作者：Ian Schreiber

就如我们看到的，玩家决策是游戏制作的核心。有些游戏的故事和玩法紧密交错。就这些游戏而言，玩家决策不仅会影响游戏的机制结果，还会左右游戏的故事叙述。（请点击此处阅读本系列第1、第2、第3、第4、第5、第6、第7、第8、第9、第11、第12、第13、第14、第15、第16、第17、第18课程内容）

对某些游戏设计师而言，真正的“互动故事”是游戏的必杀技。事实上，我们常常无法带给玩家这样的感觉：在自己创造的有趣故事中，他们是主角。我们过去如何叙述互动故事，未来要如何创造更优质的故事内容？这是尚待解决的问题，但我们至少可以谈论某些众所周知的基本要素，这就是我们今天的主题。

多数棋盘游戏都缺乏稳固的嵌入式叙述故事，所以本文主要讨论视频游戏及桌面RPG游戏的故事叙述。但有些现代棋盘游戏试图将桌面棋盘游戏体验同RPG模式结合起来，包括玩家互动玩法中的故事元素。

推荐读物

* Noah Falstein的《A Point of View》。我们将故事叙述视角划分为“第一人称”或“第三人称”——我们也以此谈论视频游戏视角。因此我们很容易会认为二者存在关联性，但其实并非如此。文章清楚阐述二者存在的差异。

* Chris Bateman的《Diversity in Game Narrative》。这主要概述游戏存在的各种故事结构。

* 《Challenges for Game Designers》第13章。

我们可以简单就游戏整体结构划分故事内容。故事结构由玩家享有的各种选择、选择的开放性或局限性及这些选择给故事带来的影响决定。每种结构都有利弊，下文我们将进行讨论。

线性模式

线性故事是传统的故事叙述模式，其包含某些不会影响故事的玩法元素。在这种情况下，故事和玩法属于独立体，因为故事没有选择，而玩法通常包含某种决策过程（游戏邦注：否则这就是故事，而不是游戏）。他们可以在主题上有所联系，故事可以影响玩法，但玩法无法影响故事，因为故事只有一个。

线性故事相比其他故事结构存在一大优点：能够轻松采用已有数千年历史的传统故事叙述方法。这类故事存在强烈情感冲击——我们依然将《堕落星球》的Floyd之死及《最终幻想 7》的Aerith之死作为游戏故事叙述发展史中的关键时刻，虽然这些事件都不在玩家的控制之中。

线性故事由于缺乏决策元素存在明显弊端，其游戏色彩并不突出。就如上面说到的，线性故事和游戏机制存在天然屏障，这会限制故事的效果。

分支型

若我们想要添加决策因素，最明显的办法就是在线性故事中添加各种决策点。当玩家到达特定地点时，他们需决定怎么选择，然后故事就会沿某路线发展，直至到达另一选择点。典型例子就是Sega Genesis的《梦幻之星III》；主角有两次在两位女孩中选择一位作为结婚对象的机会，这就带来4个分支路线，每个路线都有自己的故事和结局。

分支故事的优点是具有互动性。若开发者设置众多选择，覆盖玩家的所有期望，游戏就能够有效回应所有玩家决策。乍看之下，这似乎是游戏叙述的最终解决方案，因为这能够应对所有问题。

但分支故事存在一大缺点：成本很高。就两种选择而言，《梦幻之星 III》故事创作者就得编写4种故事。若是存在第三种选择，他们就得编写8种故事，而10种选择则需要1024种故事！想想典型策略游戏玩家所做的决策数量，你会发现编写分支故事任务瞬间变得难以驾驭。

更糟的是，只玩一次的玩家就无法浏览所有游戏内容。要查看所有游戏发展路线需进行多次重复体验，即使如此，玩家还需判断哪个是真正的故事，哪些是没有发生的额外故事情节。

平行型

用Bateman的话说，这是自行分解的分支故事，允许玩家做决策，最终将全部内容压缩成若干命令事件。例如，在《Silent Hill》中，玩家享有若干选择， 或发展故事，或沿途呈现某些额外故事元素，这些都会影响结局。但无论玩家有没有进行额外操作，有些事件他们无法回避。

