挖掘数据提升游戏设计编辑本段回目录
玩家花大量的时间在大型多人在线游戏（游戏邦注：下文简称“MMOG”）中探索最优化的战略。他们从你的MMOG中获得了最大的投资回报（游戏邦注：下文简称“ROI”）。那么，你是否跟他们一样？在这篇文章中，我将向你展示数据挖掘将如何改善游戏设计，然后我将阐述所获信息的4种实际应用：
1、平衡经济
2、寻找作弊者
3、削减制作成本
4、提升用户表现
这篇文章针对的是MMOG，但也适用于其他多人和单人游戏。我在举例时主要使用MMORPG的术语，因为熟悉和理解这个题材内容的人较多。不过在我们开始学习这些技术之前，还是得先理解为何数据挖掘技术是一项有效工具。
data process(from techpubs.sgi)
挖掘数据的原因
因为玩家会撒谎。单靠玩家反馈来审视游戏设计是不全面的。玩家口头反馈所描述的情况可能与现实情况存在较大偏差。在调查或用户反馈中，玩家并没有精确地报告他们自己的行为。他们的说法与做法并不一致。比如，人类学家William Rathje博士调查了一栋房子中住户饮用啤酒的数量，然后去查看他们扔出的垃圾。垃圾中空酒瓶的数量是调查数据的两倍左右。后者比调查更具有洞察力，这就是传统的数据搜集方法。另外，心理和社交行为会让玩家和开发者潜意识地修改自我报告内容。
调查 vs 数据(from gamasutra)
从政治角度来看，玩家和开发者也会修改他们的报告。玩家属于不同的特别兴趣小组，这会让他们的报告产生偏差。政治团体这种人类特征在在线社区中同样存在。只要MMOG中含有公会、家族或其他社交组织，就会出现特别兴趣小组。这些小组中的成员会将小组的兴趣置于整个社区之上。每个人都声称正是由于游戏平衡不佳才产生这种现象。但是，深受游戏不良平衡性影响最深的玩家往往是那些沉默的大多数人。最大的受害者往往是默默地退出游戏世界。
对于许多玩家来说，将在线时间花在游戏上也是一种投资。他们期望自己的投资能够有良好的表现。无论自己的技能以及时间投入如何，如果发现有人在游戏中选择较好的职业、道具或其他选项而超越他们，他们就会变得烦躁不安。数据挖掘要的是精确的、以观察为依据的数据。有了这些数据，游戏设计师可以做出有根有据的决定。他可以找出游戏不良平衡性的受害者，对游戏进行修改，使所有玩家实现最大化表现的机会都是平等的。
数据挖掘还可以构建出较好的理论。它让游戏设计师可以洞察到玩家如何使用和滥用游戏。它拓宽了眼界，证明或反驳假设，用事实证明了自己想法的对错。随着游戏开发专业化的提升，游戏设计师已无法再总览全局，因而，游戏开发者对游戏本质的看法出现偏差是很普遍的事情。假信息、最佳情况场景和部分自我臆断扭曲了我们的理论。但是，如果我们可以扩宽视野，就可以开始挑战那些不可靠的观点。让我们学学如何审视这个大全局。
从数据到设计
游戏设计循环的最开始或许是设计，但是现在让我们以数据挖掘为起点来开始这个循环，这样我们就能够指导如何将过时的数据重新循环运用于新的设计：
将旧数据循环运用于新设计(from gamaustra)
1、在线：从在线服务中挖掘出大量原始数据。
2、存档：整理并将其储存成档案。
3、统计：过滤数据创建统计模型，这样能够获得的信息要超过原始数据
4、分析：进行真正的勘探，从玩家的表现中获取知识。
5、假设：提出如何改变游戏的假设。
6、测试：测试每个假设，然后将其运用到新的设计中。
最后的步骤便是结束循环。此循环的每次重复都能够改善游戏平衡。让我们深入挖掘这些步骤的深层次内容。
在线服务
大型多人游戏都有数千个游戏资产，甚至更多。每个职业、道具、怪物、任务、技能、区域或其他游戏实体都是游戏资产。在数据中，这些游戏资产是死物，但是在在线服务中，这些资产会焕发生机。让它们获得生命力的人是玩家。玩家行为能够产生有关游戏平衡性的丰富信息，因而应当尽可能多地挖掘数据，要收集大量的样本。与其他统计数据收集一样，样本收集应当是随机的，否则就无法代表玩家在游戏中的真实比例。样本组越庞大，你看到的游戏表现情景就越清晰。理想情况下，如果可以汇集起无数的样本，你就能够更精确了解玩家行为。小群体样本得出的统计结果是毫无意义的。我们很容易收集基于服务器的游戏数据，因为数据就在服务器上。
我们应当何时收集数据？诸如季度、每周某天和每天某时等时间循环会让数据收集变得复杂。这些循环中最基本和最有益的是周循环。一旦你理解了周循环，你就能够掌握月、季度或者假期循环。玩家一周中每天玩游戏的时间不同。他们都有现实世界中需要做的工作。因而，他们一周中每天玩游戏的时间会逐渐变动。下图显示玩家一周每天花在游戏上的时间数值。对于某个玩家群体而言，一周中某天或者一天中某时玩游戏的可能性会较高。比如，用户量可能会在周六、周天和周五晚上达到顶峰。
玩家每周每天的表现(from gamasutra)
除了数量差异之外，一周时间内每天的游戏质量也存在差异。当有些玩家有更多的时间花在游戏上时，他们或许会选择时间跨度更长的冒险。当他们只有少数空闲时间时，他们或许会停下游戏，更乐于同自己的好友保持联系。因而，为了避免这种日常变动，应当每周收集玩家表现数据。这可以让你了解到玩家整周的平均行为。确保在每周的同一天和同一个时段进行测量。你应当将这个过程自动化，比如在Unix环境中用“crontab”，或者使用可以提供这种服务的软件。如果你将衡量周期定为每周而不是每天时，你可以同时实现3个目标：消除了工作日变动对收集数据的影响、减少了数据收集的工作量和减少所需的档案存储空间。如果你衡量的数据不是玩家平均表现，那么你可能需要更加频繁地收集数据。
数据预处理
在获得原始数据之后，我们将它们进行预处理，使后期的分析更加容易。就像处理原始矿物质一样，数据的预处理也需要多个步骤。这个过程可用的方法有很多。以下这个较为简单的方法可以节省你的存储空间并减少勘探计算。这个预处理过程通常分为5个步骤：
1、浏览数据库。
2、确认所采集数据纯净且适合进行分析。
3、将数据整合成一个存档。
4、将数据分解，找出你所需要的领域。
5、将所得到的数据转变成易于分析玩家表现的形式。
这些步骤的细节内容会因为系统配置的不同而不同。以下内容以简单系统为例解释每个步骤：
准备可挖掘的原始数据(from gamasutra)
假设你正在运营一款玄幻类MMORPG游戏。
1、从账号数据库开始。这是首个步骤，因为账号数据库上有每个记录的ID，而这些记录正是你所需信息的来源。设置自动在周天清晨0:00记录用户数据。
2、确认与分析相关且纯净的数据。这个步骤可以在短时间内清除垃圾数据，这样你就不用去储存或分析无用的数据。从账户数据库开始，去除那些未注册账户或管理账户，比如那些人工修改过属性的测试和管理账户角色。对于账户中的每个有效角色，在登录数据库中查看其活跃程度。如果某角色上周并没有登录，那么其记录中就不含有玩家表现信息。
3、备份有效用户、登录和账户记录到存档数据库中。这是个很有用的步骤，将来你可以到这个存档中搜寻之前未曾考虑过的数据。这个备份对你来说很宝贵，其重要性就类似于考古学家的发现和侦探找到的证据。
4、但是，你现在手头有大量的数据。数据量大大超过你分析某个具体问题（游戏邦注：比如每小时游戏时间可以获得的经验值）所需的内容。因而，需要将数据进行解析，获得你所分析领域的数据。在这个例子中，我们需要选择的是角色ID、等级、职业、经验值和游戏时间。根据这些数据值建立表格。
栏目有：ID，等级，职业，经验值，时间。
5、将这些得到的数据转变成易于分析的形式。因为存档中有每周的数据，使用上周的数据可以获得新的信息。你可以看到经验值和游戏时间之间的差异。将这些栏目添加到表格上。如果角色刚刚于本周创建，可能就没有之前数周的信息。如果角色已经很长时间没有上线，那么就搜索存档更早的数据库信息。
