different scale.

Read Full Post »

Due to some unfortunate data loss several days ago, this is the inferior, reconstructed version of this post. I hope it serves well enough.

Sid Meier is often quoted as having said that a game is a series of interesting choices. It can be fairly argued that a game can be other things as well, such as a series of fine motor skill challenges of escalating difficulty, and debating this point gets into semantics. But if we accept the idea that a strategy game, at least, is about choices, then a very pure implementation of this can be found in the game Hidden Agenda.

Hidden Agenda is a game about taking on the role of the newly-appointed Presidente in post-revolutionary Chimerica, a fictional Central American country. Chimerica has long been under the corrupt governance of the right-wing Farsante dictatorship, but the people have risen up and overthrown the old government. Now in de facto control of the mechanisms of government, as well as the extensive estates of the Farsante family, a junta formed by the three key opposition parties has been formed. These parties represent Chimericans from all walks of life, from the wealthy cotton farmers and army brass to the poor campesinos and revolutionaries. Even the military has been boldly merged into an extremely fragile reconciliation army, with two army chiefs of staff, representing the traditional army and the revolutionaries. You have been selected as the compromise pick to lead the intrim government.

Your job is to restore order and maintain unity during the fragile post-revolutionary years, balancing the contrary political and social interests while acting to secure a better future for the people and the country. Do do this, you first must appoint a cabinet of ministers to advise you and make policy. Your cabinet has four posts — Agriculture, Military, Internal Affairs, and External Affairs. The candidates for these positions are drawn from the three political parties, National Liberation (communists), Christian Reform (compromisers), and Popular Stability (libertarians). Your choices of these ministers determines, to a great extent, the direction you will go in each of these departments. Each of the parties has the support of a substantial faction within Chimerican society, and neglecting any of them simply because you disagree with their beliefs risks your overthrow.

Once your ministers have been hired, you then play the game in week-long turns, each turn meeting with either one of your ministers, to discuss an issue they think is important, or one of twenty-one sample members of Chimerican society, ranging from foreign diplomats and army leaders to labor organizers and coffee workers, who will all petition you with an issue important to them and the people they represent. Regardless of who you meet with, they will bring up an issue, explain what they believe is the best course of action on the issue in a usually reasonably persuasive argument, and give you an opportunity to accept their advice. If you are meeting with a minister, you can call the entire cabinet to meeting and take the input of other ministers on the issue, possibly going with an alternative proposed by another minister instead. If you are meeting with a petitioner, you may ask the opinion of the single most relevant minister, but may not hold a full cabinet meeting on the issue unless you elect to instruct the minister to place the item on his or her agenda to address at a later time. This broadens your options, but you waste a turn by not taking action that week. Finally, you can request cabinet members resign or appoint new cabinet members as a free action, and review news excerpts and an almanac of quarterly statistics about the country, also as a free action.

If, one way or another, you can make it through four years in office without being voted out of office in an election, thrown out by the national assembly in a no-confidence vote, or, more likely, forced to resign in a coup d’etat, the game ends, and you might consider yourself to have “won” the game. Certainly, to simply maintain stability for that long is the most pressing goal for a new player. But there are other priorities as well, such as preparing for elections, and managing the future of the country. Even if you make it to the end of year four, are you really happy with your accomplishments? As you read an encyclopedia entry giving a historical perspective on your first four years in office, will you like what they had to say about you?

A full game of Hidden Agenda probably takes ten minutes to half an hour to play, depending on how thoroughly and carefully you read all the arguments presented to you by petitioners and your advisors. I have probably played Hidden Agenda about fifty times, and have explored the possibility space in it fairly comprehensively. Despite my attempts to treat the game like a puzzle to be solved, there is no “Golden Path”, or solution to the game that is clearly dominant. I can survive, easily; I can even swing the government far right or far left without losing my grip on power. But still, there is no one right way to solve the country’s problems. This is partially because there is no objective standard to pass judgment on the player’s performance with, and partially because the game is designed to make a comprehensive victory in all areas impossible, even by most subjective standards — you must weigh your own values against each other, and make choices.