平行路线通过保持玩家决策及控制故事数量解决分支模式叙述存在的问题。所以，乍看之下这似乎是最根本的故事结构。

也许你早猜到，此模式依然存在一个问题：由于玩家被迫面对某些事件，整个故事情节就变成线性模式。游戏丧失玩家主导感觉，因为无论他们进行什么操作，某些故事内容依然会出现。

一个潜在解决方法是，让玩家决策改变游戏结局。玩家依然会遇到同样的故事情节，但最终结果由玩家所做决策决定。遗憾的是，因果关系将很快丧失——玩家决策只有到尾声阶段才会显露出来，玩家对特定结局所造成的影响通常非常模糊。

此外，此结构还存在这样的问题：玩家需反复体验整款游戏数次才能获悉所有结局。

交错型

Bateman用此形容分解成小块内容的故事，也许还有若干故事情节（游戏邦注：这些故事也许相互交错）同时进行。玩家选择以何种顺序进行，采用哪个路线。例子之一就是《魔兽世界》，其中玩家能够选择任意关卡，体验不同故事元素。另一例子是《上古卷轴》系列，玩家能够基于自己的期望选择特定故事情节，玩家会沿途操作其他关卡或次要情节。这些次要情节也许会给主要故事情节带来影响。

交错叙述故事的优点是富有表达性。设计师可以选择让各种故事情节同时发生，就像《黑色追缉令》和《真爱至上》之类的非线性电影。

此外，故事也许还会有各种开端、主体和结局，但玩家能够接触所有内容，能够以任何顺序体验任何内容组合，所以强制性重复体验问题便迎刃而解。若玩家足够心思缜密，只要通过一次体验就能够浏览所有游戏内容。

这依然存在一个问题。由于某些事件会影响其他事件，通过故事查看各种可能路线的过程要比分支叙述模式复杂许多。

编写交错内容也颇有难度，事件能够以任何顺序发生，因此就存在玩家以不合理顺序操作内容的可能。故事创作者需确保玩家所访问的故事事件富有意义。追踪故事内容活跃性的决定变量就会立即变得非常复杂。

最后，交错故事存在混淆玩家的危险，若同时存在过多个并行情节，玩家就无法立即辨别它们之间的关系。这也是书籍、电影试图同时讲述多个故事所面临的风险。

面向对象的动态叙述

此术语源自Bateman对《Façade》的描述。意思是游戏存在若干迷你故事，每个故事都有特定进入点和退出点。单个迷你故事的退出点会引出最终结局，或另一迷你故事。迷你故事可看作是书籍的“章节”，戏剧的“幕”。

这类故事有平行模式的优点，但缺乏先线性故事的特色。每个迷你故事都有自己的选择，整个迷你故事集合就像是更大分支或平行故事。每个迷你故事都有独立性，这能够减少编写完整故事所需的时间。

这类游戏存在两个弊端。首先，这类故事依然存在强制重复体验问题：玩家需体验游戏数次方能获悉所有故事路线。这也是试验中的故事结构，目前这类游戏数量不多，不足以让我们分析其利弊。《Façade》是由若干计算机博士共同完成，所以这不是传统故事创作者能够轻松驾驭的故事结构类型。

视角剖析

视频游戏的摄像头或第一人称，或第三人称。

摄像头视角

第一人称是指摄像头置于主角的额头前。玩家所见即角色所见。这促使玩家觉得自己同角色的联系更密切，因为可以这么说，玩家就处在角色的头部里面。此模式的缺点是，这不是真正的第一人称视角，因为人类的周边视觉要比屏幕宽许多，有时这会损害玩家的沉浸性，因为他们不习惯此处视野的局限性。

而在第三人称摄像头视角中，玩家看着主角，通常从角色背后切入，这样你就能够看到角色的背部。这让玩家得以更广泛地浏览游戏空间，向玩家提供的角色视觉信息更逼真。当然通过保持注视角色，玩家会感受到他者性——没有镜子的帮助，我们无法看到自己的背部，所以摄像头视角提醒我们主角不是自己。