栏目有：ID，等级，职业，经验值，时间，经验值，经验值差（游戏邦注：下文以“Δ exp”来表示），时间差（下文以“Δ time”来表示）
根据EPH将玩家表现数据存档(from gamasutra)
统计
使用基本的统计方法就可以从这个新鲜且准备充分的数据中解析出信息。因为需要解析的原始数据很多，所以需要对数据进行分类。举个简单的例子，我们以4种玄幻类游戏职业对数据进行分类：战士、牧师、盗贼和法师。
我们要衡量的是表现。不要被类别的角色数量所误导。游戏中某个职业的角色数量或者所选择的战略取决于与最优化表现相关的多个变量。文化偏好、美学、时尚、谣言和其他趋势也都有可能影响玩家的选择。
衡量的目标是比率而不是某个特定时刻的值。高表现指的并不是某个特定值，而是在短期内从较低的值提升到较高值的改变。衡量的时间阶段以周为周期。正如之前所提到的，以周为单位得到的数据比以天为单位更加稳定。
我们以每种职业每小时获得的经验值同等级的比值为例。“每小时经验值”是个很有用的数据，下文中我将简写为EPH。类似汽车的MPH（游戏邦注：即“每小时英里数”），玩家的EPH显示玩家在游戏中的进展速度和比率。我们要计算的是“经验值”这个表现指示器，而不是某个职业角色的数量。计算每周的经验值变更。计算角色真正存在于游戏中的时间，而不是现实生活中过去的时间。比如，如果角色每周的游戏时间是20个小时，那么我们要用的就是这个数值，而不是每周168小时。EPH的计算公式为：
EPH = Δ exp/Δ time
EPH计算公式(from gamasutra)
将得到的结果绘制成图表。纵轴就是EPH，横轴是等级范围。如果每个等级的样品过少，那么可以将相近等级划分成组。
比较选择不同策略的玩家表现(from gamasutra)
接下来，将每个类别定义为数据系列。在这个例子中，每个系列对应的是玩家的职业：战士、牧师、盗贼和法师。从纵轴上，我们可以看出每个职业表现的不同之处。如果这种差异很小，那么其并不具有统计学意义。如果差异很大，那么就需要引起重视。基于样本大小和其他数据质量，统计可以呈现出最小化的表现差异。在这个例子中，最明显的是高等级的战士与其他3种高等级职业间的差异。因而，统计结果发现，由于当周的表现较差，所以高等级战士的玩家数量受到影响。
分析
统计结束之后，数据挖掘的核心内容才真正开始。在这个阶段，我们将从之前收集到的原始数据中提出有价值的内容。有多种技术可以在这个阶段运用。以下是套较为简单的技术。
计算最大和最小表现值，包括表现比率和表现成长。在这个例子中，EPH可以通过经验值来计算，EPH本身可以被视为由等级和职业形成的函数：
EPH = f（等级）
微积分提供了衍生函数：
EPH’ = f’（等级）
因为样本大小有限，所以并不会存在精确的限制和衍生。但是，这种近似衍生可以让我们窥视到游戏平衡问题。在最大值的衍生中，玩家迅速发展。在最小值衍生中，玩家的发展停滞。他们玩了数个小时的时间，却只有小幅的提升。了解这些内容可以帮助我们将低表现游戏片段分离出来。
比较之前和之后的时段可以让我们了解到发展趋势。在这个例子中，可以用EPH减去上周的数值，得出新函数：
Δ EPH = f1（等级） – f0（等级）
如果得到的数值是正数，那么就表示玩家表现比上周好。这可以帮助我们认识到某项设计修改对游戏有何影响。玩家会调整自己的行为来应对设计的修改。通常只有少部分玩家会一开始便使用新功能。如果新功能确实比老功能要好，那么多数玩家便会迁移。在迁移之后，两个功能之间的数据对比就会趋于稳定。
以上两项技术均可以结合起来，用于分离和跟踪具体的低表现行为。比如，跟踪高等级战士从一周到下一周的改变，可以看出他们的表现是否得到提升。
EPH = 战士1(80%) – 战士0(80%)
将这个数值与其他职业的数值相比，就可以了解到相对变化。数值逐渐接近就表明职业开始趋于平衡。
自上而下与自下而上分析法(from gamasutra)
数据挖掘可以将至上而下的分析技术同至下而上的分析技术结合起来。从下往上看，我们的游戏就像是不存在分等级组织的游戏资产合集。从上往下看，同样的游戏可能看起来就是游戏资产的固定容器。群组分析能够提升职业或战略设计，因为它可以通过映射个体游戏资产的差异来产生群组。这可以比较不同类别中的相似资产。而且，群组分析可以识别多种战略共享的资产。
假设
作为游戏设计师，认为自己对游戏充分了解是种很危险的想法。分析应当能够激发假设的灵感，因为分析玩家行为可以证实或推翻某种对游戏资产的假设。这里提到的假设满足两个标准：
1、解释游戏资产的现有趋势。
2、预测修改、加入或移除某些游戏资产的结果。
以下是两个游戏资产假设示例：
1、在《无尽的任务》中，玩家偏好漂亮的种族。
2、在《黑暗时代》中，诱捕技能可以增加中级盗贼的表现。
everquest_races(from gamasutra)
定义域会限制假设适用的范围。在这个例子中，定义域便是具体的MMORPG游戏，即索尼在线娱乐的《无尽的任务》或Nexon的《黑暗时代》。确定你认为自己的发现能够运用到的定义域或范围。
假设当你在与美术人员谈论某MMORPG中种族的外观时，团队划分成了两个阵营。一个阵营认为，面目可怕的种族游戏资产数量应当与漂亮的种族相同。另一个阵营认为，选择漂亮种族的玩家更多，因而绝大多数资产应属于较为漂亮的种族。Nick Yee在他的《无尽的任务》研究文章《Norrathian Scrolls》中提供的调查数据或许可以激发此类假设。《无尽的任务》的玩家通常更偏爱精灵，其玩家数量与游戏中长相最为丑陋的两个种族——洞穴巨人和巨魔的玩家数比例为10:1。为了让这种假设显得更严密且更有说服力，应当分析玩家的数量和种族表现，因为正如上文中提到的那样，数据挖掘能够比调查更精确地展现出玩家表现。
dark_ages(from gamasutra)
在第二个例子中，假设你已经分析了《黑暗时代》的玩家表现。你注意到，根据衡量得到的EPH，与其他4中职业相比，中等级盗贼的表现较弱。我在1999年的时候就面临这种选择。我假设加入一套中等级诱捕技能能够通过改善伤害比例来提升表现。随后，我使用这篇文章所阐述的技术来测试我的假设。在转变过程中，部分玩家对盗贼的表现有所抱怨，尤其是那些玩非盗贼职业的玩家。但是实验最后成功了，在1个月的时间内，中级盗贼的EPH得到了平衡。
测试
测试假设是该循环中最为关键的步骤。当你意识到自己视为珍宝的假设是错误的时候，那种感觉很糟糕，因而要测试你的每个假设。如果是正确的，它的价值会在测试中体现出来。如果是不正确的，那么请抛弃它，为团队节省资源。
优秀的测试有且只有两种可能的结果：假设是真实的；假设是错误的。优秀的测试几乎不会得出模棱两可的结果，这意味着我们需要不断对测试进行重复或修改，直到得出确切的是非结果。该循环中的基础想法是：反复试验并从失败中寻找方法。正是因为测试能够发现游戏中的不当之处，游戏设计才能够得到改善。
以测试结果验证假设(from gamasutra)
在之前我们提到的示例中，高等级战士出现EPH较低的情况。假设有人提议增加新的游戏资产，用新的技能来增加战士的战斗效果。由此，你创造出“剑专精”技能。你可以收集测试服务器上的数据，比较每个职业以往和现在的EPH，来审视该技能是否提升了高等级战士的EPH以及出现了哪些其他的结果。
在测试中，应当尽可能地使用真实的条件。完全相同的真实条件固然不存在，但是你可以尽可能地接近。测试相同的配置、版本、功能，选取与每周收集数据相同的那天和那个时间。而且，测试人数规模可能会较小，这意味着结果的准确性较低。但是，测试中最难以掌控的因素就是玩家。你的测试玩家群体将不再是随机取样。这种自我选择出来的样本得出的动机和行为结果可能有所偏差。因而，测试可能会有错误。更为糟糕的是，发现误差方向是件很难的事情。
