1
00:00:04,050 --> 00:00:07,250
只要你在某个多人游戏社区待过

2
00:00:07,250 --> 00:00:08,850
这些词应该不陌生

3
00:00:08,850 --> 00:00:09,900
太强

4
00:00:09,900 --> 00:00:10,700
太便宜

5
00:00:10,700 --> 00:00:11,870
不公平

6
00:00:11,870 --> 00:00:14,020
其实这些玩家争议的

7
00:00:14,020 --> 00:00:16,250
是游戏平衡性

8
00:00:16,250 --> 00:00:21,270
“平衡性”是确保多人游戏一切选项公平的艺术

9
00:00:21,270 --> 00:00:24,550
没有太弱、不值得一用的设计

10
00:00:24,550 --> 00:00:28,950
也没有太强、统治一切的存在

11
00:00:28,950 --> 00:00:30,100
但话说回来

12
00:00:30,100 --> 00:00:32,970
大多数电子游戏不只是为了平衡

13
00:00:32,970 --> 00:00:37,270
而是追求众多差异化选项间的平衡

14
00:00:37,270 --> 00:00:41,000
做好一款“对称游戏”的平衡不难

15
00:00:41,000 --> 00:00:44,860
因为所有玩家的开局条件完全相同

16
00:00:44,860 --> 00:00:46,950
但多数游戏是“非对称”

17
00:00:46,950 --> 00:00:51,250
即玩家可能面对与自身完全不同的对手

18
00:00:51,250 --> 00:00:55,070
试想一款游戏包含74名不同斗士，或

19
00:00:55,070 --> 00:00:57,250
140位独特英雄

20
00:00:57,250 --> 00:01:00,700
开发者期望对于水平相近的玩家

21
00:01:00,700 --> 00:01:03,220
每种选项都各有潜力

22
00:01:03,370 --> 00:01:05,000
他们是怎么做的？

23
00:01:05,000 --> 00:01:05,970
不过我想说

24
00:01:05,970 --> 00:01:08,870
追求平衡性并非易事

25
00:01:08,870 --> 00:01:12,450
公司可能为此成立整个部门，而且

26
00:01:12,450 --> 00:01:18,220
Riot的《英雄联盟》在过去十年已发布超过200个平衡性补丁

27
00:01:18,220 --> 00:01:21,450
另外它不仅关乎数值，还有玩家心理

28
00:01:21,450 --> 00:01:23,700
《守望先锋》的杰夫·卡普兰说

29
00:01:23,700 --> 00:01:27,520
“对平衡性的感知比平衡性本身更重要”

30
00:01:27,520 --> 00:01:31,100
所以本期节目并非细致教程

31
00:01:31,100 --> 00:01:34,720
“游戏制作工具箱”将带你走马观花

32
00:01:34,720 --> 00:01:36,600
关于游戏如何去做平衡

33
00:01:36,600 --> 00:01:38,120
再平衡

34
00:01:38,120 --> 00:01:39,270
再平衡

35
00:01:39,270 --> 00:01:40,310
再平衡

36
00:01:40,310 --> 00:01:41,500
再平衡

37
00:01:42,250 --> 00:01:45,770
那开发者一开始是怎么做平衡的？

38
00:01:45,770 --> 00:01:48,870
他们首先考虑“制衡”

39
00:01:48,870 --> 00:01:54,470
也就是让角色的竞争优势能被缺点所制约

40
00:01:54,470 --> 00:01:56,520
以“马力欧赛车”为例

41
00:01:56,520 --> 00:02:00,170
森喜刚那种重型车手极速很高

42
00:02:00,170 --> 00:02:01,920
但加速很慢

43
00:02:01,920 --> 00:02:04,900
奇诺比奥那种轻量级车手正好相反

44
00:02:04,900 --> 00:02:06,400
在设计平衡的赛道上

45
00:02:06,400 --> 00:02:08,500
他们几乎旗鼓相当

46
00:02:08,500 --> 00:02:11,620
你可以假设角色有个“能力预算”

