作者：Simon Strange

从根本上来说，所有系统都处于以下3种状态之一：成长、衰退或平衡。

电子游戏可以视为系统中的系统，开发过程中都会出现成长和衰退状态。随着系统的增加、移除和调整，游戏愈发趋近于其最终形态。在发布之时，游戏系统达到平衡。

当然，设计师不能单凭一时兴致来改变和调整游戏系统。开发早期便应当认识到目标平衡点（游戏邦注：游戏系统的最终状态），这样在开发期间便可以将目标锁定在所需的个体系统（游戏邦注：以及个体系统中提供支持的资产）上。这是种非常实用的降低风险和管理项目的方法。

不幸的是，这个趋势会使设计师发挥作用的能力与开发时间相冲突。这会降低项目后期设计师投入到游戏系统中的效能，这确实令人倍感沮丧。

过去数年来，我开发出某种识别和定义低风险设计改变的系统。我的目标是，允许设计方案在项目多数开发时间内发生改变，而无需执拗于坚持之前计划的设计元素。

以系统化的方式识别低风险选择，使用图表和视觉教具（游戏邦注：本文第2和第3部分内容将具体阐述这两方面内容），我可以预先向开发商和发行商解释为何某些改变对项目的长期稳定性毫无影响或影响不大。这种方法让

我有更多时间来优化游戏的核心系统，最终得到更好、更平衡和更精致的产品。

第1部分：确定张力

“平衡张力”是关键概念。平衡系统可以感受到各个子系统产生的影响，但是每个子系统的影响都会被另一个子系统的影响所抵消。两个方向相反的子系统构成的系统，这是最简单的例子，但多数情况下必须考虑到多个子系统的结合。这与物理课上绘制的力学图表很相似。

假设桌面上有块砖。砖所承受的万有引力同桌面的支持力方向相反，所以砖块可以保持平衡（游戏邦注：如下图所示）。现在，假设你将左手放在砖块的左侧。如果你按压砖块，砖块就会向右侧滑动。如果你将双手放在砖块两侧，并施加相同的力，又可以重新构建起平衡。要点就是，砖块可以在多个力的作用下保持平衡，只要这些力可以相互抵消。

但是，这并不意味着所有的平衡状态都是相同的！“受挤压”的砖块无疑感受到来自你双手的张力。同样，通过改变游戏平衡状态的“张力”，你可以使电子游戏的体验发生显著的改变。

假设2D平台游戏中某个角色便是上文中的砖块。玩家可以将砖块向左或向右移动，跳过小障碍在游戏世界中进展下去。如果玩家放下手中的控制器，这个砖块也就停止了移动，就像你从双手从真正的砖块上移开，让它静止地放在桌面上。

现在，让我们添加个子系统，开始每3秒从屏幕右侧投掷火球。火球算是股新出现的力量，如果不得到平衡，就会打破砖块的平衡状态。要平衡这个力量，玩家必须在火球出现是按后退键跳开。只要玩家做出正确的跳跃动作，

游戏仍然会保持之前的平衡。不同之处在于，现在玩家需要主动做出动作来实现游戏的平衡。我们对玩家动作的要求越多，在平衡状态上施加的张力也就越多。

接下来我们来看看，在某些著名游戏中张力是如何实现增加或减少的，对于这种增加或减少是否产生有利影响暂不作讨论。

传送门

《传送门》的核心系统可以浓缩为两个问题：我要在何处放置蓝色传送门？我要在何处放置橙色传送门？

在《传送门》中，并非每个墙面都光滑到足以让玩家开启传送门。因为通往各个房间的可行解决方案随可开启传送门地点数量而增加或减少，所以减少光滑墙面的数量可以让玩家更快地找到解决方案。也就是说，较少的光滑墙面就等同于较低的张力。

《传送门》现在的设计是，不会对玩家创造过多传送门进行惩罚。

但是，假如我们跟踪每个房间中的光枪使用次数，比如限制枪支的弹药或告诉玩家我们期望他们能够有更好的表现（游戏邦注：比如玩家射击16次，而系统预期为8次），那么玩家就会将每次射击视为需要管理的资源。管理的资源越多，就意味着张力越高。