虽然完美测试是无法实现的，但是包含实验性错误的测试或许仍然可以大幅提升游戏的平衡性，因为这个过程可以不断重复。如果单次循环可以将游戏的不平衡性削减一半，那么两次循环就可以让游戏的不平衡性减少到1/4,。这比完全没有提升要好得多，当然，也比基于不准确信息进行设计要好，这种不准确信息包括竞争性特别兴趣小组提供的反馈。
在新设计通过测试之后，将设计重新投放到在线服务中。将整个过程再循环一遍，为得到最佳的结果，这种大循环应当每月进行一次。
不挖掘数据的原因
我们已经阐述了整个过程。现在，我们回头看看数据挖掘的健康范围。数据挖掘可以提供其他进化型游戏设计方法无法提供的信息。但是，它也不是万能的。
数据挖掘采集数字，将其处理并得出新的数字。这些数字并不能告诉你玩家的感觉。玩家或许对游戏的平衡性的理解出现偏差，但是玩家总是能够清晰地明白自己的感觉。有些玩家的感觉或许不成熟，有些玩家或许会有相互矛盾的反应。但是，他们的感觉都是真实的。每个玩家的情感回应都是鲜活的。数据无法反映玩家对每种游戏资产的感觉，无法显示哪个资产有着漂亮的模型、令人印象深刻的动画或吸引人的故事。
先入为主分析数据会干扰用户(from gamastura)
要慎用先入为主或预防性的数据挖掘手段，后者的目标在于识别和防止游戏中的作弊、骚扰或搞破坏行为。这等同于专利和盗版侵权二者间的关系。除了不道德之外，先入为主的数据挖掘还会对玩家协作产生影响。它还会干扰玩家的游戏过程。开发者难以发现其干扰性，而且会付出巨大的成本，因为这种干扰是由开发团队施加于用户。
数据挖掘也称为知识发现。虽然你从数据中挖掘出了知识，但是你无法挖掘出明智选择。你必须先对结果进行分析，决定哪些游戏不平衡性应当保持不变。数据挖掘在放大有效设计过程的同时，也会让游戏本身的问题加剧。
4种实际运用
现在你已经知道了整个过程，让我们将其运用到普通的MMOG问题中。以下是数据挖掘的4种实际运用：
1、平衡经济
2、寻找作弊者
3、削减制作成本
4、提升用户表现
平衡经济
每个游戏资产都会以商品或货币的形式在玩家之间交易。这些可交易的游戏资产奠定了游戏的经济。商品和货币不只限于金钱和装备。比如，在Nexon的《黑暗时代》中，我设计并执行了劳动、政治和宗教3种货币。
宗教、政治和劳动力也可以作为MMORPG中的货币(from gamasutra)
在衡量角色个体的得失时应当谨慎。关注商品或货币之间转换的交易。比如，角色有可能在一周后金钱数变少但是财富量增多。他可能用自己的金钱来换取其他有更大价值的商品。
跟踪游戏的宏观经济指示器。看看货币的供给量是增加还是减少。与现实世界中的货币供给相同，你可以通过这项数据了解到货币的通货膨胀率。衡量关键的表现指示器，然后产生如何提升游戏平衡性的假设。
你可以考虑使用改变游戏资产价格这种较为简单的平衡技术。相比其他数值的更改而言，玩家更能够接受价格的更改。比如，2002年时当Stewart Steel注意到Nexon的《风之国度》中法师的数量较少时，他通过增加该职业初始道具的方式来增加法师在游戏中的比例。事实上，游戏增加的是NPC支付给法师玩家的金钱数量。
相对而言，玩家比较能接受游戏调整物品售价(from gamasutra)
在测试假设之后，以月份为单位重复循环。虽然有些改变看似并不重要，但是可能会对经济系统其他部分产生极大的影响。在上述例子中，如果法师这个职业只是初始价值较高而后期依然没有较大改变，那么法师玩家的留存率可能随后还是会下降。
在平衡战略时，比如玄幻背景下的玩家职业，应当确保每个战略具有独特性。让我们看看之前的例子。低表现的高等级战士有许多特殊和共享资产。当添加新资产来平衡其表现时，最好不要选择让战士拥有“毒药抗性”这种资产。如果牧师职业有治疗中毒的技能，那么上述做法就是多余的。这会减小玩家选择牧师的概率，开始向其他两个职业聚集。反之，提供“剑专精”新技能是个不错的选择，这是其他职业所没有的能力。这种控制资产供应的方式能够确保各职业游戏中资产群的独特性。
在保持各种角色独特性的同时，平均每种角色的表现(from gamasutra)
寻找作弊者
MMOG中的作弊者不仅欺骗自己，而且他的表现还伤害了所有诚实的玩家。有些玩家可能会通过作弊来获得特别高的表现。玩家接受高表现战略，无论设计师是否愿意他们这么做。作弊还使得所有非作弊者的相对表现处在下风。如果没有得到迅速的修正，作弊会像野火那样扩散开来。一周甚至数天的作弊就可能影响到整个游戏经济。以下技术可以帮助你在经济受到影响之前找出作弊者。
调查可疑的玩家表现(from gamasutra)
从预处理阶段制成的表格开始。这里列举出了每个角色的ID和他们的表现。以表现这一栏为目标整理列表。现在，列表顶端的便是最令人起疑的角色ID。调查其另类的表现。
我们也可以通过EPH这栏来分类样本表格。顶端的角色是最可疑的。即便他获得的总经验值较低，但是他的比率较高，因为他在短时间内获得了这些经验值。调查登录来查看他的表现。这样你就能够知道玩家如何使用和滥用你的游戏。
这样找到的答案并不能展示出玩家的真实意图。玩家可能使用的是游戏中合情合理的功能。事实上，玩家可能会辩驳称，除非他修改了软件，否则自己所有的行为都是合理的。他按照你提供的方式来玩游戏。无论玩家的动机如何，发现作弊者并不能解决问题。系统的不平衡性才是催生作弊者的根源，因为你可以通过设计本身来杜绝作弊。
削减制作成本
每个游戏资产都需要编程、艺术、设计、测试和客户服务来开发和维护。但是，部分职业、道具、怪物、任务、技能、区域和游戏中的其他物品有可能被浪费了。这会影响开发者的斗志。
低表现的游戏资产的投资回报值接近于0。随着时间的推移，玩家做出决定并且不断优化决定，玩家间的交流会加速大家涌向这种最优化战略的过程。他们迅速地接纳了最高表现游戏资产，舍弃低表现资产。从竞争的角度来看，后者已经成为了累赘，所以可以抛弃。举个明显的例子，如果有两个属性几乎相同的武器，如果其中一个有着较高的伤害率，那么另一个武器就是无用的。在游戏中，经常出现这种新旧资产的交替现象。但是，这会使你的游戏中的内容无足轻重，因为没有人会使用这些资产。
要放弃无人问津的资产(from gamasutra)
但是，情况可以都得到改善。设计精巧的补丁会将资产重新放入玩家的选择列表中。要尽可能地重新循环美工、程序员和测试者付出艰辛劳动的成果。创造出新颖且平衡性良好的态势。保持这些资产的样式。使用适度的修正将这些美术资产循环运用于游戏，这样才不会浪费团队付出的人力和资源。但是要注意，我们只需要重新循环那些被玩家抛弃的资产，而非其他内容。因为玩家排斥开发者更改那些并不算过时的资产，他们想要的是更多新资产。
提升用户表现
如果玩家无法意识到自己的选择要如何行动才能够提升自己的表现，那么他就麻烦了。比如，如果所有的高等级战士的表现都比其他高等级角色要差，那么所有的战士都将陷入麻烦中。如果开发者不快速采取措施，玩家的这种麻烦感会转变成开发者的麻烦。
如果玩家无法提升表现，游戏就会流失用户(from gamasutra)
当玩家在单人游戏中面临表现不良的情况时，遭受不幸的只是他一个人。但是在大型多人游戏中，他归属的整个队伍都将遭受不幸。不幸的是，对于这个队伍来说，排除那些表现较差玩家是提升整体表现的绝佳选择。当低表现不是玩家的个人错误时，这就会催生出挫败感。假设排除低表现玩家可以让某个玩家小组的EPH提升20%，那么许多小组都将会做出这个选择。假设被排除出的低表现玩家无法改变自己的EPH能力，就像无法猎食和进化的濒危物种一样，这类玩家就会逐渐消失。游戏世界失去的只不过是角色，但是玩家玩游戏的动力也会就此消失。以此类推，如果玩家不想继续玩游戏，他们最终也不会为游戏付费。