47
00:02:11,620 --> 00:02:14,070
至少Riot是这么想的

48
00:02:14,070 --> 00:02:15,920
“优势”是一种开销

49
00:02:15,920 --> 00:02:18,320
“劣势”是一种补贴

50
00:02:18,320 --> 00:02:22,370
如果所有角色都花掉相同的“能力预算”

51
00:02:22,370 --> 00:02:24,620
他们就离平衡更近一步

52
00:02:24,620 --> 00:02:26,450
当然一般没那么简单

53
00:02:26,450 --> 00:02:27,500
没错

54
00:02:27,500 --> 00:02:31,120
有时一张牌对所有随从造成1点伤害

55
00:02:31,120 --> 00:02:34,500
另一张对所有随从造成4点伤害

56
00:02:34,500 --> 00:02:35,570
这个不难

57
00:02:35,570 --> 00:02:38,870
只要让第二张牌费用更高

58
00:02:38,870 --> 00:02:43,300
但你该如何计算那些不可比较的事物，比如

59
00:02:43,300 --> 00:02:44,650
《守望先锋》的英雄？

60
00:02:44,650 --> 00:02:47,320
或各种参数组成的选项？

61
00:02:47,320 --> 00:02:52,220
例如，Bungie时代的《光环3》最初狙击步枪过强

62
00:02:52,220 --> 00:02:54,220
它包含大量微调参数

63
00:02:54,220 --> 00:02:55,570
比如视野大小

64
00:02:55,570 --> 00:02:56,800
放大时间

65
00:02:56,800 --> 00:02:57,900
装弹时间

66
00:02:57,900 --> 00:02:59,100
最大携弹量

67
00:02:59,100 --> 00:03:02,900
他们最后发现，最佳做法是调整“射击间隔”

68
00:03:02,900 --> 00:03:06,800
将0.5秒调成0.7秒

69
00:03:06,800 --> 00:03:07,920
其中关键在于

70
00:03:07,920 --> 00:03:11,020
你得突出不同选项的差别

71
00:03:11,020 --> 00:03:13,020
狙击步枪和霰弹枪

72
00:03:13,020 --> 00:03:17,070
应该是令人兴奋的选择，而非简单两种突击步枪

73
00:03:17,070 --> 00:03:19,820
哪怕后者更容易做平衡

74
00:03:20,220 --> 00:03:23,320
所以我赞同前暴雪设计师罗布·帕尔多

75
00:03:23,320 --> 00:03:28,270
他曾提醒设计师，别用数学公式把“平衡”变“平庸”

76
00:03:28,270 --> 00:03:28,770
他说

77
00:03:28,770 --> 00:03:33,120
你最后得到的游戏会处处平淡无奇

78
00:03:33,120 --> 00:03:36,220
你们可以击掌相庆说平衡了，但这有趣吗？

79
00:03:36,220 --> 00:03:37,200
恐怕不是

80
00:03:37,900 --> 00:03:40,520
另一个考虑是“相克”

81
00:03:40,520 --> 00:03:45,500
即赋予角色抵消对手操作和战术的能力

82
00:03:45,500 --> 00:03:46,270
举个例子

83
00:03:46,270 --> 00:03:49,770
“星际争霸”的虫族速攻很好用

84
00:03:49,770 --> 00:03:54,350
但前提是对手未能抢先造出防御性地堡

85
00:03:54,350 --> 00:03:58,470
我们期望的是“一物降一物”

86
00:03:58,470 --> 00:04:00,820
所以星际中的防御打法

87
00:04:00,820 --> 00:04:04,320
也会被一种经济型战术克制

88
00:04:04,320 --> 00:04:10,350
即囤积资源，建造最终能碾压地堡的大型单位

89
00:04:10,350 --> 00:04:13,470
然后我们得为这也造个克星

90
00:04:13,470 --> 00:04:15,720
但这样会没完没了

91
00:04:15,720 --> 00:04:17,650
其实办法已经有了

92
00:04:17,650 --> 00:04:21,770
你该如何对付那些悠哉着存钱的对手？

93
00:04:21,770 --> 00:04:23,500
嗯，速攻

94
00:04:23,500 --> 00:04:25,360
你应该想到了

95
00:04:25,360 --> 00:04:28,250
这是“石头剪刀布”