将此同麦克劳德的图标&现实主义理论结合起来，我们会发现第一人称角色更接近图标，更像是个“白板”角色，其中玩家能够将自己的个性影射到上面，而逼真的第三人称角色就有强烈的特征描述。

假设存在第二人称摄像头视角，它主要指玩家从正面观察主角。毋庸置疑，这会令多数游戏陷入混乱局面。最贴近的例子就是晦涩的掌机游戏《Robot Alchemic Drive》，其中玩家控制人物角色（第三人称），而角色反过来会通过遥控装置控制庞大机器人（第二人称）。从此意义上看，机器人就是主角，游戏就是采用第二人称摄像头视角。

故事视角

谈论文学时，我们也会使用“第一人称”和“第三人称”。这存在不同之处。

第一人称故事是指叙述者以自己的角度叙述故事。主角直接面对读者，告诉读者自己所遭遇的事件。

第三人称故事更常见，是指故事以旁观者角度切入。这也许是“第三人称全知视角”，其中读者能够看到发生的所有事件，甚至读出多个角色的想法，而在“第三人称局限角度”中，有些信息则不为玩家所知。

文学中鲜有的故事类型是第二人称模式，这里的故事从读者视角切入。这类故事非常少见，因为以可信角度编写此故事颇具难度。

你也许会发现此刻几乎所有游戏叙述都基于第二人称。这就是玩家控制角色会出现的情况，这种情形很常见。

这也是故事叙述颇具难度的另一原因。

互动角色

有时游戏的整个故事脉络已确定，呈线性模式，但依然存在若干供玩家进行互动的角色。在视频游戏中，人际关系通常需要量化，NPC同玩家的关系有两种常见模式。

标记

此背景下的“标记”具有二元值。某活动或发生，或未发生。玩家同某NPC或交谈，或没有交谈，若他们有进行交谈，他们或选择，或没有选择某沟通路线。因此，特定新NPC或者沟通选择，或出现，或消失。

标记的优点是操作简单。所有情况或发生，或没有发生，这简化玩家的操作逻辑。这同时也非常容易在代码中落实；程序员能够轻松制作二元逻辑。

缺点是若这些角色非黑即白，中间没有灰色元素，便缺乏足够深度。设计师可以通过结合二元值提高复杂性，但此时操作会变得非常复杂，丧失之前提到的简单性优点。

亲密性

亲密性不是基于是/否二分法，而是通过数值测量角色对另一角色的喜爱或厌恶程度。若你需要判断某NPC是否以特定方式表现，不妨查看其亲和力值是否超越特定界限值。

亲密性机制的优点是其依然保持简单，比标记机制更能应对复杂内容，更富有表达性。

但开发者需谨慎使用亲密机制。他们得到的也许是非常模糊的机制，尤其是当亲密变量或随机摇骰机制参与决定最终结果。总之，我们很难在未进行广泛测试的情况下，“准确把握”此机制。

角色对话

有些游戏包含对话机制，角色具有话语选择权，所瞄准的NPC会给出相应回应。

我们可以通过相同方式思考游戏叙述机制。对话可以是：

* 线性模式。我走进NPC，选择“交谈”指令。它们告知我某些情况。

* 分支型。我发起对话。NPC陈述某事，我随后享有系列回应选择。根据我的回应，NPC也许会陈述其他内容，给我全新的系列选择。

* 平行型。我发起对话。NPC陈述某事，我享有系列回应选择，然后它们做出相应回应。但我可以通过若干回应组合得到相同结果。

* 交错型。我发起对话，可通过各种方式开始谈话。NPC做出相应回应，我在对话中采用的每个行动都会开启新的对话路线，关闭不相关的对话主题。

* 动态型：我同时经历分支或平行对话。我基于结果继续新对话。各模式下的对话编写难度同对应故事编写难度成正比。

构造优质角色

有些故事角色非常“丰满”——它们有很多层面，深刻的角色通常富有层次，非常精细，会随故事的发展而发展。

有些角色则非常“单调”。单调角色是否总是不好？其实不然；不是所有角色都要复杂、深刻或定义清楚。即便是在莎士比亚的戏剧中，也有某些非常单调的次要角色，但主要正反角色要非常丰富。