为避免这种现象，开发者需采用平衡战略。不要改动游戏资产的现有状态，这会惹恼使用其他战略的其他玩家。他们会将这种改动视为不公平待遇，视为一种特别关照某个职业的行为。但你可以增加新的资产。
如果你的游戏已经商业化运营，那就应当提升玩家表现而不是让其变得更糟糕。如果某些资产过于优秀，但是玩家很喜欢，那就让其保持原样。只有那些长期造成用户损失的资产才需要移除，因为移除资产会影响用户对游戏的信赖。移除游戏中某些受到玩家喜爱的内容，无异于在对玩家说“我不想要你的钱，走开”。玩家需要预先为游戏付费或者持续支付订阅费用的理由，他们期望游戏每个月都能有所改善，他们的评判标准很简单——游戏应当提升他们自己的角色以及角色所属的特别兴趣小组。
游戏中存在的错误可能会影响玩家对游戏的信任。面对这些令人不悦的情况，开发者应向玩家证明你愿意与他们协商。
想象下这种最糟糕的场景：某个早上你的薪水减少了50%。你的汽车驾驶速度降为原来的一半，但需要的修理次数是原来的两倍，而费用也是原来的两倍。所有的这些噩梦般的事情都只是因为“造物主”希望如此。而游戏用户的感受也同此理，他们对游戏中的内容也会很较真。
关于数据挖掘和游戏设计，本文所探讨的只是冰山一角。这两个领域都是值得进一步调查、实验和研究，希望数据挖掘就是此类工具，能够帮助你改善游戏设计。
游戏邦注：本文发稿于2003年8月15日，所涉时间、事件和数据均以此为准。（本文为游戏邦/gamerboom.com编译，作者：David Kennerly）
Better Game Design Through Data Mining
David Kennerly
Players spend millions of man-hours selecting optimum strategies in massive multiplayer online games (MMOG). They are getting the best return on investment (ROI) from your MMOG. Are you? In this article, I will show you how data mining can improve game design, and then I will present four practical applications for applying that information:
1. To balance the economy
2. To catch cheaters
3. To cut production costs
4. To increase customer renewal
Although this article is written for MMOGs, you will find that most of these techniques can be adapted to multiplayer and single-player games. I will give several examples using fantasy MMORPG terms, since that genre is probably the best understood. However, these techniques apply to most MMO genres; I have even used these techniques to improve an online trivia game show. But before we learn the techniques, let’s understand why data mining is a good tool for these jobs.
Why Mine Data?
Because players lie. Player feedback alone provides a poor diagnosis of game design. The picture a player’s verbal feedback paints is not even an approximate guide. It is a distorted portrait of psychological and social forces. Players do not accurately report their own behavior in surveys or customer feedback. They may say one thing but do another instead. For example, anthropologist Dr. William Rathje surveyed the amount of beer people drank in a household and then went through their garbage. The garbage revealed twice as much consumption as the surveys had. This method was more insightful than surveys, which had been the traditional method of data collection. As psychological and social creatures, players, and developers, subconsciously revise their self-reports.
As political creatures, players, and developers, also revise their reports. Players belong to special interest groups, which bias their reports. Political ganging, a human trait, exists in online communities, too. Wherever a MMOG has guilds, classes, or any social organizations it has special interest groups. The members of these groups put their own group’s interests before those of the entire community. Each claims that it is the victim of poor game balance. But the players that actually suffer the most from poor game balance are the most silent. The greatest victims are ending their days in your game in quiet desperation.
To many players, the time spent online in your game is an investment. They expect their investment to perform well. They become upset if, despite their skill and time commitment, someone who happened to pick the better class, item, or other option in your game surpasses them. Data mining begins with accurate, empirical data. With this the game designer can make informed decisions. He can identify the victims of poor game balance, and he can correct it so that all players have an equal opportunity to achieve maximum performance.