96
00:04:28,250 --> 00:04:32,820
这种决定谁去洗碗的傻游戏可能过于简单

97
00:04:32,820 --> 00:04:34,950
也缺乏策略深度

98
00:04:34,950 --> 00:04:37,200
但它是完美平衡的

99
00:04:37,200 --> 00:04:39,220
因为一切都有克星

100
00:04:39,220 --> 00:04:40,920
一切都是克星

101
00:04:40,920 --> 00:04:43,870
伙计，赢了你的剪刀

102
00:04:43,870 --> 00:04:48,100
这就是为何它成了许多多人游戏的基石

103
00:04:48,100 --> 00:04:50,820
几乎每个格斗游戏都有类似系统

104
00:04:50,820 --> 00:04:52,070
比如“死或生”

105
00:04:52,070 --> 00:04:54,820
它所夸耀的三角系统就是

106
00:04:54,820 --> 00:04:56,150
“打击”克“投技”

107
00:04:56,150 --> 00:04:57,320
“投技”克“返技”

108
00:04:57,320 --> 00:04:59,000
“返技”克“打击”

109
00:04:59,000 --> 00:05:02,470
在策略游戏中，三角相克的不仅是战术

110
00:05:02,470 --> 00:05:04,320
还有作战单位

111
00:05:04,320 --> 00:05:08,200
虽然种类繁多的宝可梦分布于一张相生相克的大网

112
00:05:08,200 --> 00:05:12,370
但开局的御三家，是火、水、草

113
00:05:12,370 --> 00:05:16,350
“石头剪刀布”是很好的平衡性框架起点

114
00:05:16,350 --> 00:05:19,750
因为你能确保没有选项过强

115
00:05:19,750 --> 00:05:21,120
克星总是存在

116
00:05:21,120 --> 00:05:22,970
也没有选项无关痛痒

117
00:05:22,970 --> 00:05:26,170
一个起码能克制另一个

118
00:05:26,170 --> 00:05:28,550
另外至少对策略游戏而言

119
00:05:28,550 --> 00:05:30,310
它会鼓励混合策略

120
00:05:30,310 --> 00:05:32,870
它要求你成为多面手

121
00:05:32,870 --> 00:05:37,250
迫使你在不同战术之间随机应变

122
00:05:37,250 --> 00:05:39,000
对于多职业游戏

123
00:05:39,000 --> 00:05:42,300
它很容易自动促成混合队伍

124
00:05:42,300 --> 00:05:43,920
以《军团要塞2》为例

125
00:05:43,920 --> 00:05:46,150
九个职业中的七个

126
00:05:46,150 --> 00:05:51,920
组成一张“石头剪刀布”相生相克的复杂关系网

127
00:05:51,920 --> 00:05:57,200
所以，团队必须挑选互补职业来相互照应

128
00:05:57,200 --> 00:06:00,700
比如你是工程师，间谍不断瘫痪你的步哨枪

129
00:06:00,700 --> 00:06:04,720
这时你就需要队友切换成火焰兵

130
00:06:04,720 --> 00:06:07,970
这些通常被称为“硬相克”

131
00:06:07,970 --> 00:06:10,000
它们能彻底击败对方

132
00:06:10,000 --> 00:06:13,620
比如《ARMS》的“出拳”硬克制“投掷”

133
00:06:13,620 --> 00:06:16,870
因为前者每次都能化解后者

134
00:06:16,870 --> 00:06:21,720
而“软相克”则意味着一个面对另一个有优势

135
00:06:21,720 --> 00:06:24,020
比如麦克雷胜过猎空

136
00:06:24,020 --> 00:06:27,620
但前者胜率不可能100%

137
00:06:27,770 --> 00:06:29,370
说到相克

138
00:06:29,370 --> 00:06:32,600
很重要的一点是分清，什么是“手”

139
00:06:32,600 --> 00:06:34,620
什么是“出手”