一般原则认为，角色需随故事的发展而发展。因此，“放映时间”较少的次要角色需保持单调，因为它们的发展时间很短；常见角色需要富于变化。若你希望创造强烈角色，设计单调主人公不是个好主意。

原型 & 固定模式

我曾在采访中被问道：能否描述“原型”和“固定模式”及二者的差异。

McKee认为故事主要涉及模式，而非公式。原型是模式，而固定模式是公式。

“恶棍”是原型，而Snidely Whiplash样式，留小胡子，戴高顶帽，穿黑斗篷，自私自利的坏家伙则是固定模式。

原型作用很大。它让我们得以基于符合熟悉模式的可信角色叙述故事。固定模式则使用过度，通常会让故事缺乏可信度，因为用户之前已从其他故事中看到类似角色，除非此设计有非常充分的理由。

Freytag的三角模式

这也许曾出现在你小学语文课堂中。其理念是故事通过少量描述奠定基调，然后逐步融入上升动作，内容开始变得日益强烈，接着就达到高潮，然后随着尾声阶段的到来，故事开始融入下降动作和解决方案。

这不仅适合故事叙述，玩法亦是如此。游戏包含陈述内容、上升、高潮、下降和解决方案。

若故事和玩法的情节脉络保持协调和同步，我们能够得到更富戏剧色彩的体验。实际操作要困难许多；在许多游戏中，最困难的玩法通常发生在游戏中间，也就是敌人和挑战变得更加困难，而你尚未找到新武器、新升级渠道的时候。随着玩家的日益强大和熟练，游戏的挑战性就会逐步降低，虽然此时故事的紧张度日益提高。我们能够基于大量的戏剧元素和预兆判断游戏何时进入最终boss战斗，然后玩家挥舞宝剑几秒钟后便获得胜利（游戏邦注：这听起来有些虎头蛇尾，缺乏可信度）。

总之，设计师需在平衡游戏的过程中留心故事内容，及故事戏剧情是否适合玩法的戏剧化时刻。

游戏依然是主导

记住，作为游戏设计师，你的主要任务依然是制作游戏，而非故事内容。在多数情况下，故事要能够兼容玩法。至少，你要在通过牺牲玩法成全故事发展时保持谨慎，或是当你计划向玩家呈现过多非互动故事内容。

创作者在撰写故事叙述内容时要时常回归到根本美学效果，确保故事能够支撑游戏的美学画面。

所学经验

让传统线性故事融入更多互动元素的方法很多。各种方法都有其利弊。

我们应不断进行尝试，探索各种适合游戏的非线性故事叙述模式。（本文为游戏邦/gamerboom.com编译，拒绝任何不保留版权的转载，如需转载请联系：游戏邦）

Level 10: Nonlinear Storytelling

Last time, we learned some basic linear storytelling principles, as told to us by people that worked with books, plays, and movies. And this is fine and good for games that have a linear story. Many video games work this way, where the story is essentially told as a movie broken up into small parts, and the player has to complete each section of the game to see the next bit of movie. I do not mean this in any kind of derogative way; many popular games work like this, and many players find these games quite compelling. Even personally, I have had times when I would be messing about in the subscreens optimizing my adventuring party, only to have my wife call from across the room: “stop doing that and go fight the next boss so you can advance the plot, already!”

However, not all games are like this. As we’ve seen, player decisions are the core of what makes a game. Some games have a strong narrative intertwined with gameplay. For these games, it would make sense for player decisions to not only affect the mechanical outcome of the game, but to affect the narrative as well.

For some game designers, a true “interactive story” is something of the Holy Grail of games. In reality, we often fall short of giving the player the feeling that they are actually the starring role in a compelling story of their own creation. How have we told interactive stories in the past, and how can we make better ones in the future? It’s largely an unsolved problem, but we can at least cover the basics of what is already known, and that is our focus today.

Note that most board games do not have strong embedded narrative, so this entire discussion is mostly relevant to video game narrative, as well as tabletop RPGs. However, some modern board games are in fact attempting to combine the tabletop board game experience with that of the RPG, including story elements that the players must interact with as part of the gameplay.

