游戏邦在:
杂志专栏:
gamerboom.com订阅到鲜果订阅到抓虾google reader订阅到有道订阅到QQ邮箱订阅到帮看

Playrix首席设计师谈如何创造三消游戏的元素以及关卡

发布时间:2019-10-30 09:00:04 Tags:,

Playrix首席设计师谈如何创造三消游戏的元素以及关卡

原作者:Aleksandr Shilyaev 译者:Willow Wu

你怎么创造三消游戏中的新元素?还有,关卡是如何设计出来的?你要怎么利用分析?Playrix的首席游戏设计师Aleksandr Shilyaev以他们的热门游戏《梦幻花园》为例,在圣彼得堡的White Nights大会上阐述了工作室是如何应对这些挑战的。

以下是编辑整理后的文字记录:

在Playrix,我负责设计《梦幻花园》的三消玩法。我想跟你们分享三消游戏中最有意思的设计阶段之一:创造新元素。除此之外,我还会讲到分析方法和A/B测试。

接下来的这些内容,我将按照开发里程碑分为若干部分。我们需要记住所有开发阶段,并提前考虑将来可能出现的任何问题,这很重要。

回到2016年,游戏还在测试发行的时候,《梦幻花园》的300多个关卡包含了16种不同的玩法。半年之后,关卡又增加了1000个,玩法增加到了29种。每次更新,我们都尝试增加一两种新机制。

1.创造新物品

第一阶段:研究竞争对手

我们需要创造一个新元素,但是从哪里开始着手呢?首先观察一下竞争对手。似乎有不少人都忽视了这个极为重要的阶段。三消市场有很多不可错过的好游戏。无论是设计新元素还是关卡,在构建自己的三消冒险之前你得意识到竞争对手已经做过类似东西。所以你可以看看他们的成果,如果觉得好,那就拆解分析,用自己的方法加以打磨,为己所用。

第二阶段:产生想法,加以过滤.

我们已经对竞争产品有所了解,可以开始设计新元素了。在这个阶段,我们把所有的初期想法整理到一个文档中。我们经常会有20~30个点子。把好的留下来,其它的都淘汰掉。

如何判断好坏?很简单:你只需要记住游戏的细节。就比如《梦幻花园》,让玩家不断匹配元素,没有任何停顿。奖励没有针对特定颜色,游戏过程中可以给彩虹炸弹充能。这些特色可以帮助我们缩小机制选择范围。比如说,我们必须要排除不能即刻触发的机制,因为它会阻塞游戏面板,影响效果,并给我们的游戏一种完全不同的感觉。

township-on-facebook(from games)

township-on-facebook(from games)

我们整理出一个主要清单并过滤掉那些不适用的点子,然后请专业人员选择2~3个看起来最有前景的机制。这是一个关键的阶段,我们只会关注两三个设计元素。

第三阶段:制作原型

在专家敲定点子、我们选择特定的机制之后,就到了下一阶段:制作原型。在这一阶段,我们能看到元素跟单位方格的结合效果、跟其它三消元素的结合效果,会对游戏体验产生怎样的影响。在制作原型期间,你得留心所有的这些方面。

最后得到的结果?元素已经准备好了,并且在原型阶段测试反馈良好。

第四阶段:画面创作

确定设计概念之后,我们就接下去创作画面内容。这个阶段不比之前的简单。主要的难点在于玩家与元素图像进行交互,必须表现出所有与概念相关的特点。

在开始创作之前,你需要明确游戏的设定。《梦幻花园》的主题是园艺,所有的三消元素都应该紧紧围绕这个核心。它们或者是跟花园相关的元素,或者是来自游戏主人公Austin的真实化生活环境,不存在任何虚构的幻想物品。