140
00:06:34,620 --> 00:06:38,570
“手”是指比赛开始前就已确定的东西

141
00:06:38,570 --> 00:06:41,800
比如角色和种族

142
00:06:41,800 --> 00:06:44,850
“出手”是指比赛过程中的抉择

143
00:06:44,850 --> 00:06:47,550
包括操作、单位和战术

144
00:06:47,550 --> 00:06:50,020
对于团队游戏，如《守望先锋》

145
00:06:50,020 --> 00:06:52,150
整支队伍是“手”

146
00:06:52,150 --> 00:06:55,180
单个玩家的表现是“出手”

147
00:06:55,180 --> 00:06:59,400
“出手”在设计上会突出对抗的不平衡

148
00:06:59,520 --> 00:07:02,820
以此营造胶着战况和团队协作

149
00:07:02,820 --> 00:07:05,660
但“手”的设计旨在平衡

150
00:07:05,660 --> 00:07:08,700
确保玩家能用上所有招式

151
00:07:08,820 --> 00:07:12,120
如果桑吉尔夫不能格挡

152
00:07:12,120 --> 00:07:13,620
他就没法打了

153
00:07:17,400 --> 00:07:20,970
于是有了一堆角色，考虑了“制衡”和“相克”

154
00:07:20,970 --> 00:07:23,270
你觉得这平衡了

155
00:07:23,270 --> 00:07:25,750
但你该如何确保这点？

156
00:07:25,750 --> 00:07:28,200
嗯，是时候收集数据了

157
00:07:28,200 --> 00:07:30,050
无论是来自内部测试

158
00:07:30,050 --> 00:07:32,950
还是成千上万的在线玩家

159
00:07:32,950 --> 00:07:37,850
你可能觉得，要做的也就是跟踪每名角色的获胜表现

160
00:07:37,850 --> 00:07:39,970
也即“胜率”

161
00:07:39,970 --> 00:07:42,430
如果一名角色有50%胜率

162
00:07:42,430 --> 00:07:43,520
就平衡了

163
00:07:43,520 --> 00:07:45,220
但正如所有统计那样

164
00:07:45,220 --> 00:07:46,720
这也会误导人

165
00:07:46,720 --> 00:07:49,400
想想一款格斗游戏有三名角色

166
00:07:49,400 --> 00:07:52,120
如果隆每次都能赢春丽

167
00:07:52,120 --> 00:07:54,050
却每次都输给嘉米

168
00:07:54,050 --> 00:07:56,660
他的胜率也是50%

169
00:07:56,660 --> 00:07:57,970
但这平衡吗？

170
00:07:57,970 --> 00:07:59,120
非也

171
00:07:59,120 --> 00:08:00,700
这就是“对阵胜率表”