Course Announcements

I will be at SIGGRAPH all next week. While I will make every effort to post on time, I may not be able to respond quickly to email and I may be slow in validating blog comments or new forum accounts, so please be patient. Naturally, if you are attending there yourself, feel free to say hello in person – I’ll be speaking on Monday morning about the results of the Global Game Jam.

Readings

Read the following:

* A Point of View, by Noah Falstein. We talk of the point of view of a story as “first person” or “third person” – and we also talk of the point of view of a video game in similar terms (First Person shooter, Third Person stealth, etc.). It is easy to assume that the two are equivalent, but they are not. This article makes the differences clear.

* Diversity in Game Narrative, by Chris Bateman. This is a brief overview of the different kinds of story structures encountered in games.

* Challenges for Game Designers, Chapter 13 (Designing a Game to Tell a Story). This short chapter covers today’s topic and can also serve as a review of some topics from last time.

Kinds of Stories

We can roughly classify different stories by their overall structure. The structure is determined by what kinds of choices are available to the player, how open or constrained those choices are, and what effect those choices have on both the ongoing story and the final ending. Each structure has its advantages and disadvantages, which we will discuss below.

Linear

Linear stories are the traditional narrative, with some gameplay elements thrown in that do not affect the story. In this case, the story and gameplay must be separate entities, because the story has no choices and the gameplay must include decision-making of some kind (else it is just a story and not a game). They can be thematically linked, and the story may influence gameplay (perhaps when a certain pre-scripted story element happens, it causes a new gameplay effect to come into play), but the gameplay cannot influence the story because there is only one story.

Linear stories have one major advantage over all other story structures: it is easy to apply traditional storytelling techniques, which have been developed over thousands of years. Such a story can have a powerful emotional impact – witness that we still talk of Floyd’s death in Planetfall and Aerith’s death in Final Fantasy 7 as key moments in the history of game narrative… even though neither events was under player control.

Linear stories have the obvious disadvantage that, due to a lack of decisions, they are not very game-like. As stated above, there is a natural barrier between linear stories and game mechanics, which limits the effect the story can have.

Branching

The first and most obvious thing to do to a linear story, if we want to add decisions, is to add choice points at various places. When the player reaches a certain point, they decide what to do, and then the story goes down one of several continuing paths until another choice point is reached. An example is the old-school Phantasy Star III for Sega Genesis; twice in the game, the main character is given a choice of which of two girls to marry (and then the story continues to the next generation of characters), leading to a total of four branches, each with its own story and its own ending.

Branching stories have the advantage of being interactive. If you include a sufficiently large number of choices and your choices cover all of the things that a player would want to do, the game can respond believably to any number of player decisions. At first, this would appear to be the ultimate solution to game narrative since it can handle just about anything.

However, there is one major drawback of using a branching story: it is expensive! With only two choices (which is not very many), the story writers of Phantasy Star III still needed to write four stories. A third choice would have had them writing eight stories, and including a mere ten choices would require writing 1024 stories! Consider the number of decisions you make as a player in a typical strategy game, and you’ll see that the amount of work to write a branching story can quickly explode into something unmanageable.

To make things worse, note that a player that goes through the game once will never even see the vast majority of content. It requires multiple replays just to see every path through the story… and even then, the player must decide which is the “real” story and which are the alternate timelines that didn’t actually happen. (If the developers ever make a sequel to the game, they must deal with this problem as well.)

Parallel Paths

This is Bateman’s word for a branching story that collapses in on itself, allowing the player to make choices but then collapsing all of them eventually into several mandatory events. In Silent Hill, for example, there are several choices the player can make that may advance the story or reveal some additional story elements along the way, and these will influence the ending. However, there are certain events that the player is forced to encounter no matter what else they have or haven’t done.

Parallel paths solve the problem of branching narrative by keeping the advantage of player decisions while still keeping the total amount of story manageable. So, at first, this would appear to be the ultimate story structure.