洛克人

《洛克人》游戏中总是会设计较为困难的连续跳跃。在多数情况下，跳跃失败就意味着需要从头开始整个系列的连续跳跃。

但是在部分地点，跳跃失败就意味着死亡。跳跃让你有通过关卡的机遇，但跳跃也会导致你在关卡中的失败。

成功跳跃就不会导致角色的死亡，这种设计只是对失败的惩罚。减少此类情况可以让玩家在无需为惩罚感到恐惧的前提下掌握充满技巧性的跳跃。这会改变跳跃的“张力”，减少玩家对失败的恐惧。恐惧较少就等同于张力较低。

《洛克人》中充满技术性的跳跃、死亡和有限的攻击选择使该游戏难度适中。但是，游戏对于每个关卡的时间并没有做出限制。增加关卡时间限制会增加玩家的失败可能性，增加游戏中每个动作的张力。时间限制就等同于较高的张力。

异形大战铁血战士

Rebellion于1999年发布的PC游戏《异形大战铁血战士》中有着某些设置简单的关卡，这些关卡中只会出现数个敌人。游戏的睿智之处在于，无论玩家何时加载或重新加载关卡，敌人都会在随机位置刷新，他们出现的位置会被重置，这对当时的FPS来说几乎是史无前例的设计。这意味着尽管你可以熟悉关卡的地理形态，但你永远都不知道敌人会于何时在何处出现。这种创新令人印象深刻，也是《异形大战铁血战士》之所以成为拥有如此高声望的惊悚游戏的重要原因。不可预测性就等同于较高的张力。

《异形大战铁血战士》的另一个创新之举是3个角色间的不对称平衡。Predators和Marines各自都有一套强大的武器，但是他们的弹药都受到严格的限制。这迫使角色不得不思考接下来可能面对的威胁，更换相应的武器。

另一方面，异形完全没有装备、弹药以及其他的限制因素。异形时刻都可以保持100%的攻击力。这正是玩家喜欢扮演异形来玩游戏的重要原因。以异形来玩游戏时，较少的游戏玩法选择就等同于较低的张力。

摇滚乐队

《摇滚乐队》的核心游戏玩法从未改变。这是该系列游戏所需秉承的重要准则，因为公司想要向用户出售更多的音乐，所以后台的兼容性不可打破。但是，《摇滚乐队》成功地在其他方面实现改变，这些都是维持不同张力层次下系统平衡的绝佳范例：

失败。《摇滚乐队》原作的设计是，如果玩家错过的节拍过多就会失败。《摇滚乐队2》设置快速游戏模式，玩家可以在“无失败”的环境下体验游戏，但这不计入“真正的”游戏进程中。《乐高摇滚乐队》允许濒临失败的玩家拯救自己，降低了失败的冲击力。《摇滚乐队3》引入菜单选项，可以将实现不受惩罚的失败。

乐队构成。《摇滚乐队》原作设定各个阶段的乐队成员数量维持不变。《摇滚乐队2》允许玩家带领多个乐队，但是通过主菜单后就可能锁定在所选择的乐队上。《摇滚乐队3》允许玩家在任何时刻改变乐队，增加或移除乐队成员。

成就/胜利标准。为获得所有的成就，《摇滚乐队》原作要求玩家掌握每种乐器，期望所有玩家都能够成为专家级玩家，而且玩家还需要多名相同地区玩家玩游戏。《摇滚乐队2》需要玩家掌握每种乐器，甚至要求玩家连续玩游戏7个小时，期间不暂停、不断开控制器。《摇滚乐队3》的玩家玩联系模式也可以获得奖励，玩每个难度的关卡都可以获得奖励，还跟踪玩家在每首歌曲中的进展情况。

《摇滚乐队》显然正朝着张力越来越低的状态发展。有些人将其视为进步，但有些人并不喜欢这种改变。但是Harmonix成功地实现在不打破核心平衡的前提下调整游戏，通过管理玩家张力层次的方法实现。

在上述列举的这些具体游戏范例中，设计元素让游戏变得更好或更差并非讨论重点。要点在于，这些改变都没有打破游戏已经建立起来的系统平衡。游戏或许会变得更具挑战性，玩家更容易失败或吸引力降低，但从本质上说依然是同一款游戏。我并非旨在宣传应该在游戏中制造出更多或更少的张力，这取决于各个游戏设计团队自己的想法。我只是想说明，通过对张力和系统平衡的考量，我们可以找到不影响项目稳定性的设计机遇，同时实现对玩家游戏体验的改变。