Data mining also builds better theories. It gives the game designer insight into how players use and abuse the game. It broadens perspective, proves or disproves hypotheses, and substitutes facts in place of opinions. With increasing specialization of game development, a game designer no longer sees the big picture. It is all-too-common for any game developer to acquire a skewed view of the nature of his game. Disinformation, best-case scenarios, and a dose of self-hypnosis distort our theories. But if we can see the big picture, we can begin to challenge our own misinformed opinions. Let’s learn how to scan this big picture.
From Data To Design
In the beginning, there may have been the Design, but let’s start the cycle where data mining begins so we can discover how to recycle old data into new design:
1. Live: Scoop up lots of raw data in the live service.
2. Archive: From here, clean it up and store it for safe keeping in an archive.
3. Statistics: Sift through the data to create statistics, which are more informative than the raw data.
4. Analysis: Then apply the actual mining, which yields knowledge about player performance.
5. Hypothesis: Propose hypotheses about how to tune the game.
6. Test: Test each hypothesis and then introduce the new design into the live service.
The final step closes the loop. Each iteration of this cycle evolves game balance. Let’s dive into the details.
Live Service
A massive multiplayer game has thousands of game assets, or more. Every class, item, monster, quest, skill, zone, or any other game object is a game asset. In the data these game assets are dead; in the live service these assets come to life. It is the players that animate them. Player behavior generates rich information about game balance, so scoop up as much data as possible. Collect a large sample. Like any other statistical data collection, the sample should be random or otherwise representative of the actual proportions of player population. The larger the sample, the clearer the picture becomes. In a perfect game, an infinite number of players would render a perfect portrait of player behavior. On the other extreme, a small or biased sample generates no meaningful statistics. Given that this is a server-based game, collecting data is convenient. The data is already on your server.
When should data be collected? Temporal cycles, such as the season, day of the week, and the time of the day, complicate data collection. The most basic and instructive of these cycles is the weekly cycle. Once you understand the week, you can grasp the effect of a month, season, or holiday. Players cannot play as often as they wish on all days of the week. They have real-world schedules. So their playing volume varies depending on which day of the week it is. A graph depicts when most players participate. For a given player demographic, it might be higher on certain days of the week and certain times of the day. For example, usage might peak on Saturdays, Sundays, and Friday evenings.