As you might guess, there is still a problem: since the player is forced into certain events, the entire plot arc is now essentially linear again. We have lost the player’s feeling that they are directing the story, because no matter what they do there will be certain parts of the story that are the same no matter what.

One potential solution is to have the player decisions alter the game’s ending. The player may still encounter the same plot arc, but the final outcome is determined by the choices the player made. Unfortunately, that means the relationship between cause and effect can be easily lost – the player’s decisions are (by definition) not seen until the very end, and it is often unclear what the player did to cause a certain ending.

And, we still have the problem that the player must replay through the entire game multiple times just to see all the endings.

Threaded

This is Bateman’s term to describe stories that are divided into small pieces, perhaps with several plot arcs going on at once that may or may not intersect. The player then chooses which paths to follow and in what order. One example of this is World of Warcraft, where the player can accept any of several quests to advance different elements of the story. Another example are the Elder Scrolls series of games (like Morrowind and Oblivion), where the player may follow certain storylines based on how they want to advance their character, and the player may find other quests or subplots that they choose to pursue (or not) along the way, and these subplots may or may not have their own effects on the main plot line.

The advantage of a threaded narrative is that it is extremely expressive. You can have multiple storylines happening concurrently, as with nonlinear films like Pulp Fiction or Love, Actually (except, unlike those movies, have the plot lines advance according to player intent).

Also, the story may have multiple beginnings and middles and endings, but the player has access to all of them and can advance any combination of them in any order, so we have finally solved the problem of forced replays. The player can see everything there is to see with a single play-through, if they are thorough enough.

As you might guess, there is still a problem here. First, since some events may affect others, testing out all possible paths through the story can get even more complicated than with a branching narrative. (For programming geeks, a branching story with two choices per choice point is O(N^2), while a threaded narrative is potentially O(N!). Yes, factorial.)

Writing a threaded narrative is hard, because events can happen in any order, leading to the potential for the player to do things in an order that doesn’t make sense (for example, perhaps they are given a quest to assassinate a rival leader in the middle of a war, before the war actually breaks out, or after the war is concluded). The story writer must be careful to allow access to certain story events only when it would make sense to do so. Keeping track of all the variables that determine when a story event is or isn’t active can get very complicated very quickly.

Lastly, a threaded narrative runs the risk of confusing the player, if there are too many concurrent plots happening at a single time and the player does not immediately see the relationships between them. This is also a danger with books and movies that try to tell several stories at once.

Dynamic Object-Oriented Narrative

This last structure is Bateman’s term that, as far as I can tell, he invented to describe the game Fa?ade (which you should absolutely download and play if you have not yet seen it). The idea is that there are several mini-stories, each with potentially several entry points and exit points. A single mini-story’s exit point may lead to a final ending, or to another mini-story. The mini-stories may be thought of as “chapters” in a book or “acts” in a play (except that you may not “read” all of the chapters or may read them in a different order, depending on the choices you make and how you exit each chapter).

This kind of story has the advantages of parallel paths, but without a linear story arc. Each mini-story has its own choices, and the overall collection of mini-stories itself acts like a larger branching or parallel path story. Each individual mini-story is self-contained, which reduces the required time to write the complete story.

This kind of story has two disadvantages. The first is that there’s still the forced-replay problem: a player must play many times to see all of the story paths (which is perhaps why Fa?ade needs to last about ten or twenty minutes, and not ten or twenty hours). It is also a highly experimental structure, so we do not yet have enough games to really analyze what does and doesn’t work in this form. Fa?ade itself took a couple of guys with PhDs in Computer Science to develop, so this is not the kind of story structure that is easily accessible to a traditional story writer.

Points of View

The camera in video games is generally either first-person or third-person.

Camera Views

A first-person view means the camera is, essentially, glued to the main character’s forehead looking forward. The player sees what the character sees. This can lead to a greater sense of connection with the main character, because the player is inside the character’s head, so to speak. The disadvantage is that this is not truly a first-person view, because a typical human’s peripheral vision is wider than the screen (and a typical human can turn their head faster than a typical first-person video game camera), which can break immersion at times for players who are not used to the limitations.