第2部分：绘制张力图表

我已经指出，游戏中某些系统层次的改变可以在较低风险下实现，我已经列出了些许此类改变的范例。但是，我们要如何区分保留平衡的改变和破坏平衡的改变呢？如果没有学习这些结果的合理做法，那么单纯列举范例是无济于事的。所以，接下来我将要分享的是绘制张力图表的各个步骤。

1、尽量识别更多子系统。

2、尽量识别更多的玩家动作。

3、将子系统连接起来。

4、将玩家动作与子系统连接起来。

5、识别未发现的子系统和玩家动作

6、重复第3到第6个步骤。

作为练习和示例，我将在下文中为《毁灭战士2》制作张力图表。以今天的标准来看，《毁灭战士2》是款简单粗糙的游戏，但是游戏有限的系统依然足以让我们完成图表。如果你对《毁灭战士2》已经很熟悉，或许可以自行完成前两个步骤，无需参考我的内容。

《毁灭战士2》的子系统。FPS元素：关卡，计时器，可见度，生命值挑战，弹药挑战，护甲挑战，炮弹，道具收集，敌人，胜利条件，难度设置，加载游戏，保存游戏和死亡竞技。

敌人。战士，散弹枪士兵，纳粹士兵，链炮士兵，小恶魔，恶魔，幽灵，报复者，人球，焰魔，独眼魔，苦痛之源，遗魂，地域骑士，地域爵士，机械巨魔，机械蜘蛛，蜘蛛魔。每种敌人包含以下子系统：移动速度，生命值，攻击力，击退几率。因为游戏支持死亡竞技，我们在这个列表中增加了“玩家敌人”，但玩家敌人的子系统比NPC敌人更多。

武器。拳头，狂暴之拳，链锯，手枪，散弹枪，链炮，火箭筒，电浆枪和BFG9000。

环境。门，锁住的门，移动炮塔，碾压机器，腐蚀地板，爆炸桶。

可收集物品。治疗药水，小急救箱，大急救箱，头盔，背心，弹药，灵魂球，超级护甲，超级球，无敌球，夜视镜，弹药背包，狂战箱，隐形球，地图，防护衣和钥匙卡。

《毁灭战士2》的玩家动作。4个方向的移动，转向，机枪扫射，奔跑，开门，射击和更换武器。

连接子系统：生命值。我以玩家生命值为图表中心开始绘制。我列举出所有与玩家生命值有关的子系统，然后安排它们围绕该机制呈图表式循环。有4种子系统可以减少生命值，7种可以增加生命值。当然，那些系统中多数会产生更多的效果，这些我们也需要呈现出来。在这个初始图表中，向右移动等同于生命值增加，向左移动等同于生命值减少。

有3个道具（游戏邦注：治疗药水、灵魂球和超级球）可以直接增加生命值，因而可以直接将它们绘制在右边。小急救箱、大急救箱和狂战箱只能将生命值提升到上限，所以从某种程度上来说，它们对生命值的提升不及前面提到的3个道具。所以，这些道具被放置在偏右的位置。

护甲不会增加或减少玩家生命值，但是它能够减少玩家受到的伤害，因此也算是对抗4种伤害来源的力量。无敌球也是如此，防护衣也一样，只是效果较差。

连接子系统：可见度。“可见度”指玩家能够看到的事物清晰度。影响可见度的子系统有3个：光照、障碍物（游戏邦注：比如墙）和击退几率。光照影响了玩家能否看到眼前的事物，只受夜视镜的影响。门和墙阻挡了视线，也可防御敌人。

这样，围绕可见度，我们便拥有了两个正交的系统群组。虽然它们并没有互动，但是却能够通过普遍标准连接起来。

增加玩家动作

玩家的移动选择让我们可以绕开墙和门之类的障碍物。我们在每个关卡中的移动最终会引导我们去的胜利。移动还可以让我们躲开敌人发射的子弹。

射击可以削减我们周围敌人的数量，但需要耗费弹药。如果再考虑到每种武器对抗各类型敌人的效用，图表会变得更为复杂，但这部分内容并不是必要的，而且需要用到重叠线或3D模型。因为《毁灭战士2》中的敌人没有抵抗系统，因而所有武器对多数敌人都有效。