172
00:08:00,700 --> 00:08:04,850
一张每个角色对阵其它所有角色的胜率表

173
00:08:04,850 --> 00:08:06,370
为何如此重要

174
00:08:06,370 --> 00:08:08,750
但即使这张表也不全面

175
00:08:08,750 --> 00:08:12,150
Riot曾遇到《英雄联盟》角色阿卡丽的问题

176
00:08:12,150 --> 00:08:16,220
数据显示她非常平衡，44%的胜率

177
00:08:16,220 --> 00:08:18,070
可能还有点偏弱

178
00:08:18,070 --> 00:08:22,950
所以她是怎么在2018全球总决赛稳获72%胜率

179
00:08:22,950 --> 00:08:26,120
并在禁选率上拔得头筹？

180
00:08:26,120 --> 00:08:29,020
这是因为她虽然很强

181
00:08:29,020 --> 00:08:31,220
却难以精通

182
00:08:31,220 --> 00:08:34,250
换言之，她的技巧门槛很高

183
00:08:34,250 --> 00:08:38,100
所以尽管顶级玩家能用她横扫千军

184
00:08:38,100 --> 00:08:42,500
低级玩家用阿卡丽却被虐来虐去

185
00:08:42,500 --> 00:08:45,950
因此，她的胜率被拖累了

186
00:08:45,950 --> 00:08:51,370
这就是为何要在不同水平线上考察一名角色的对阵胜率

187
00:08:51,370 --> 00:08:56,250
最后，胜率并不完全体现游戏实际情况

188
00:08:56,250 --> 00:08:59,570
我们还得知道玩家会选哪些角色

189
00:08:59,570 --> 00:09:03,120
人们可能会嫌弃某些平衡的角色

190
00:09:03,120 --> 00:09:05,570
因为角色不够有趣

191
00:09:05,570 --> 00:09:08,660
或只适应特定场景

192
00:09:08,660 --> 00:09:11,020
暴雪发现《守望先锋》的秩序之光

193
00:09:11,020 --> 00:09:13,200
是个大体平衡的角色

194
00:09:13,720 --> 00:09:17,820
但她出场率不高，因为适用场景很有限

195
00:09:17,820 --> 00:09:20,050
所以在首次重做中（2016）

196
00:09:20,050 --> 00:09:24,070
暴雪为了让她更流行，给了她两个终极技能去选

197
00:09:24,070 --> 00:09:26,500
“传送面板”和“护盾发生器”

198
00:09:26,500 --> 00:09:30,470
这就是为何玩家反馈和出场率如此重要

199
00:09:30,470 --> 00:09:34,370
后者能体现一名角色实际被用的频率

200
00:09:34,370 --> 00:09:35,770
对于《彩虹六号：围攻》

201
00:09:35,770 --> 00:09:40,200
育碧用一个矩阵来对照胜率和出场率

202
00:09:40,200 --> 00:09:45,070
加上针对落入这四个区间每种干员的不同考虑

203
00:09:45,070 --> 00:09:48,720
出场率还能反映当前趋势

204
00:09:48,720 --> 00:09:52,950
也就是比赛环境下流行的角色、牌组、战术等等

205
00:09:52,950 --> 00:09:57,270
那些被主流社区所发掘的最高效和常用搭配

206
00:09:57,270 --> 00:09:59,820
它们通常可见于论坛帖子

207
00:09:59,820 --> 00:10:01,470
粉丝自制排行

208
00:10:01,470 --> 00:10:02,770
YouTube视频

209
00:10:02,770 --> 00:10:04,650
和电竞赢家那

210
00:10:04,650 --> 00:10:08,220
当名为詹森的小伙赢得《皇室战争》赫尔辛基锦标赛后

211
00:10:08,220 --> 00:10:11,900
他的牌组突然成了热门

212
00:10:11,900 --> 00:10:15,150
趋势其实也是自我平衡的力量

213
00:10:15,150 --> 00:10:18,770
假设人人都知道特定角色太强

214
00:10:18,770 --> 00:10:20,720
人人都去用

215
00:10:20,720 --> 00:10:22,600
那现在大家的关注点

216
00:10:22,600 --> 00:10:26,770
就是拼命找出克制或针对那个角色的打法

217
00:10:26,770 --> 00:10:29,770
一旦找到，趋势就可能变化

218
00:10:29,770 --> 00:10:32,170
这种风水流转让游戏常葆新鲜

219
00:10:32,170 --> 00:10:36,700
赋予那些找到克制之道的玩家极大满足感

220
00:10:36,700 --> 00:10:38,670
《守望先锋》的杰夫·卡普兰说

221
00:10:38,670 --> 00:10:42,470
“想想那些趋势转变是源于玩家发明的新战术”

222
00:10:42,470 --> 00:10:45,320
“嗯，这是最理想状况”

223
00:10:45,320 --> 00:10:48,050
“但我们已看过它一次次上演”