In a third-person camera view, the player is instead looking at the main character, usually from a behind-the-shoulder perspective so that you usually see the character’s backside. This gives a wider view of the area and is more realistic in terms of giving the player the visual information that the character would have. Of course, by looking at the main character all the time, it brings a sense of otherness – you can’t look at your own backside without the aid of mirrors, so this camera view is a reminder that the main character is not you.

Combining this with McCloud’s point on icons versus realism (from last time), we might guess that a first-person character is closer to an icon and tends to do better as a “blank slate” character that the player can project their own personality onto, while a third-person character that is drawn realistically should have strong characterization.

A second-person camera view, if it existed, would involve the camera that spends the entire time looking at the main character from the front. Needless to say, this would be confusing in the vast majority of games. The closest I’ve seen is an obscure console title, Robot Alchemic Drive, where you control a human character (in third person) that is in turn controlling a giant robot by remote control (in second person). To the extent that the robot is the main character, this game uses a second-person camera for control.

Story Views

We also use the words “first person” and “third person” when discussing literature. There, it works a bit differently.

A first-person story is when the story is told from the narrator’s own point of view. The main character is addressing the reader directly, telling you a story of what happened to them.

A third-person story, which is more common, is when the story is told from an outside perspective (much like a typical movie camera, where the view just represents a “fly on the wall” and not an actual character’s viewpoint). This may be “third-person omniscient” where the reader can see things that happen and even hear several characters’ thoughts, or “third-person limited” where some information (such as character thoughts) is concealed from the reader.

A rare kind of story in literature is that of second-person, where the story is told from the reader’s point of view. This kind of story is rare because it is very hard to write in a way that is believable.

You might be realizing about this time that almost every game narrative is told in second person. This is the case whenever the player is controlling the main character, which is most of the time.

And this is another reason why storytelling in games is so hard.

Interactive Characters

Sometimes, the overall plot arc of the game is fixed and linear, but there are a number of characters (specifically, “non-player characters,” referred to as NPCs) that the player can interact with. In video games, where interpersonal relationships must be quantified, there are two common ways to treat NPCs and their relations with the player.

Flags

A “flag” in this context is just a binary (yes-or-no) value. An action did or did not happen. The player did or did not talk to a certain NPC, and if they did, they either did or did not choose a particular conversation path. As a result, certain new NPCs or conversation options may appear or disappear.

The advantage of flags is that it is very simple to do. Everything either happens or it doesn’t, making the logic that drives the characters pretty easy to follow. It’s also easy to implement in code; programmers can do binary logic easily.

The disadvantage is that there is not a lot of depth to these characters, if there is only black and white and no shades of gray in between. You can add more complexity by combining binary values (“only make Aragorn appear if Frodo chose to travel to the Prancing Pony and he successfully avoided capture by the Ringwraith and he puts on the Ring”) but at this point it can get complex to follow, reducing the benefit of simplicity mentioned earlier.

Affinity

Instead of a yes/no dichotomy, use numeric values to measure things like how strongly a character feels affection or hatred towards another character. If you have to decide whether an NPC behaves in a certain way, check to see if its affinity value is past a certain threshold value. (In tabletop RPGs, it is common to also add a die-roll to the mix, so that things may or may not happen based partly on chance.)

The advantage of an affinity system is that it is still fairly simple, it can handle more complexity than a flag system, and it is much more expressive.

However, you must be careful with affinity systems. There is a danger of having a very muddy system (where the player can’t tell why a character behaved in a certain way), especially if several affinity variables or a random die-roll are included in determining the outcome. In short, it is hard to get this kind of system to “feel right” without a lot of playtesting.

Character Conversations

Some games include a conversation system, where the player is given a choice of things to say, and then the NPC they’re talking to responds.

Conversation systems can be thought of in the same way as game narrative systems. A conversation can be:

* Linear. I walk up to an NPC and choose the “Talk” command. They tell me something. That is all.

* Branching. I initiate a conversation. The NPC says something, and I am then given a choice of responses. Based on my response, the NPC may say something else and then give me a brand new set of responses.

* Parallel. I initiate a conversation. The NPC says something, and I am given a choice of responses, then they respond, and so on. There are several combinations of responses that can get me to the same outcomes, however.