图表遗漏的内容

这个图表中不包含来自故事、主题、菜单的任何系统，而且还忽略了玩家期望这个层面。这些内容都很重要，但是它们不会影响到我们在这里讨论的核心系统机制。事实上，看看Raven Software的《Heretic》就会明白，几乎相同的系统设计加上不同的故事系统就能够产生完全不同的游戏感觉。

第3部分：隔离变量

我已经对自己有关系统平衡、张力和如何绘制两者间关系图的想法做了解释。我们也已经看许多具体的范例，单个子系统的调整需要相反方向的子系统同样做出调整才能够保持游戏的平衡。那么，这种分析会如何帮助设计师预测和最小化改变的风险，尤其那些开发后期设计改变带来的风险呢？

要找到这个问题的答案，我们需要考虑每个子系统的张力如何对玩家体验产生影响。从《毁灭战士2》子系统的已完成图表上，可以得出如下结论：

1、玩家生命值是个受众多张力影响的子系统。

2、敌人和玩家生命值通过各种不同的子系统连接起来。

3、与其他子系统相比，胜利的张力相对较小。

4、可见度的张力相对较小，夜视仪可能是图表上影响力最小的元素。

这意味着，改变生命值获得或失去的方式会影响到游戏中几乎每个系统，改变敌人的行为同样也很糟糕。调整这些系统中的任意一个都可能导致制作人不希望看到的重大改变，这也是为何他会逼迫设计师尽早确定系统的原因所在。

胜利的张力很有趣，似乎打败敌人（游戏邦注：从而保住玩家自己的生命值）是个比完成关卡更有驱动力的目标。事实上，在没有故事或关卡的框架下玩《毁灭战士2》（游戏邦注：也就是所谓的死亡竞技模式）是游戏中最有意思的部分。张力图表帮助设计师理解到，故事和线性过程对产品的重要性并不如系统中的攻击和防御。

从张力图表上我们明显可以看到，在《毁灭战士2》的整体体验中，可见度系统是个几乎被设计师所遗忘的元素。如果我是设计师，我会尝试将这个系统完全抛弃，或寻找方法让其更能够影响到核心体验。

科学研究员认为，只有隔离变量，才能够在不受其他元素的干扰下测试变量。为游戏系统创建张力图表可以帮助你识别隔离系统，这些系统在项目开发的任何时候都可以安全地进行重新设计。也就是说，如果你拥有个庞大的系统性问题，比如“游戏难度过高”、“AI设置不当”或“我不知道应当如何做才能提升自己的技能”，那么张力图表在识别正在运行的子系统时便是个极有价值的工具，可以让你了解到哪些设计上的改变有较高风险。

当然，识别出安全的调整只能算是成功了一半，设计师还需要有效地同程序员、艺术师、制作人和客户交流调整范围。此时张力图表的可视化本质也很有帮助，因为图表可以让深入发掘和探讨变得更容易。对子系统的划分可

以让工程师对游戏认识更为清晰，而制作人也希望能够在改动之前对需要改变的元素数量有一定的了解。最棒的是，此类图表在一张纸内就可以绘制完成，所以所有人都愿意去观察阅读！

第4部分：艺术与科学

科学是个协作性过程。数据必须精确记录，实验必须不断在多个独立团队间进行。单个人或者单次实验可能会出错，但是随着时间推移，科学的协作性本质会纠正所有的问题和错误。

艺术则是种强烈的个人化过程。艺术受情感和感情所驱动，艺术的价值取决于不同的人和不同的事例。趋势可能会随时间逐渐显现，但艺术在本质上不受传统、进程或解释说明所限制。

游戏设计的吸引力在于，这个领域里艺术和科学并重。困难的技术问题需要创造性方案才能解决。对于每款优秀的游戏来说，它既包含完全从其他游戏处模仿过来的元素，也包含此前从未出现过的独特系统。

许多情况下，游戏设计的艺术冲动会同已知的科学准则发生直接的冲突。正因为此，许多艺术家才发现自己脱离了社会，在关键的时候发现他们自己是孤立的。

关于我提出的子系统张力图表，这是种深层次的科学化方法。但是，它并不能为我们所有人遇到的任何设计问题提供出解决方案。它的作用在于，为我们提供艺术化表达的自由，同时保持科学化结构的稳定。（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）

Tension Maps: A Process for Identifying Low-Risk Design Opportunities

Simon Strange

All systems are fundamentally in one of three states: growth, decay, or equilibrium.