一旦大家明白了设定,我们就可以选择元素的特点了。元素一般都有多个特点,具体要取决于个体本身。就比如《梦幻花园》的障碍物:箱子沾满了方格,所以玩家明白箱子是固定在这的,不会随着其它元素移动。这就是其中一个特点。另一个例子就是强化道具。《梦幻花园》中的每个强化道具都有各自的炫酷小动画。当玩家在游戏面板看到它时,他们马上就能识别出来,思考使用策略。

元素特点设计完成,接下来的还有一件重要的事情就是查看它们与游戏整体视觉效果的契合程度,物品不能孤立存在,要与其它元素、特色共存。记住这一点并提前考虑是非常重要的。

你应该谨慎挑选元素的颜色,不会太突出也不会完全不起眼。规模和细节方面也应该注意。在这个阶段,你最好咨询专业的画师,为你指明方向。就拿游戏中的烟花来说吧(或许你已经玩了很久,非常了解这个道具),我们需要在一个元素的帮助下对整个面板造成影响。我们想到的最好方式是把它做成一个大烟花效果,Austin的柜子里也有烟花。

我们找到了参考资料,画了一些草图。之后,我们开始设计元素特点。这个元素是固定在面板上的,不能移动到别的位置,不能用来匹配。因此,我们将元素周围的空间压到最少,添加了部分与游戏棋子相对应的颜色,并添加了一个炸弹图标,让玩家知道它的效果。

第五阶段:引入元素

一旦元素准备就绪,我们就需要把它加入引导系统。在游戏初期,新元素出现的频率会比较高,元素特点也各不相同。你玩的时间越久,元素就越多,关卡也会变得更加丰富。然而,到了一定阶段,新元素就没有那么频繁地被引入了。

在制定决策之前,我们并不是就跟着感觉走。我们分析竞争对手。下图展示了《梦幻花园》与其它三消手游的引导过程对比。你可以看出大家的曲线是基本一致的(游戏邦注橙色的是《梦幻花园》)。

所以,我们做好了一个元素,也把它加入了玩家引导系统。现在我们可以开始设计关卡了。

2.关卡设计

在开始设计之前,有两件事情需要做决策:1.元素被引入的顺序,也就是你要在哪个关卡介绍哪个元素;2.难度曲线设计,不同阶段的关卡难度要怎么调整。

我们有一份文件显示了每个关卡所在的区域,以及过关条件。这其中包括难度、目标、关卡所使用的元素数量等等。事实上,我们整理成了一份说明供关卡设计师们参考使用。

现在设计师们有了指南,我们需要讨论的是他们在设计某个特定关卡时应该考虑的事。

我们只有四个基本要求:

a)每个关卡不能超过3种或者4种元素。

你可以偶尔不遵循这个规则,但最好不要太过头。不然的话玩家会觉得很迷茫。

b)元素必须互相兼容

并非所有的元素都能顺利组合。我们一直在跟踪记录哪些组合是可行的,哪些不能。

比如说,在《梦幻花园》中没有一个关卡是藤蔓和蜂蜜同时出现的。

c)每个关卡都应该有对应的想法。

例如,特定关卡或许需要特别的策略,或者它可能是完全随机的,只需要激活大量的强化道具。

关卡可以有一个有趣的开头,你需要炸毁一些东西,每次爆炸后你能得到新的掉落物。

d)每个三消关卡都可以在不使用强化道具的情况下通关。

根据难度曲线制作完关卡后,我们要做什么?开玩!

设计师调整关卡时,他们会根据我们主文档中的“预期难度曲线”进行设计。但实际怎么操作那就是另外一回事了,这完全取决于关卡设计师的技能。他们知道这一关的难度应该是80%,可以通过调整移动步数、炸弹充能速率等设置来实现这一点。

但是要实现整体目标不能只依靠一个关卡设计师。我们有一个专门的游戏测试部门,不断有新人加入。他们的任务是像普通玩家那样玩游戏。我们会记录下相关数据,据此调整关卡难度。

如何的分析关卡?