* Threaded. I initiate a conversation, and am given a set of different ways to begin. The NPC responds to what I say, and each action I take within the conversation may open up new paths of conversation and close older, no-longer relevant conversation threads.

* Dynamic. I go through a branching or parallel conversation. Depending on the outcome, I proceed to a new conversation (or a new area of exploration within the current conversation).
Writing dialogue in each of these methods is proportional in difficulty to writing a story of each form.

Writing Good Characters

Some characters in stories are “round” – they have many facets to them, a deep character with many layers, and are highly detailed and developed over the course of the story.

Other characters are “flat” (or “shallow”) – they are very simple, don’t have a very deep personality (at least, not that the audience can see), and do not develop or change much (if at all).

Normally if we say a character is “flat” it sounds a bit derogatory. Are flat characters always bad in stories? I don’t think so; not all characters need to be complicated, deep, or well-defined. Even in Shakespearean plays, some minor characters are flat (I’m looking at you, Rosencrantz and Guildenstern), but the main protagonist and villain at least should be round.

A rule of thumb is that a character should develop over time as we seen them through the story. As a corollary, minor characters with very little “screen time” can be flat since they don’t have much time to develop anyway; the characters who we see the most often (generally the main characters) should develop a great deal. A flat protagonist is probably not a good idea if you are trying to make a strong character.

Archetypes versus Stereotypes

Here’s a question I was once asked in a job interview: describe the terms “archetype” and “stereotype” and the difference between them.

McKee says that story is about forms, not formulas. Archetypes are forms; stereotypes are formulas.

“Villain” is an archetype. The Snidely Whiplash-esque, mustache-twirling, top-hat-and-black-cape-wearing, purely-mean-for-its-own-sake bad guy is a stereotype.

Archetypes are useful. They allow us to tell a story with believable characters that fit familiar forms. Stereotypes are overused, and typically make a story less believable because the audience has seen these same characters before from other stories, so they do not seem unique. Generally, avoid stereotypes, unless there is very good reason (such as if your story is a parody that is making fun of a particular character stereotype).

Freytag’s Triangle

You might have seen this before in a grade school literature class. The idea is that stories start with a small amount of exposition to set the tone, then they have rising action where things get more intense, then they reach some kind of climax (a final battle or confrontation, etc.), then there is falling action and resolution at the end as the story reaches its conclusion.

Note that this is not just true for story, but for gameplay as well. Games have exposition (we call it the “opening cinematic” and “tutorial level”), rising action (most of the game), climax (final boss fight), falling action (occasionally some post-fight sequence like a timed escape from the building before it blows up), and resolution (end cinematic).

You get a more dramatic experience if the dramatic arcs of story and gameplay are aligned and in synch with each other. This is a lot harder than it first appears; in many games, the most difficult part of gameplay occurs somewhere in the middle of the game, when the enemies and challenges are getting harder but you haven’t yet found new weapons and abilities to power yourself up to compensate. As the player gets more powerful and better at playing, the challenge of a game often decreases over time, even as the intensity of the story is increasing. You can tell these games when the final boss fight is introduced with this high amount of drama and foreboding, and then the player wins after swinging their sword around for a few seconds – it feels anticlimactic and not very believable.

In short, as you balance your game, pay attention to the story and whether the story drama is proportional to the dramatic moments in gameplay.

It’s Still a Game

Remember that, as a game designer, you are still making a game and not a story. In most cases, the story should not preclude gameplay. At the very least, you should be extremely careful when you are tempted to make a concession in gameplay for the sake of the story, or when you plan to overload the player with non-interactive story bits.

When writing the narrative for a game, return frequently to your core aesthetic – your overall vision of the optimal experience your game will offer – and make sure that the story supports it!

Lessons Learned

There are lots of ways to take a traditional linear story and make it more interactive. All of these have advantages and drawbacks. If the Holy Grail of a full interactive story is even achievable, we have still not discovered how.

But we should keep trying, and exploring the various non-linear story forms that are uniquely suited to games.（Source：gamedesignconcepts）