For a video game, which can be viewed as a system of systems, growth and decay both happen during development. As systems are added, removed, and adjusted, the game more and more resembles its final shape. By the time you ship, your game is (hopefully!) at equilibrium.

Of course, designers cannot simply tweak and tune game systems on a whim. The target equilibrium point (the final state of the game’s systems) needs to be identified fairly early on, so that individual systems (and their supporting assets) can be locked down during development. This is a very practical way to reduce risk and manage a project.

Unfortunately, this tends to create an antagonistic relationship between a designer’s ability to effect change and the amount of development time left. This reduces the designer’s ability to work on game systems during the latter half of the project, which can be very frustrating.

Over the last few years, I’ve developed a system for identifying and defining low-risk design changes. My goal is to allow design changes during the majority of a project instead of being forced to lock down design elements early on.

By identifying low-risk options in a systematic way, using charts and visual aids (which I discuss in parts 2 and 3), I have been able to describe to producers and publishers in advance exactly why certain changes pose little to no risk to the project’s long-term stability. This has afforded me almost twice as much time for fine-tuning our game’s core systems, which has resulted in better, and more balanced, more polished products.

Part 1: Defining Tension

The key concept is “equilibrium tension.” A system in equilibrium feels the effect of many sub-systems, but each “pull” is balanced by an inverse “pull” of equal magnitude. In the simplest cases, this means two sub-systems opposing one another’s effects, but in most cases a combination of sub-systems must be considered. This is exactly analogous to the force diagrams you might have drawn in physics classes.

Imagine a brick resting on a table. The gravitational force on the brick is exactly opposed by the table, so the brick remains in motionless equilibrium (See Figure 1). Now imagine your left hand on the left side of the brick. If you press on the brick, the brick will slide to the right (See Figure 2). If you use both hands, one on either side, and apply an equal amount of force, you can re-establish equilibrium (See Figure 3). The point is that the brick can remain in equilibrium with any magnitude or combination of forces, so long as each force is counteracted by other forces.

This does not mean that all equilibrium states are the same! The “squeezed” brick can absolutely feel the tension from your two hands. In the same way, you can make significantly different experiences within a video game by changing the “tension” on that game’s equilibrium state.

Imagine our brick as a playable character in a simple 2D platform game. A player could move the brick left or right, jumping over small obstacles to progress through the world. If the player puts the controller down to take a break, nothing would happen to disturb the brick, just as you might have removed your hands from the physical brick and left it lying on the table.

Now let’s add a sub-system, and start throwing fireballs from the right side of the screen every three seconds. The fireballs are a new force which, if unbalanced, would “push” the brick out of equilibrium. To balance this force, the player must simply “push” back by jumping over the fireballs as they appear. So long as the player jumps properly, the game remains in the same equilibrium as before. The difference is that the player is now actively working to balance the game’s equilibrium. The more we demand of the player, the more tension we place on our equilibrium state.

Let’s look at a few examples of how tension might be increased or decreased in some well-known games, without passing judgment in regard to whether this increase or decrease would be a good thing.

Portal

The central system in Portal can be boiled down to “Where do I place the blue portal?” and “Where do I place the orange portal?”

Not every surface in Portal is smooth enough for the player to open a portal through. Since the possible solutions to each room increase or decrease with the number of possible portal locations, reducing the number of smooth walls makes the solution more readily apparent to the player. So fewer smooth walls equals lower tension.

As Portal exists now, there is no penalty for creating more portals than necessary.

But if we were to track the number of shots used in each room — by limiting the gun’s ammo, or simply telling the player that we expected them to be more prudent (16 shots/8 expected) — players would become aware of each shot as a resource to be managed. Managing more resources means higher tension.