From my meagre experience developing games, I can’t help but be impressed by Hidden Agenda. There are very few logical hiccups, despite the extremely open-ended storytelling — the game appears to have been tested pretty thoroughly. The quality of the writing and the limited black and white art and animation is high. The game has a noticeable bias toward the left, but other than the seemingly constant inability of right-wing ministers to acquit themselves rhetorically, and the general tone of the end-game descriptions of your accomplishments, it does not greatly affect the game, and in all cases the outcomes of your actions seemed generally believable under the given scenario. Ultimately, what’s most startling, however, is that Hidden Agenda, released in 1988, managed to do what game designers today still struggle with, though many still strive for it — it gives real moral choices to the player, choices that matter, and choices for which reasonable people may disagree on the best option. It’s also one of the best persuasive/serious games I’ve ever played, despite it predating the entire movement. For me, Hidden Agenda evokes an image of an alternate reality in which an entire evolving genre is modeled along the same lines of dynamic storytelling and strategic and moral depth.

Hidden Agenda is currently distributed by the game’s author, Jim Gasperini, as described at the web site of the Maricopa Center for Learning and Instruction.

Read Full Post »

I’ve been reading a lot about permadeath over the last couple weeks, from various blog posts and discussions. The discussion of the feature has been percolating in the blogs I’ve been following for over a month; in particular, Andrew Doull wrote about it here in an article that showed up on Gamasutra two weeks later. The comments there were an interesting contrast to those following the blog, as they represent a more mainstream audience; Andrew Doull is a roguelike developer, so his audience (myself included) skews toward favorable views of permadeath. One of the comments on the Gamasutra article, by Aaron Knafla, caught my eye:

The method of game design outlined here is known (now and forever) as “cult” or “indie” gaming. There’s a very good reason for that; most people are not masochists.

There’s more to the comment too; I just excerpted the part that struck me and is relevant here. While I’m not sure what precisely in Andrew Doull’s article invited this particular comment, I suspect it may have a lot to do with the mention of Dwarf Fortress’s striking philosophy that “Losing is Fun” — a seemingly masochistic assertion that is displayed as a reminder every time you start a new game. Regardless, it got me thinking about the sentiment that permadeath is an inherently masochistic mechanic. I think this is an idea that comes out of a simplistic understanding of what permadeath is, and more importantly, how it’s used.

Mechanically, permadeath is some failure state in which all material, mechanical gains the player has acquired over the course of play are lost, and the internal game state (aside from score boards) is reset. Another way to put it is that permadeath represents complete failure; you lose the game. You can try again, but that’s it, you lost. Standard death in a game, which may display a “you lose” screen but allow restoring a saved game, is always just veiled partial failure.

Many games have this mechanical permadeath, though it’s less common than it used to be. Arcade-style multi-level shooters, brawlers, and platformers may have a lives and continues system even today, in which you have several chances to make mistakes for partial failure before you’re judged to have totally lost the game. My own Politician Paul, as a steeply difficult single-level shooter, has this total failure implicitly; every time you die, you start again from the beginning, with no legacy benefits from your last play. There’s nothing crazy or hardcore about this; at the extreme, a game like Politician Paul is only a few minutes long from beginning to end, so complete failure in that game is no more repetitious or lossy than partial failure in longer games.

But permadeath usually refers to larger losses than just restarting the sole level in the game; it means dropping back to level one from level four, losing your hard-earned equipment, and reverting material gains made only after hours of play, thereby creating a sort of treadmill in which it becomes very difficult to ascend the “ladder” of material progression in the game.

If the game has a linear story or mission structure, this can easily be the death of the game as a source of entertainment. In the movie Groundhog Day (see Jonathan Morin’s talk on predictability in games, discussed last post), Phil Connors is fairly quickly driven to repeated suicide to attempt to get out of the endless cycle of the same repeating day. Instead of suicide, Players will just turn the game off in utter frustration — a sort of complete failure for the designer. If you have permadeath in this kind of game, this is where seeming masochism comes into play.

But Roguelikes, which seem to be the nerdy poster child for permadeath, don’t have this kind of linear story or mission structure. They do have dungeon levels, but it’s so quick to get to the next dungeon level that this is an insignificant loss. The actual loss is in the accumulated experience and gear — loss of power, as opposed to losing your place in a content progression. For these games, the enjoyment in the game is primarily from the strategic and tactical play emerging from the game itself, rather than from consumable one-shot content that forces the player to progress if they wish to continue to be entertained. Power is gratifying and motivates play, but is unnecessary to enjoy it. Left4Dead single player, for example, could tolerate permadeath, because its level progression is rewarding but not absolutely necessary; Half-Life 2 single player could not.

For some people, it’s probably too much trouble to worry about permadeath. There are successful models for games with very light death penalties, and the alternative of trying permadeath is fraught with danger, without guaranteed benefits, so why worry about a risky idea like permadeath? If your only goal is to advance your games as a business, I can understand that. For me though, that seems like denying yourself knowledge, and limiting your understanding of games and therefore you ability to synthesize new designs that go beyond parroting existing conventions.