In addition to quantity discrepancies, the quality of play differs depending on the day of the week. Some players might go on an extended adventure when they have more hours to spend. They might just stop to keep in touch with friends when they have little time to spend. So to avoid daily variation, collect player performance data once per week. This provides you with the average behavior for the whole week. Be sure to measure at exactly the same day and time of the week. You should automate this process, such as with “crontab” in the Unix environment, or whatever scheduling tools your database management software supports. When you measure once per week instead of once per day, you achieve three ends simultaneously: you eliminate weekday variation, you reduce the data collection workload, and you reduce the required archive storage space. If you are measuring data other than average player performance, then you may need to collect more often. But that is beyond the scope of this introduction.
Preprocess Data
After scooping up the raw data, let’s make it easier to analyze. Like processing a raw mineral, there are several steps that will prepare your data for mining. Many alternate methods can do this. Here is a simple method that economizes storage space and reduces mining computation. This preprocess has five general steps:
1. Take a snapshot of the database.
2. Validate that the data is clean and appropriate for analysis.
3. Integrate the data into a central archive.
4. Reduce the data down to just the fields you need.
5. Transform the reduced data into a form that is easy to analyze for player performance.
The details depend on the system’s configuration. This example explains each step in a simple system:
Suppose you are operating a fantasy MMORPG during its commercial service.
1. Start at the accounts database. This will be the first step to economy, since the accounts database has the ID of every record that you want information about. Schedule an automated snapshot of the user data at 00:00 on Sunday morning.
2. Validate which data is relevant and clean. This eliminates garbage as soon as possible, so that you are not storing or analyzing unusable data. Starting at the accounts database, exclude unregistered accounts or administration accounts. For example, exclude test and admin characters that have artificial attributes. For each valid character in an account, query for activity in the log database. If the character has not been active during the previous week, then its record contains no player performance information.
3. Backup valid user, log, and accounts records into an archive database. This will be a useful warehouse that you may return to in the future to mine for data you have not considered yet. Treat this backup preciously; if you were an archaeologist, this would be your find; if you were a detective, this would be your forensic sample.
4. However, you are now overwhelmed with a deluge of data. There is much more than you need to analyze a particular problem, such as the amount of experience points earned per hour of play. So reduce the data down to the fields you need. In this example, select the character ID, level, class, experience points, and number of hours played. Create a table of these values.
ID, level, class, exp, time
5. Transform this reduced data to make it easier to analyze. Since this archive has weekly versions of the data, use last week’s data to create new information. Get the difference of the experience points and the difference of the time played. Append these columns to the table. If this is the character’s first week, then there will be no information from the previous week. If the character has not played a while, then search backward through each prior week’s archive.
Δ exp = exp1 – exp0
Δ time = time1 – time0
ID, level, class, exp, time, Δ exp, Δ time
Statistics
Basic statistics can extract information from this fresh, well-prepared data. Since there is too much raw data to draw conclusions from, categorize or aggregate this data. For a simple example, let’s categorize the data by one of four fantasy player classes: fighter, priest, rogue, or wizard.
We will attempt to measure performance. Do not be misled by the popularity of each category. The number of characters that fit into a certain class or choose a strategy in the game depends on many variables irrelevant to optimum performance. Cultural preferences, aesthetics, fads, rumors, and other trends sway players’ choice. Chasing popularity as a measure of performance, leads to a vicious circle. Like a cat chasing its own tail, balance would never be achieved.
Measure rates instead of instantaneous values. High performance is not any particular value. It is measure of change from a low value to a high value in a short period of time. The period of time to measure is the week. As noted earlier, the week is more stable than the day.
Let’s take experience points per hour versus level for each class as an example. “Experience points per hour” is such a useful indicator that I will abbreviate it as EPH. Like a car’s MPH (miles per hour), a player’s EPH indicates his speed or rate of progress. Count the “experience points,” which is a performance indicator, instead of the population of a class. Count the change in experience points from one week to the next week. Count the time that the character actually played, instead of the total amount of time that has passed. For example, if the character played twenty hours in a week then use this value, instead of the 168 hours in a week. This gives the following formula:
EPH = Δ exp / Δ time
Let’s graph the results. On the vertical axis is the EPH, and on the horizontal axis is the level range. If there are too few samples per level, then group nearby levels together.
Next, plot each category as a data series. In this example, each series is a player class: fighter, priest, rogue, or wizard. Along the horizontal axis we can see the difference between the heights of each class’ performance. If the difference is small, then it is statistically insignificant. If the difference is large, then it is statistically significant. Based on the size of the sample and other qualities of the data, statistics defines the minimum gap that indicates significantly low performance. In this example, the most significant gap is between the high-level fighter and the other three high-level classes. So statistics discovered that the high-level fighter segment of the player population suffered from low-performance during that week.
Analysis
The core of data mining begins where statistics ends. Here we can extract golden knowledge from the raw mineral that we began with. Several techniques can be applied, most of them particular to the data and the purpose. Here is a simple set of techniques.
Calculate the maximum and minimum performance values. Do this for performance rate and performance growth. In this example, EPH is calculated from the experience points, and the EPH itself can be viewed as a function of class and level:
EPH = f(level)
Calculus provides the derivative:
EPH’ = f’(level)
Because of the finite sample size, the precise limit and derivative does not exist. However, the approximate derivative will provide insight into the game balance. At the maximum derivative players rapidly advance. At the minimum derivative players suffer stagnation. They play for hours with little advancement. Knowing this helps isolate low-performance segments of the player population.
Comparing a previous and subsequent period can identify a trend. In this example, the EPH can be subtracted from its value last week, creating a new function:
Δ EPH = f1(level) – f0(level)
Where the change is significantly positive, that segment of players is performing better than the previous week. This helps isolate the effect of a modification to a game’s design. Players’ adjustment to the modification delays full impact. Usually only early adopters will use the new feature at first. If it outperforms an old substitute, then most players will migrate. After migration, the empirical comparison between the two features stabilizes.
Both of the above techniques can be combined to isolate and track specific low performance. For example, tracking the change in high-level fighters from one week to the next indicates if their performance is improving or not.