Mega Man

Mega Man games have always included difficult jump sequences. In most cases, missing a jump means you have to start the sequence over again.

But in a few spots, missing a jump means instant death. Jumps always provide opportunities to progress through the level but jumps over spikes or pits also offer an opportunity to fail the level entirely.

Spikes and pits have absolutely no consequence on a successful jump, as they are simply an extra harsh penalty for failure. Reducing these cases allows players to master the tricky jumps without fear of penalty. This changes the aggregate angle of the jump “tension” and reduces the player’s fear of failure. Less fear equals lower tension.

Mega Man games have always been moderately difficult due to tricky jump timing, instant deaths, and limited attack options. But there have never been strict time limits within individual stages.

Simply adding a countdown clock would give players yet another way to fail, increasing the tension of every action in the game. Time limits mean higher tension.

Aliens vs. Predator

Rebellion’s 1999 PC game Aliens vs. Predator had some wonderfully austere levels, in which only a handful of enemies would appear. The clever twist — nearly unprecedented for a FPS at the time — was that enemies were randomly distributed, and their locations were reset whenever you loaded or re-loaded the level. This meant that while you could learn the geography of the level, you never knew when or from where the enemies would be coming. This was a dramatic departure from the norm, and was a big part of AvP’s reputation as a very scary game. Unpredictability equals higher tension.

Another innovation in AvP was the asymmetrical balance between the three characters. Predators and Marines each had a suite of powerful weapons, but they were all strictly limited by ammunition.

This forced the characters to constantly guess which sort of threat they would be facing at any moment, and to switch weapons accordingly.

The Alien, on the other hand, had no equipment, ammunition, or other limiting factors at all. Aliens were always working at 100 percent effectiveness. This cognitive simplicity was a big source of the appeal for playing as the Alien. Fewer gameplay choices means lower tension while playing as an alien.

Rock Band

Rock Band’s core gameplay has never really changed. That’s actually an important principle of the franchise, because they want to sell additional tracks to customers, and breaking backward-compatibility would be a big deal. But the Rock Band games have managed to change in other significant ways, all of which are excellent examples of maintaining system equilibrium at different levels of tension:

Failure. Rock Band forced players to fail if they missed too many notes. Rock Band 2 allowed quick games to be played in a special “no fail” mode, but separated those sessions from “real” RB2 sessions. LEGO Rock Band weakened the impact of failure by allowing failing players to save themselves. Rock Band 3 introduced a menu option which simply turns off failure at no penalty.

Group composition. RB1 forced bands to keep essentially the same members in every session. RB2 allowed players to lead multiple bands, but locked your selected band once you moved past the main menu. RB3 allows a group to seamlessly change bands at any time, and to add or remove players mid-song.

Achievement/Trophy criteria. In order to get all achievements, RB1 required players to master each instrument, expected all players to play at expert level, and required playing with multiple local players. RB2 required players to master each instrument, and even asked players to play for seven hours without pausing or disconnecting controllers. RB3 rewards you for playing in practice mode, gives rewards at every difficulty level, and tracks individual progress in every song (so you can fail or excel independently of your group).

Rock Band has clearly been moving toward lower and lower tension states. Some people consider that an improvement, while others are less happy with the change. But Harmonix has done an excellent job of adjusting its games without breaking their core equilibrium, and has done so by managing player tension levels.

The point of all these game-specific examples is not that these design elements necessarily make the games any better or worse. The point is that all of these proposals do nothing to break the equilibrium of the game’s established systems. The game might be more challenging, more frustrating, or less compelling, but it would mechanically be the same game. I’m not trying to evangelize creating games with more or less tension — that question needs to be addressed by each game’s individual design team. I’m simply posing that by thinking in terms of tension and system equilibrium, we can identify design opportunities which do not threaten the stability of the project but still impact the player’s experience in significant ways.

Part 2: Drawing Tension Diagrams

I’ve made a point that some system-level changes in your game can be made with minimal risk, and I’ve identified a few examples of such changes. But how, exactly, does one go about separating equilibrium-preserving changes from equilibrium-destroying changes? Without a formal process to repeat these results, it’s meaningless to simply point out examples. To that end, I’ll share my step-by-step process of drawing a Tension Diagram.