224
00:10:48,050 --> 00:10:50,820
当然，它并非总是奏效

225
00:10:50,820 --> 00:10:54,270
有时，设计师得动手做些修改

226
00:10:54,270 --> 00:10:56,220
如果战术过于强势

227
00:10:56,220 --> 00:10:58,220
如果角色无人问津

228
00:10:58,220 --> 00:11:00,470
如果某种打法破坏氛围

229
00:11:00,470 --> 00:11:02,570
是时候动刀了

230
00:11:02,570 --> 00:11:03,170
首先

231
00:11:03,170 --> 00:11:08,970
开发者需要搞清角色或战术不平衡的具体原因

232
00:11:08,970 --> 00:11:12,450
你很容易看出某个角色称霸胜率表

233
00:11:12,450 --> 00:11:14,770
但不容易找出原因

234
00:11:14,770 --> 00:11:18,120
对于《任天堂明星大乱斗X》的梅塔奈特

235
00:11:18,120 --> 00:11:21,200
不平衡很可能是因为他极快的攻速

236
00:11:21,200 --> 00:11:25,420
和半空中取消操作、规避击倒的能力

237
00:11:25,420 --> 00:11:28,050
他优势满满，制衡不足

238
00:11:28,050 --> 00:11:31,320
其他角色也没法克制他

239
00:11:31,320 --> 00:11:32,820
一旦找到根源

240
00:11:32,820 --> 00:11:35,950
你就得琢磨削弱什么、强化什么

241
00:11:35,950 --> 00:11:38,420
“削弱”是指降低某物威力

242
00:11:38,420 --> 00:11:39,850
比如减慢速度

243
00:11:39,850 --> 00:11:41,050
限制射程

244
00:11:41,050 --> 00:11:42,670
或削减攻击

245
00:11:42,670 --> 00:11:46,000
“强化”正相反：让某物更强大

246
00:11:46,000 --> 00:11:50,400
但你不一定要强化弱势角色、削弱强势角色

247
00:11:50,400 --> 00:11:52,820
你可以把强势角色留着

248
00:11:52,820 --> 00:11:54,850
然后强化克制他的角色

249
00:11:54,850 --> 00:11:56,700
一样能解决问题

250
00:11:56,700 --> 00:12:02,100
请务必看看Core-A Gaming这期视频：为什么更应Buff而非Nerf？

251
00:11:56,700 --> 00:12:02,100


252
00:12:02,200 --> 00:12:06,520
平衡性调整多种多样，无论是微调角色移动速度

253
00:12:06,520 --> 00:12:09,150
还是回炉重做角色设计

254
00:12:09,150 --> 00:12:11,950
它甚至可能改变游戏基本规则

255
00:12:11,950 --> 00:12:15,120
《彩虹六号：围攻》让攻守更平衡的做法

256
00:12:15,120 --> 00:12:17,470
是将比赛缩为三分钟

257
00:12:17,470 --> 00:12:20,520
有时你甚至得砍掉整个东西

258
00:12:20,520 --> 00:12:25,400
就像Epic移除《堡垒之夜》的神器“无尽之剑”

259
00:12:25,400 --> 00:12:27,870
任何改动都会影响玩家

260
00:12:27,870 --> 00:12:33,170
尤其是那些已非常习惯特定角色或相克的玩家

261
00:12:33,170 --> 00:12:34,620
所以当游戏更新时

262
00:12:34,620 --> 00:12:38,820
通过更新日志或视频来解释改动非常重要

263
00:12:38,820 --> 00:12:41,270
其实，更新日志太重要了

264
00:12:41,270 --> 00:12:44,670
Riot曾发布日志，说一名英雄被削弱

265
00:12:44,670 --> 00:12:47,350
却忘了对代码做实际修改

266
00:12:47,350 --> 00:12:50,450
尽管如此，英雄出场率仍然暴降

267
00:12:50,450 --> 00:12:53,070
就连胜率都少了点

268
00:12:53,070 --> 00:12:56,650
我说玩家心理也是个重要因素对吧？

269
00:12:57,570 --> 00:12:59,500
那么在节目开头

270
00:12:59,500 --> 00:13:03,020
我曾说“平衡性”是关于对水平相近的玩家

271
00:13:03,020 --> 00:13:05,820
让每种选项都各有潜力

272
00:13:05,820 --> 00:13:09,300
但如果玩家水平参差不齐会怎样？

273
00:13:09,300 --> 00:13:12,800
嗯，许多重度竞技游戏都有匹配系统

274
00:13:12,800 --> 00:13:15,000
配对水平相近的玩家

275
00:13:15,000 --> 00:13:17,870
但对于更易上手的聚会游戏

276
00:13:17,870 --> 00:13:20,520
我们也许想有个“负反馈循环”

277
00:13:20,520 --> 00:13:21,950
或“追赶机制”

278
00:13:21,950 --> 00:13:25,020
向表现不佳的玩家伸出援手

279
00:13:25,020 --> 00:13:28,120
例子包括《现代战争2》的“连死补助”