即使到开发的后期阶段,我们也没有停止关卡设计的相关工作。我们先从分析、优化单个关卡以及整体曲线开始。

前两三周是分析关卡,玩家的原始数据和游戏发送到我们的分析数据库的数据是非常关键的资料。所以你玩游戏,通关,游戏会在这个过程中给我们发送信息,让我们知道你做了什么。

然而,分析并没有告诉我该如何解决一切问题。

有什么是分析不能解决的?

1)Bugs

Bugs会对游戏体验造成干扰,或者让你无法完成关卡。通常情况下,你可以从关卡的某些明显变化中发现bugs,例如难度、强化道具的使用、玩家无法达到的目标等等。

2)难度曲线不合理

在制作关卡的时候,设计师会设定一个难度。当测试员完成一个关卡后,他们就对难度加以调整。当然,这个难度系数跟最终发行的很不一样。

所以我们就会对比这两个曲线,第一条是预期难度,第二条是基于玩家的实际关卡体验。当曲线出现偏离时,就表示这有可能是个问题,我们要特别注意这些地方。

3)连续多个关卡没有难度变化

有时你会遇到连续好几个关卡都是一样的难度,特别是难度达到92%、93%或95%的关卡,你得注意这个问题。《梦幻花园》曾经就出现类型的情况,我们得把问题揪出来解决掉。

如果你连续向玩家抛出几个高难度的关卡,他们会丧失斗志、停止游戏。所以一定要避免。

3.A/B测试

你分析了关卡、找到问题了,接下来你应该怎么调整关卡?我们的回答是A/B测试系统。

先跟大家透露下我们关注的几项数值。

·第一是难度。就如我之前所说的,这是失败次数与尝试次数之比。
·第二是流失率,或者卡在某个关卡七天以上,活跃度大大降低的玩家比率。
·第三是顺利通关所需的尝试次数。
·第四是转化率和盈利率。转化率显然就是有多少玩家在关卡中消费,而盈利率所指的是玩家为了通关所花费的金币数量。不是购买强化道具,而是花费金币的总数量。

怎么测试?我们把玩家分为三组。我们让第一组玩家玩这一关,设定都保持不变;第二组玩家玩的是难度降低的版本;第三组玩家玩的是调整后的新版本。然后我们进行观察。几个星期后,测试结果逐渐显现出来。这时候我们就可以决定是要保持原样还是要降低难度还是采用全新的版本。

就比如游戏的第77关,这是游戏中最早的付费墙之一。参考版本的问题之一是流失率太高,达到2.29%。我们必须重视这个问题,因为它属于游戏早期的关卡。如果我们在此阶段就失去这么多玩家,游戏长期的盈利无疑会受到影响,所以必须要降低流失率。

我们做了一个简单版本,又做了一个大改版本,然后测试对比。测试结果表明新版本的流失率最低。然而,盈利率、转化率和ARPU也有略微下滑。简单版本提升了盈利率但是流失率还是居高不下。我们认为这阶段更重要的任务还是压低流失率,所以我们选择了新版本。

游戏的第93关出现了更加棘手的状况。问题依然是流失率很高,经过A/B测试,我们发现全新版本的关卡流失率非常低,玩家基本都不会放弃,但是盈利率也非常低,我们需要一个折衷的办法。简单版本降低了流失率,盈利率只降了一点点,所以我们决定选用这个版本,因为它能留下更多玩家,盈利方面也跟原版基本一致。

要把三消玩法调整到理想的状态,你得在盈利和留存之间找到一个平衡点。如果你设计了一个高难度关卡,但又不想让玩家放弃游戏,单单增加移动步数降低难度是不够的,因为你的盈利率也会跟着降低。反过来也是同一个道理:一个简单的关卡,你想提高盈利率,光靠提升难度是不够的,因为流失率就会随之上升。

盈利率和流失率是紧密关联的。

要点

要设计游戏的三消玩法,你需要遵循以下四条规则:

·通过增加新元素和机制让玩家保持兴趣;
·注意画面的统一性;
·观察游戏过程中玩家的行为,以便及时发现、解决关卡设计过程中的错误;
·进行A/B测试或者其它类似测试来提高盈利率和留存率。

但最重要的前提是你有一个好游戏,有完善的系统、合理的引导过程以及酷炫的关卡,能够在早期留住玩家。

本文由游戏邦编译,转载请注明来源,或咨询微信zhengjintiao

How do you make new elements in a match-3 game? And how do you create levels and work with analytics? Aleksandr Shilyaev, lead game designer at Playrix, explains how the studio addressed these challenges while working on its hit title Gardenscapes.