Identify as many sub-systems as possible.

Identify as many player actions as possible.

Link sub-systems together.

Link player actions to the sub-systems.

Identify missing sub-systems and player actions.

Repeat steps 3 – 6.

For this exercise, I’m going to create a Tension Diagram for Doom II. Doom II is pretty primitive by today’s standards, but the limited systems will assist in our completion of the diagram. If you’re already familiar with Doom II, you may want to perform steps 1 and 2 on your own without looking at my list. You’re welcome to use online resources and so forth.

Doom II’s sub-systems. FPS fundamentals: levels, time counter, visibility, health counter, ammo counters, armor counter, projectiles, item collection, enemy spawning, victory condition(s), difficulty settings, load game, save game, and deathmatch.

Enemies. Soldier, Sergeant, Nazi, Heavy Weapon Dude, Imp, Demon, Specter, Revenant, Mancubus, Arch-vile, Cacodemon, Pain Elemental, Lost Soul, Hell Knight, Baron, Cyberdemon, Arachnotron, and

Spiderdemon. Each enemy has the following sub-systems: movement, health, attack, aggro. Because the game supports deathmatch, we add “enemy player” to this list, though enemy players have more sub-systems that NPC enemies.

Weapons. Fist, Berserk Fist, Chainsaw, Pistol, Shotgun, Chaingun, Rocket Launcher, Plasma Gun, and BFG9000.

Environment. Doors, locked doors, moving columns, crushers, acid floors, and exploding barrels.

Collectibles. Health shard, stimpack, medkit, armor shard, armor vest, ammo, soul sphere, megaarmor, megasphere, invulnerability, night goggles, backpack, berserk, invisibility, map, radiation suit, and keycards.

Doom II’s player actions. Four-way movement, turn, strafe, run, open door, shoot, and change weapon.

Linking sub-systems. Health (see Figure 4). For no particular reason, I’ve started with Player Health at the core of my diagram. I listed all the sub-systems which interact with Player Health, and then arranged them flow chart-style around that mechanic. There are four sub-systems which decrease health, and seven which increase health. Of course, most of those systems have more subtle effects, which we also need to represent. In this starting diagram, moving to the right means increasing health while moving to the left means decreasing it.

Figure 4 shows the linking sub-systems for Health in Doom II.

Three items (health shard, soul sphere, and megasphere) directly increase health, so they lay to the immediate right. Stimpacks, medkits, and berserk only increase health up to 100 percent, so in some cases, they don’t increase health as well as the first three items mentioned. To indicate this, those items are represented as mostly pulling against Player Health, but also partially pulling against the reduced maximum health for those items.

Armor does not increase or decrease Player Health, but it does mitigate damage, which I have represented by allowing it to “pull” on the four damage sources. Invulnerability (and to a lesser extent, Rad Suit) play a similar role.

Linking sub-systems: Visibility (see Figure 5). When I use the term “Visibility,” I literally mean what can be seen. There are three sub-systems which manipulate visibility: light, obstacles (such as walls), and aggro. Light affects whether the player can actually see something in front of them, and it is affected only by Goggles. Doors and walls block line of sight as well as prevent enemy aggro. Similarly, unspawned enemies cannot be seen.

Figure 5 shows the linking sub-systems for Visibility.

So around Visibility, we have two orthogonal groupings of systems. Though they do not interact, they are linked by a common metric.

Adding Player Actions

All of the player’s movement options allow us to circumvent obstacles such as walls and doors. Moving eventually leads us to victory as we navigate each level. Moving also allows us to avoid enemy projectiles (see Figure 6).

Figure 6 shows Doom II’s linking sub-systems with the addition of player actions (click for full size).

Shooting can (hopefully!) reduce the enemies around us, at the expense of ammunition. I could complicate the diagram by taking into account how efficiently each weapon combats each type of enemy, but that’s not strictly necessary, and would require overlapping lines or a 3D model. Since enemies in Doom II did not have any sort of resistance system, all weapons work reasonably well against most enemies.

Are We Missing Anything?