While Dwarf Fortress’s “Losing is Fun” may sound like a mantra of masochism, it’s really more of a recognition of the fact that through massive and proficient use of emergent gameplay and procedural and user-generated content, Dwarf Fortress is actually still fun even when you lose, to such an extent that losing becomes part of the play experience and the game’s emergent narrative. There are no doubt some masterful design ideas that you can learn from something like this, as that’s something most commercial games can only dream of accomplishing; how much are you missing as a designer if you write it off as a cultish, minority interest?

Unless you’re trying to make something akin to a bullet hell game based on exacting mastery of specific patterns, any usage of permadeath should not feel masochistic. If it does, there’s something wrong.

Read Full Post »

Back in spring of this year, I was taking a class on 3D level design that focused its practical work on building a series of levels for Half-Life 2. Our final project was to have the entire class make a campaign of individual levels, with students being divided into teams of three or so people, each team assigned a different level in the campaign. My team was assigned the final level, in which all weapons and enemies were available, and the pumped, final conflict was to take place. It could have been the most exciting level, or easily the most catastrophic failure.

We decided to implement it as a sort of “last stand” level, in which you and many allies would mount a defense of a street intersection against wave after wave of mixed enemy attackers, culminating in an attack of overwhelming force which would segway into the closing cinematic. Hanspeter Ziegler designed the street and building layout such that we had multiple enterable buildings with internal rooms, stairs, and windows to shoot out of, and I set up the positions of friendlies and enemies, laid down the AI node graph, and set up the attack patterns that would let enemy troops sweep from building to building.

Our team worked early, hard, and with intent. We cared about what we were doing and we wanted it to be fun. I spent many hours playing our level, finding bugs in pathfinding or odd geometry, areas of the flow of battle that could be improved. Despite the instructor being concerned if we had too much ahead of us, we kept to schedule and had time to refine the balance, optimize for performance, and add elements of polish.

In the end, we had a level that was a lot of fun. It was open-ended — you had hard points, medics, and weapon caches distributed throughout the level in different buildings, and waves of enemies that could be dealt with using a wide variety of strategies and weapons. Your actions would determine whether NPC allies survived or fell.

I remember mentioning to the instructor that I was really happy with how our level was turning out. I even liked it more than playing the base game of Half-Life 2. She suggested part of that might be pride; I disagreed, because I already knew what it was. Half-Life is an extremely linear, refined game. Our level was an open-ended, arena style level, where every play was new, and you used tactical strategy to learn the terrain and experiment with strategies.

Just today I was watching Jonathan Morin’s talk about predictability in games and how to avoid pre-conceived experiences in game design. He showed a clip from Groundhog Day — a great movie, but the clip showed Phil Conners bored out of his mind as he recited events before they occurred and robbed an armored car without risk of failure, because he knew the exact moments people would turn away and knew his efforts would succeed. Jonathan Morin then talked about the problem of games being just like this with repeated plays — too predictable after the player goes through the same sequence again and again. This sentence resonated very strongly with me:

[Game designers] want to control what the player does, we want to control when he’s going to do it, we want to control why he’s going to do it, how he’s going to do it, and what he is going to feel when he’s doing it.

It was an extremely well-chosen video clip that really captured the problem with highly linear game structures, and that incisive concern seemed to capture at least my personal impression of some of the highly polished, highly refined linear levels such as are found in Half-Life 2. A specific experience, a specific difficulty, a specific set of challenges. This experience might be calibrated to be incredible, but it’s still only one experience. I like these levels… once. And that’s if the designer did a good job. But I prefer a level that lets me play with strategies and tactics, that lets me experiment and decide how I want to play.

Our Half-Life 2 map let you choose if you wanted to snipe from rooftops, shoot from the fire escape on the third floor into the street below, or fight in close quarters to hold a stairwell in a storefront shop with three friendly NPCs and a medic. If one tactic wasn’t working, you could try another. If you ran out of ammo for one weapon, you could switch to a different tactic, or sprint across the street to the hard point across the way and try to hold that one. Our level encouraged intentional, strategic, and improvisational gameplay that changed with every play through.

As a player, I find those experiences more fun the first time, and far more replayable in successive times. As a designer, I find levels and mechanics that create these experiences to be more challenging to create, but far more satisfying.

Jonathan Morin was the lead level designer on Far Cry 2, a game I’m currently playing, and one that epitomizes the kind of dynamic, player-driven gameplay I’ve described here. His blog is the Game Design Cave.

Read Full Post »