280
00:13:28,120 --> 00:13:30,870
给死得太多的你一份特殊奖励

281
00:13:30,870 --> 00:13:32,870
还有“马力欧赛车”的道具系统

282
00:13:32,870 --> 00:13:36,120
像蓝龟壳那样恼人的强力道具

283
00:13:36,120 --> 00:13:38,900
只会送给跑在队尾的玩家

284
00:13:38,900 --> 00:13:42,900
这些设计很有争议，必须谨慎而行

285
00:13:42,900 --> 00:13:46,620
我们也能加入更多运气成分降低门槛

286
00:13:46,620 --> 00:13:51,070
这在《蛇梯棋》和《大富翁》等家庭桌游中屡见不鲜

287
00:13:51,070 --> 00:13:53,850
它们非常依赖掷骰子的运气

288
00:13:53,850 --> 00:13:57,420
而在电子游戏中，这可见于《Apex英雄》

289
00:13:57,420 --> 00:13:59,770
你的赢面并非固定

290
00:13:59,770 --> 00:14:03,420
得看你着陆时捡到什么好东西

291
00:14:03,420 --> 00:14:06,050
游戏还能提供“让分模式”

292
00:14:03,420 --> 00:14:06,050
（为弱者提供优势选项来弥补差距）

293
00:14:06,050 --> 00:14:07,550
而在团队游戏中

294
00:14:07,550 --> 00:14:11,770
我们可以提供更多玩法，让新手也能贡献力量

295
00:14:11,770 --> 00:14:14,850
而不一定要求冲锋在前

296
00:14:14,850 --> 00:14:17,520
比如担任医生或工程师

297
00:14:18,990 --> 00:14:22,100
所以打造平衡性颇具挑战

298
00:14:22,100 --> 00:14:24,230
角色越是差异化

299
00:14:24,230 --> 00:14:26,950
让他们公平竞技就越难

300
00:14:26,950 --> 00:14:30,500
这还没考虑玩家水平参差不齐的情况

301
00:14:30,500 --> 00:14:32,470
我们可以先从“制衡”出发

302
00:14:32,470 --> 00:14:35,020
确保角色不会过强

303
00:14:35,020 --> 00:14:36,670
然后设计相克角色

304
00:14:36,670 --> 00:14:38,300
让他们相互牵制

305
00:14:38,300 --> 00:14:43,050
但即使是最好的设计也经不起数百万玩家的挑剔

306
00:14:43,050 --> 00:14:45,320
所以我们得不断检查平衡性

307
00:14:45,320 --> 00:14:47,500
观察胜率、对阵胜率

308
00:14:47,500 --> 00:14:49,320
出场率和玩家反馈

309
00:14:49,320 --> 00:14:53,250
虽然玩家趋势变化可能会自然修正不平衡

310
00:14:53,250 --> 00:14:56,800
但有时开发者还是得动手修改

311
00:14:56,800 --> 00:15:00,620
然后你引入一名全新角色，一切又得重来

312
00:15:00,620 --> 00:15:03,350
哎，我说过这不简单

313
00:15:03,350 --> 00:15:07,000
所以你觉得最平衡的游戏是什么？

314
00:15:07,000 --> 00:15:10,620
你是否玩过平衡性差得离谱的游戏？

315
00:15:10,620 --> 00:15:14,800
请在下方评论中分享你的体验吧

316
00:15:15,550 --> 00:15:16,670
感谢收看

317
00:15:16,670 --> 00:15:20,750
本期节目得到很多人帮助，从多人游戏大神

318
00:15:20,750 --> 00:15:24,470
到一线开发者，他们参与过《英雄联盟》、《脏弹》

319
00:15:24,470 --> 00:15:26,000
和《彩虹六号：围攻》

320
00:15:26,000 --> 00:15:29,470
这个话题还有很多可聊，比如多人射击游戏的地图平衡性

321
00:15:29,470 --> 00:15:31,800
但我们可以来日再叙

322
00:15:31,800 --> 00:15:34,170
现在嘛，谢谢，拜拜