What follows is an edited transcript of Aleksandr’s presentation at White Nights St.Petersburg.

At Playrix, I’m in charge of match-3 gameplay in Gardenscapes. I’d like to talk about one of the most interesting stages in a match-3 project – creating new elements. I’ll also touch on analytics and A/B testing.

The things that I am going to talk about and that I break up into separate points are development milestones. It’s important to make it clear that we keep all the development stages in mind simultaneously and try to think ahead about any problems that may arise in the future.

Back in 2016, when we soft-launched, we had 16 kinds of mechanics across 300 levels.

Half a year later, Gardenscapes already had 29 kinds of mechanics across 1,300 levels.

With each update, we try to add one or two new mechanics.

Creating a new item

Stage One: Studying the Competition

So we need to create a new element. Where do we begin?

We start by looking at our competitors. This is a very important stage that many people seem to forget.

There are a lot of cool games on the match-3 market that you absolutely have to know.

Before embarking on your own match-3 adventure, whether it’s the development of a new element or the creation of levels, you need to remember that your competitors have already done something like this.

So you look at what they’ve done, and if it’s a good idea, you analyze it, polish it, and make it work for your project.

Stage Two: Generating and Filtering Ideas

Now that we’ve studied the competition, we can begin developing a new element.

At this stage, we put all of our initial ideas into a single document. It often happens that we have over 20 or 30 of these. We keep the good ideas and filter out the bad ones.

How do we decide which ideas we should keep and which we should reject? It’s simple: you just need to remember the specifics of your game.

Gardenscapes, for example, is a match-3 game that allows players to continuously make matches without any pauses, with non-color-specific bonuses, and a cumulative “Lightning” bonus.

These features put certain limitations on the range of mechanics we can choose from. For example, we can’t choose a mechanic that isn’t triggered immediately, since it would block the field, interfere with the effects, and basically give our game a completely different feel.

When we’ve put together a master list and filtered out the bad ideas, a group of experts selects 2–3 mechanics that seem the most promising. This is an important stage where we only focus on 2 or 3 elements.

Stage Three: Prototyping

After the experts have made their decision, and we’ve selected certain mechanics, we move on to the next stage, which is prototyping. On this stage, we can see how the element combines with the tiles, how it combines with other match-3 elements, and how it affects the experience of the game. You need to look at all of these aspects during prototyping.

What do we get in the end? An element that’s ready and has tested well during the prototyping stage.

Stage Four: Image Creation

Once we’ve decided on the concept, we move on to creating the visuals for the element.

This stage is no easier than the previous ones. The main difficulty lies in the fact that the player interacts with the image of the element, so it must reflect all the properties that you associate with the concept.

Before you can start working on the image, you need to define the setting of the game.

Gardenscapes is a game about gardening, and all match-3 elements revolve around that concept. They’re all either the elements of a garden or elements from the real world in which Austin (the game’s protagonist) lives. There are no fantastical objects with amorphous properties in the game.

Once we’ve understood the setting, we can choose the element’s properties. Each element can have a number of properties — it all depends on the individual element.

Take the blockers in Gardenscapes, for example. Blockers, like Boxes, are elements that fill a tile, leaving no empty space. So the player understands that the Box is tied to the field and isn’t going anywhere. This is one of its properties.

Another example is boosters. In Gardenscapes, each booster has a small, glowing animation. When the player sees one of these elements on the match-3 field, they immediately identify it as unique and begin to strategize about how they can use it.