EPH = Fighter1(80%) – Fighter0(80%)
Comparing this value to the other class values indicates the relative change. As the values converge, the classes are becoming balanced.
Data mining can combine top-down analysis techniques with bottom-up analysis techniques. From the bottom-up our game may appear to be a galaxy of game assets with no hierarchical organization. From the top-down the same game may appear to be rigid containers of game assets. Cluster analysis might improve class or strategy design, since it generates clusters from the bottom-up, by mapping differences of individual game assets. This can compare similar assets in different categories. As well, cluster analysis can identify assets that multiple strategies share. If you are interested, the books at the end of this article explain techniques for cluster analysis.
Hypothesis
As a game designer, it is dangerous to assume that you know your game. The analysis should inspire the hypothesis, since analyzing player behavior can prove or disprove a good hypothesis about game assets. The kind of hypothesis mentioned here meets two criteria:
1. Explain existing trends of game assets.
2. Predict the result of modifying, inserting, or removing a set of game assets.
Here are two examples of game asset hypotheses:
1. In EverQuest, players prefer pretty races.
2. In Dark Ages, a trap skill will increase mid-level rogue performance.
The domain delimits where the hypothesis applies. In this case the domain is a particular MMORPG, Sony Online Entertainment’s EverQuest or Nexon’s Dark Ages. Define the domain, or scope, that the knowledge that you believe you are discovering applies to.
Suppose when you discuss the appearance of races in an MMORPG with artists, the team divides into two camps. One camp argues for an equal number of game assets for gruesome player races as well as beautiful races. The other camp argues that many more players will choose beautiful races, so almost all assets should be devoted to the more beautiful races. Nick Yee provides survey data in his EverQuest research paper “Norrathian Scrolls” (http://www.nickyee.com/eqt/metachar.html#4) that may inspire this hypothesis. EverQuest players prefer Elves, in general, about 10-to-1 compared to the two least popular and, arguably, the most ugly races: Trolls and Ogres. To make the hypothesis rigorous, the player population and the race performances should be analyzed, because, as noted earlier, data mining more accurately depicts player behavior than a survey does.
In the second example, suppose you have analyzed player performance in Dark Ages. You note that mid-level, but not high-level, rogues have low-performance in terms of measured EPH when compared to the other four classes. In 1999 this was one of the decisions that I faced. I hypothesized that inserting a set of mid-level trap skills will improve performance, by improving their damage ratio. Then I used techniques in this article to test my hypothesis. During the transition, some players, especially non-rogues, argued about the performance of rogues. But the experiment succeeded: within a month, mid-level rogues had a balanced EPH.
Test
Testing your hypothesis is the most rigorous, sensitive, and critical step in the cycle. Although it feels good to hold a gem of wisdom, it feels bad to realize your treasured hypothesis is a false gem. So it is tempting, and sadly common, to halt the cycle before the testing stage. Test each hypothesis. If it is correct, it will survive with its value proven. If it is incorrect, then please conserve the team’s resources by discarding it.
A good test has two and only two possible outcomes: the hypothesis is true, or the hypothesis is false. A good test rarely yields an inconclusive result, which means the test needs to be repeated or modified to yield a definite true or false. This cycle is an elaboration of a basic idea: trial and error. Since testing detects error, it improves a game’s design.
In the earlier example, high-level fighters suffered from low EPH. Suppose someone suggests a new game asset, a new skill to increase the fighter’s combat effectiveness. You create a “Sword Mastery” skill to do this. After collecting data on the test server, you compare the old and new EPH for each class in order to conclude if the skill improved high-level fighter EPH and what other results it may have.
In the test, mirror actual conditions as much as possible. Just like an ideal point, or a limit, identical conditions do not exist, yet you can approximate. Test an identical configuration, build version, feature set, and at the same day of week and time of day. Additionally, the population will be smaller, which means results will be less precise. But the most uncontrollable factor of the test is the players. Your test player population is not going to be random sample. It will be a self-selected sample whose average motivations and behavior will be biased. So the test contains error. Worse than this, discovering the direction of bias may be an intractable problem.
Although a perfect test is impossible, a test that contains experimental error may still improve your game’s balance tenfold, because this process is iterative. If a single iteration cuts game imbalance in half, two iterations will quarter game imbalance, and so on. This is far better than no improvement at all, and certainly better than designing based on disinformation such as feedback motivated by competing special interest groups.
After a new design passes this test, feed the design back into live service. The process is iterative, so for best results, it should be repeated monthly.
Why Not To Mine Data
We have glossed over the general process. Let’s now step back and consider a healthy scope for data mining. Data mining provides answers that other methods of evolutionary game design cannot. However, it is not a panacea.
Data mining takes numbers, processes them, and makes new numbers. These numbers cannot tell you how each player feels. The player may be misinformed or biased about the balance of the game, but she is always right about how she feels. Some players’ feelings may be immature, and some players may have contradictory responses. Yet the paradox is that they are all right. Every player’s emotional response is valid. The data also does a poor job of revealing how players feel about each game asset. It does not indicate which asset has beautiful modeling, expressive animation, or a compelling story.
A healthy scope excludes preemptive or preventive data mining, which attempts to identify and prevent cheating, harassment, or sabotage. This equates to profiling and an invasion of privacy. Besides being unethical, preemptive data mining is disastrous. Data mining cannot establish cooperation or culpability. Not only is it prone to random error and false positives, but also it creates a new source of player harassment. This source of harassment is hard to discover, impossible to eliminate, and much more costly: Harassment by your own staff upon your customers.
Data mining is also called knowledge discovery. While you can mine knowledge from data, you cannot mine wisdom. You have to prioritize results and decide which game imbalances should be left alone. Data mining automates a process within your overall evolutionary design cycle. It amplifies an efficient design process and multiplies the problems in a poor process.
Four practical uses
Now that you have seen the general process, let’s apply it to some common MMOG problems. Here are four practical applications of data mining:
1. Balance the economy.
2. Catch cheaters.
3. Cut production costs.
4. Increase customer renewal.
Balance The Economy
Each game asset that passes hands between players is a commodity or currency. These tradable game assets define the game’s economy. The commodities and currencies need not be limited to money and property. For example, in Nexon’s Dark Ages, I designed and implemented a labor currency, a political currency, and a religious currency.
Be careful when measuring individual character gains and losses. Account for transactions that exchange one commodity or currency for another. For example, a character could have less money after one week but have more wealth. He may have exchanged his money for other commodities of greater value.
Track the game’s macro-economic indicators. See if the supply of currency is increasing or decreasing. Like a real-world money supply, this tells you about the inflation rate of the currency. Measure key performance indicators and generate hypotheses of how to improve game balance.
One simple balancing technique you can use is to change the price of a game asset. Players are more receptive to price changes than they are to other attribute changes. For example, in 2002 when Stewart Steel noticed low admittance rate for wizards in Nexon’s Nexus: The Kingdom of the Winds, he increased the rate by increasing the starting items of that class. In effect, this increased the price that an NPC paid the player for choosing the wizard career.
After testing the hypothesis, repeat the cycle each month. Each modification, although seemingly insignificant, can have a huge ripple effect on the rest of the economy. In the same example, if there were a higher starting value but a poor prospectus for the career of a wizard, then retention rate among wizards might drop.
While balancing the strategies, such as player classes in a fantasy setting, ensure that each strategy remains unique. Keep the clusters in strategic space from converging. Let’s return to the original example. The low-performing, high-level fighters have several unique and shared assets. When adding a new asset to balance their performance, it might be better not to give a fighter “Poison Tolerance.” If the Priest class has an ability to cure poison, then this would be redundant. It would reduce the group’s demand for Priests and begin to merge the two classes. Instead it might be better to provide “Sword Mastery” if no other class has this kind of ability. This controls the supply of assets so that each cluster of assets retains its unique niche in the game.
Catch Cheaters
A cheater in a MMOG does not just cheat himself. He performs an injustice to all honest players. Cheating short-circuits gameplay, so it achieves exceptionally high performance. Players adopt high performance strategies, whether intended by the designers or not. Cheating also penalizes the relative performance of all non-cheaters. If not corrected quickly, cheating will spread like wildfire. In a matter of weeks or even days a cheat can flood the game’s economy. These techniques can help catch cheating before it ruins the economy.
Start at the table that preprocessing generated. This lists each character ID and their performance. Sort the list by the performance column. Now, at the top of the list is the most suspicious character ID. Investigate his exceptional performance.
Let’s sort the example table by the EPH column. The character at the top is the most suspicious. Even though he has a lower total experience gain, he has a higher rate, since he accumulated the experience during fewer hours. Investigate the logs to discover how he performed so well. The answer will enlighten you as to how players use and abuse your game.
The answer does not indicate the player’s intention. The player may have been using a legitimate feature of the game. In fact, a player may argue that unless he modified the software, all of his behavior is a legitimate use. He played the game as it was given to him. Regardless of the motive, deleting a cheater cannot solve the problem. System imbalances breed cheaters, so the design itself can prevent cheating.
Cut Production Costs
Each game asset took some amount of programming, art, design, testing, and customer service to develop and maintain. Yet some of these classes, items, monsters, quests, skills, zones, and other objects in your game are being wasted. This lowers developer morale.
Game assets with low performance have a return-on-investment value that approaches zero. Players make decisions and optimize their decisions over time, and communication among players accelerates the migration to an optimal strategy. They quickly adopt the highest performing game assets available, and discard low-performance assets. In terms of competition, these assets are liabilities, so they become obsolete. For an obvious example, if there are two nearly equivalent weapons, except one has a higher damage rate, the other weapon is obsolete. In a game, this kind of decision, between obsolete assets and newer assets creates “fat”; there is some fraction of your game that might as well not exist, because no one uses it. Imagine having to break this news to an artist: “Thank you for the long-nights you spent making this new graveyard that we specified, but no one hunts there. Sorry about that.”
It does not have to be this way. A wise patch can put the assets back into the players’ list of options. Recycle the artists, programmers, and testers’ hard work as much as possible. Create new, well-balanced instances. Measure and prove their balance in terms of performance. Do not change the values of existing instances. Let them remain as they are. Recycle the art with modest modifications so those man-hours are not lost. But only recycle obsolete assets. Players do not tolerate recycling of assets that they do not consider obsolete. They demand fresh assets.
Increase Customer Renewal
If the player cannot or does not realize how to improve his performance with the choices he has already made, he is doomed. For example, if all high-level fighters perform worse than average high-level characters, all fighters are doomed. The players’ sense of doom will become the developer’s death knell unless you act fast.
When a player suffers from poor performance in a single-player game, he suffers alone. But in a massive multiplayer game, his whole team suffers. Unfortunately, a good choice for the team to increase their performance is to exclude low-performers. When low-performance is not the player’s fault, this breeds frustration. Suppose a group of players can increase its EPH 20% by excluding low-performance players. Sadly, many groups will. Suppose the excluded low-performer is unable to alter his EPH liability. Like an endangered species that is unfit to hunt and unable to evolve, this set of players becomes extinct. The character will not only become extinct from the playscape, but the player’s motivation to play will become extinct, too. If she will not play, eventually she will not pay, either.
To prevent this, balance the strategies. Do not edit existing instances of game assets. This will upset other players using other strategies. They will perceive the correction as an injustice, an act of favoritism. Instead of creating a perceived injustice, add new assets.
If your game is commercial, improve player performance instead of worsening it. If some asset is too good, but players love it, let it be. Only when the asset would cause long-term customer losses should it be removed, because removing or degrading an asset decreases customers’ good faith. There are few things that say, “I do not want your money, go away” as quickly as removing a beloved feature in the game. Players paid money in advance and continue to pay a subscription fee each month for a reason. They expect the game to improve each month. Their criterion is simple. The game should improve for their character personally and for the special interest group that their character belongs to.
A major error’s existence costs more than this loss of good faith. In these uncomfortable cases, prove to players you care by negotiation and diplomacy.
Imagine a worst-case scenario from the most extreme player’s perspective: This morning your paycheck was suddenly slashed 50%. Your brand of car drove half as fast, required repairs twice as often, and costs twice as much. Because the “gods” said so. Some customers take the game just as seriously.
Go Forth And Mine!
We have only touched the tip of the iceberg of data mining and game design. Both are elaborate and exciting fields for research, experimentation, and application. For years, we game designers have wanted systematic and scientific tools, and I hope that data mining is one such tool to help improve your game’s design. If you have questions, comments, or would like to discuss this topic in detail please contact me at kennerly@sfsu.edu. (Source: Gamasutra)