This diagram does not include any systems from the story, theme, menus, or taunts, and it ignores player expectations which might have been built-up from an overexposure to the original Doom. Those all have an important place, but they generally do not impact the central system mechanics that we are considering here. In fact, you can look at Raven Software’s excellent Heretic to see how a nearly identical system design can spawn a game with an entirely different feel.

Part 3: Isolating Variables

I’ve explained my ideas concerning system equilibrium, tension, and how to diagram the relationships between them. We’ve looked at a few specific examples of how adjusting one sub-system

necessitates an inverse adjustment in another sub-system in order to maintain equilibrium. So how does this analysis allow a designer to anticipate and minimize risk, especially the sort of risk that tends to follow design changes made late in development?

To dig into that, we need to consider how tension on each sub-system affects a player’s experience. Let’s take a look at the completed diagram of Doom II’s sub-systems, and make some observations:

Player Health is the sub-system under a great amount of tension.

Enemy and Player Health are linked through a wide variety of sub-systems.

Victory is relatively low-tension, compared to the rest of the sub-systems.

Visibility is relatively low-tension, and the Night Vision Goggles are probably the most obvious outlier on the diagram.

This means that changing the way health is gained or lost would undermine nearly every system in the game, and changing the behavior or treatment of enemies would be nearly as bad. Adjusting either

one of these systems is exactly the sort of gigantic change that producers try to avoid by forcing designers to lock down systems as early as possible.

The tension around victory is interesting; it seems to say that defeating enemies (and thus defending your health) is a more driving goal than actually completing the level. In fact, playing Doom

II without a story or level progression (that is to say, deathmatch) was arguably the best part of the game. The tension diagram helps the designer to understand that the story and linear progress are less significant to the product than the moment-to-moment systems of attack and defense.

The visibility system — as evidenced by the tension diagram — was a fairly forgettable element in the overall experience of Doom 2. As a designer during production, I might try to get the system cut entirely, or find a way to work it into a more central part of the experience. Consider how Looking Glass approached this in its original Thief games.

Scientific researchers talk about isolating variables in order to test them without interference from other effects. Creating a tension diagram for your game’s systems will help you to identify those systems which are isolated, and are generally safe bets for redesign at any point in the project. That said, if you do have a larger systemic problem such as “the game is too hard,” “the AI is cheating,” or “I don’t know what I should be doing to improve my skills,” then a tension diagram can be an invaluable tool for determining what sub-systems might be at play, and help you to make

those higher-risk design changes in as mindful a way as possible.

Of course identifying safe changes is only half the battle — designers also need to communicate the scope of their adjustments effectively to programmers, artists, producers, and clients. The visual nature of Tension Diagrams helps here as well, since diagrams are easy to digest and easy to talk about. Engineers appreciate the compartmentalization of sub-systems, and producers love enumerating the number of pieces under revision ahead of time. Best of all, a diagram like this fits on a single page — so people will actually read it!

Part 4: Art vs. Science

Science is a collaborative process. Data must be carefully recorded, and experiments must be reproduced by independent teams. One person or one experiment can go awry, but over time, the collaborative nature of science accounts for all errors.

Art, on the other hand, is an intensely personal process. Art is driven by emotion and expression, and the value of art is determined for each individual on a case-by-case basis. Trends can appear over time, but art is fundamentally unconstrained by convention, process, or interpretation.

Game design is fascinating because it is equal parts art and science. It’s dependent upon creative solutions to very difficult technical problems. Every good game contains both wholly derivative elements cribbed from other games, and unmistakably unique systems which have never been seen before.

At many times, the artistic impulses of game design exist in direct conflict to the scientific principles of what is already known. Because of this, many designers are marginalized, become rampaging tyrants, or find themselves isolated at critical moments.

My process for discussing sub-systems in terms of tension diagrams is deeply scientific because it begs for reproduction and refinement. But it does not, at the end of the day, offer any solutions to the design problems we all face. Rather, it focuses our artistic efforts on well-constrained elements in an effort to provide freedom to our artistic expression while simultaneously providing guidance through structure.

Will this process help you? I sincerely hope that it will. Game designers are very much in need of some common processes by which we can compare and contrast our efforts. Have you experimented with some similar process? Have you tried to create sub-system diagrams in the prototyping stage? I would love to expand upon this discussion with all who have read this. (Source: Gamasutra)