Once we’ve prepared the properties of the element, it’s important to check them against the general visuals of the game. The item won’t exist in isolation. It’ll coexist with other elements and features. It’s very important to keep that in mind and take it into account in advance.

You should make sure that the element doesn’t stand out too much and doesn’t get lost because of its color. It should also fit in in terms of scale and level of detail. At this stage, it’s best to consult an artist who can tell you which directions to go in.

Let’s take the example of our fireworks element. You probably know what it is if you’ve played long enough.

We had the following task: we needed to make an impact across the entire field with the help of one element. The best idea we came up with was to make it in the form of a big firework, an item that could realistically be found in one of Austin’s closets.

We found references and made some sketches. Next, we started working on the properties of the element.

The element is tied to the field, doesn’t move anywhere and responds to matches. Therefore, we minimized the space around the element, added colors to its sections that correspond to the game’s pieces, and added a bomb icon, to give players the idea of an explosion.

Stage Five: Introducing the Element

Once the element is ready, we need to add it to our introduction system. At the start of the game, new elements are introduced at a very high rate, and the elements themselves are very diverse. The longer you play, the more elements there are, and the levels themselves become more interesting. However, at some point, new elements are no longer introduced as frequently.

Before we make a decision, we don’t just go with our gut. We analyze our competitors. The graph below shows a comparison of Gardenscapes to three popular match-3 games. You can see that our introduction curve is consistent with that of our competitors (Gardenscapes is yellow).

So, we’ve prepared an element and made it part of our element introduction system. Now we start to make levels.

Level design

Before you start creating levels, it’s essential to decide on two things: the order in which elements will be introduced (which elements are you going to introduce and on which levels) and the difficulty curve (how exactly the difficulty increases or decreases from level to level).

We have a document that says where in the game each level is located and what requirements it needs to meet. This includes difficulty, goals, the number of elements used in this level, and so on. In fact, we put together a statement of work for the level designers which they use when making levels.

So we have this document and passed it on to the level designers. Now we need to discuss what they should take into account when designing a particular level.

There are only four basic requirements:

a) There should be no more than three or four elements per level.

You can break this rule occasionally, but it’s best not to overdo it. Otherwise, the level starts to feel chaotic. It’s confusing for the player, and it can be tough to navigate.

b) The elements need to be compatible.

Not all elements combine well with each other. We keep track of which combinations are possible and which are not in a master document.

For example, you won’t find a level in Gardenscapes where Ivy and Honey are used together. These elements are locked in the combination document.

c) Each level should have an idea.

Each new element creates new ideas.

For example, a certain level might require a special strategy, or it can be completely random, where you just need to set off a lot of boosters and collect bonuses.

A level can have an interesting beginning, where you need to blow something up and each time you get a new drop, or it can have a pre-set match that you have to make, and every time you play it, it’ll be exactly the same.

d) You should be able to beat every match-3 level in your game without using boosters.

What do we do once we’ve prepared a level that fits our difficulty curve? We start playing it.

When the level designer fine-tunes the levels, they design them according to the “projected difficulty curve” as described in our master document. How the designer does it is a whole other matter. It all depends on the skills of the level designer. Knowing that the level should have an 80% difficulty, the designer can achieve this by using settings such as the number of moves, the Lightning fill rate, and so on.

But it takes more than a level designer to get the full picture. We have a special department of playtesters with new people constantly coming on board. Their task is to play levels in the same way that real players would. We take their stats to adjust the difficulty of each level.

How do we analyze levels?

We don’t stop working on the level even once it goes into production. At that point, the work begins on analyzing and improving the overall curve and the individual levels.

We begin analyzing levels 2–3 weeks into production. To conduct our analysis, we need raw data from our players and the data that the game sends to our analytical database. So you play, you go through the game, and as you do that, the game is sending us information about what you do.

However, analytics don’t tell how to fix everything.

What can’t be fixed with the help of analytics?

1) Bugs

Bugs are errors that interfere with the game or make it impossible to complete a level. Very often, you can spot levels with bugs by seeing drastic changes in the level properties, such as difficulty, booster usage, goals that players fail to meet, and so on.

2) Inconsistencies in the difficulty curve

When a level designer makes a level, they set the difficulty. When the playtesters complete a level, they adjust the difficulty. Still, this often will be very different from what we ultimately get in the game.

So what we do is compare the two curves. The first curve is the projected difficulty, and the second is the difficulty curve based on players’ actual results in the level. Where the curves diverge, there is likely to be a problem, and we pay special attention to those places.

3) Long level sequences with the same difficulty

Again, sometimes playtesting is not enough. Sometimes you get several levels in a row that have the same difficulty. This is especially critical for levels that have a 92%, 93%, or 95% difficulty. This is something that happened in Gardenscapes, and we had to find it and fix it.

If you throw several difficult levels in a row at the player, they can get demoralized and stop playing. So you have to really look out for chains of levels like that.

A/B Testing

What should you do if you’ve analyzed a level, found the problem, and want to decide how to change the level? To answer this, we use the A/B testing system.

Let me tell you a little about the level metrics we look at.

·The first one is difficulty. As I said, this is the ratio of the number of losses to the total number of attempts.
·The second is the dropout rate or the percentage of players who have not completed this level for 7 days and have not been active in the game since.
·The third is the number of attempts before winning.
·The fourth is conversion and monetization. With conversion, it’s clear when the player made a purchase in the level, and by monetization, we mean the number of coins that the player spent to complete the level. So it’s not boosters, but the total amount, the number of coins.

How do we test the level? We divide the players into three groups. We let the first group play the level in its current state, the second one plays an easier version of the level, and the third group of players gets a completely new version of the level. Then we watch them play these levels. After a few weeks, we start to see the real picture. That’s when we decide whether to keep the level as it is, take the easier version, or go with a completely new version.

Here is level 77, for example. This is one of the first paywalls in the game. The main problem of the reference level (which is on the left of the slide) is a very high dropout rate, 2.29%. This is critical because we have this level early in the game. If we lose many players here, it affects the earnings of the whole game in the future.

So we needed to decrease the dropout metrics. We made a simplified level, we made a whole new level, and then we conducted a test. The test results showed that the dropout rate was lower for the totally new level. However, monetization, conversion, and ARPU also slightly decreased. The simplified version improved monetization, but the dropout rate was still very high. We decided that at this stage, it is more important for us to reduce the dropout rate. So we went with the new level.

The next example is level 93, a more difficult situation. In this level, we had a high dropout rate. So we did A/B testing and here’s what we saw: the new level (on the right) has a very good dropout situation, the players almost never quit. But the monetization metrics are very low. We tried to strike a balance here. The simplified version reduced the dropout rate, and the monetization metrics only decreased a little bit. So we decided to keep this level because it retains more players, and we earn almost as much as with the original level.

To fine-tune match-3 gameplay, you have to strike a balance between monetization and retention. If you have a difficult level where you want to reduce the dropout rate, it’s not enough to just add moves and make it easier, because your monetization metrics will also decrease. The opposite is also true: if you have an easy level, and you want to improve monetization, it’s not enough to just make it more difficult because you would also lose players.

Monetization and dropout metrics are closely connected.

Takeaways

To develop the match-3 part of your game, you need to follow four rules:

·add new elements and mechanics to keep the player interested in the match-3 mode;
·keep a close eye on the visual integrity so that the game doesn’t become visually messy;
·analyze the level flow so you can respond timely to errors made during the level design stage;
·conduct A/B testing or similar tests to improve monetization and retention.

But most importantly, you need to have a good game to begin with. The game should have a well-developed system, smart element introduction, and cool levels to keep players engaged in the early stages.

(source: gameworldosberver


上一篇:

下一篇: