提交 bf27db8d 编写于 作者: V viviwong

add subtitle

上级 fe978ea8
1
00:00:00,210 --> 00:00:03,009
we will be talking about self-interested
我们将谈论自私的
2
00:00:03,209 --> 00:00:04,658
agents and their interactions
代理及其互动
3
00:00:04,859 --> 00:00:07,149
so let's first speak about what we mean
所以我们先说说我们的意思
4
00:00:07,349 --> 00:00:14,290
by self-interested agents we don't mean
对于自私的经纪人,我们并不是说
5
00:00:14,490 --> 00:00:17,829
necessarily that agents are adversarial
必然是代理商具有对抗性
6
00:00:18,028 --> 00:00:21,010
or don't care about what happened to
或不在乎发生了什么
7
00:00:21,210 --> 00:00:24,220
other agents what we mean by that is
其他代理商,我们的意思是
8
00:00:24,420 --> 00:00:26,318
that agents have opinions have
代理商有意见
9
00:00:26,518 --> 00:00:28,949
preferences and so there's some
偏好,所以有一些
10
00:00:29,149 --> 00:00:31,568
description of the world how the world
世界的描述世界如何
11
00:00:31,768 --> 00:00:33,518
could be and in different descriptions
可能是并且在不同的描述中
12
00:00:33,719 --> 00:00:36,669
the agents have different preferences
代理商有不同的偏好
13
00:00:36,869 --> 00:00:39,608
and different utilities as we'll say and
以及我们将要说的不同的实用程序
14
00:00:39,808 --> 00:00:43,329
so what we mean by utility function is a
所以我们所说的效用函数是
15
00:00:43,530 --> 00:00:47,619
mathematical measure that tells you how
告诉你如何的数学测量
16
00:00:47,820 --> 00:00:50,709
much the agents likes or does not like a
代理商非常喜欢或不喜欢
17
00:00:50,909 --> 00:00:52,119
given situation
给定情况
18
00:00:52,320 --> 00:00:57,428
it describes not only their attitude
它不仅描述了他们的态度
19
00:00:57,628 --> 00:01:00,279
towards a definite of events so for
对于一定的事件,所以
20
00:01:00,479 --> 00:01:02,198
example tomorrow the temperature will be
明天的温度将是
21
00:01:02,399 --> 00:01:07,988
exactly 25 degrees centigrade but but in
正好25摄氏度,但
22
00:01:08,188 --> 00:01:09,730
fact it will describe the preferences
实际上它将描述偏好
23
00:01:09,930 --> 00:01:15,129
towards a distribution of such outcomes
分配这样的结果
24
00:01:15,329 --> 00:01:18,039
so it really captures their attitude
所以这真的抓住了他们的态度
25
00:01:18,239 --> 00:01:21,399
towards uncertainty about events so for
对事件的不确定性
26
00:01:21,599 --> 00:01:26,200
example if I tell you that it will be 25
例如,如果我告诉你它将是25
27
00:01:26,400 --> 00:01:30,069
degrees with probability 0.7 and 24
度为0.7和24
28
00:01:30,269 --> 00:01:33,308
degrees with probability 0.3 you might
度,概率为0.3,您可能
29
00:01:33,509 --> 00:01:35,500
have an opinion about how much you like
对你喜欢多少有意见
30
00:01:35,700 --> 00:01:38,399
that versus some other distribution and
与其他发行版相比,
31
00:01:38,599 --> 00:01:42,429
the decision theoretic approach which is
决策理论方法是
32
00:01:42,629 --> 00:01:45,759
what underlies modern game theory says
现代博弈论的基础是什么
33
00:01:45,959 --> 00:01:47,469
that you're going to try to act in the
你将要尝试采取行动
34
00:01:47,670 --> 00:01:50,528
way that maximizes your expected or
最大化您的预期或
35
00:01:50,728 --> 00:01:54,640
average utility and so this is a concept
平均效用,所以这是一个概念
36
00:01:54,840 --> 00:01:59,439
we need to get comfortable with and and
我们需要对and和
37
00:01:59,640 --> 00:02:02,230
it's not obvious that one would want to
人们是否愿意
38
00:02:02,430 --> 00:02:06,128
use such an approach so for example we
使用这种方法,例如
39
00:02:06,328 --> 00:02:08,800
are going to look at a single dimension
将要看一个单一的维度
40
00:02:09,000 --> 00:02:11,170
so your preferences will
所以你的喜好会
41
00:02:11,370 --> 00:02:15,849
all beyond a scale as we'll see the
一切都超出规模,我们将看到
42
00:02:16,049 --> 00:02:17,920
scale that is not that important unlike
规模不那么重要
43
00:02:18,120 --> 00:02:19,750
probabilities utilities don't have to
概率实用程序不必
44
00:02:19,949 --> 00:02:22,810
lie in the 0-1 scale but they will lie
说谎在0-1规模,但他们会说谎
45
00:02:23,009 --> 00:02:26,830
on a linear dimension and maybe maybe
在线性维度上,也许
46
00:02:27,030 --> 00:02:29,860
that's inappropriate or for example you
那是不合适的,例如你
47
00:02:30,060 --> 00:02:32,950
might have some level of wealth and some
可能有一定程度的财富和一些
48
00:02:33,150 --> 00:02:37,270
degree of health and for a certain level
健康程度和一定水平
49
00:02:37,469 --> 00:02:41,679
of of each one you'll have some some
每一个中,您都会有一些
50
00:02:41,878 --> 00:02:46,179
some some notion of well-being but is it
一些关于幸福的概念,但这是
51
00:02:46,378 --> 00:02:48,670
appropriate to put the two together and
适合将两者放在一起
52
00:02:48,870 --> 00:02:50,439
have a single scale you might you might
有一个单一的规模,你可能会
53
00:02:50,639 --> 00:02:54,819
question that similarly why is looking
同样为什么要看的问题
54
00:02:55,019 --> 00:02:56,649
at the expected value when you're
当您处于期望值时
55
00:02:56,848 --> 00:02:58,659
looking at your uncertainty while
看着你的不确定性
56
00:02:58,859 --> 00:03:00,939
looking at the expected value an
看期望值
57
00:03:01,139 --> 00:03:02,860
appropriate way to capture your attitude
捕捉态度的适当方法
58
00:03:03,060 --> 00:03:07,330
and so these are not trivial statements
所以这些不是琐碎的陈述
59
00:03:07,530 --> 00:03:10,209
and in fact are not tautological they
而且实际上不是重言式的
60
00:03:10,408 --> 00:03:13,569
make a subjective claim but there's a
提出主观主张,但有一个
61
00:03:13,769 --> 00:03:16,599
very long tradition and maybe the most
悠久的传统,也许是最悠久的
62
00:03:16,799 --> 00:03:20,319
famous reference is to phenomena
著名的参考是现象
63
00:03:20,519 --> 00:03:26,560
Morgenstern simul book on on which is
Morgenstern simul上的书是
64
00:03:26,759 --> 00:03:28,179
really in some ways the introduction to
确实在某种程度上介绍了
65
00:03:28,378 --> 00:03:31,000
modern day in game theory that derives
博弈论中的现代派生
66
00:03:31,199 --> 00:03:34,990
the these utility function from more
这些实用功能更多
67
00:03:35,189 --> 00:03:38,649
basic assumptions one makes and we won't
一个基本假设,我们不会
68
00:03:38,848 --> 00:03:40,990
go into that but we just wanted to flag
进入那个,但是我们只想标记
69
00:03:41,189 --> 00:03:43,840
this issue as something that will
这个问题会
70
00:03:44,039 --> 00:03:48,550
underlie everything we say about game
成为我们谈论游戏的基础
71
00:03:48,750 --> 00:03:51,340
theory and which really underlies one
理论,这实际上是一个基础
72
00:03:51,539 --> 00:03:56,539
game theory
博弈论
1
00:00:00,030 --> 00:00:03,969
in the celebrated example of prisoner's
在囚犯的著名例子中
2
00:00:04,169 --> 00:00:05,859
dilemma we have this general scheme
困境,我们有这个总体方案
3
00:00:06,059 --> 00:00:08,530
where the two prisoners can either
两个囚犯可以
4
00:00:08,730 --> 00:00:12,550
cooperate or not defect as is called if
合作或不存在缺陷
5
00:00:12,750 --> 00:00:15,460
they both cooperate they get some payoff
他们俩都合作,他们得到了一些回报
6
00:00:15,660 --> 00:00:17,530
a-and if they both defect they get a
-如果他们俩都叛逆,他们会得到一个
7
00:00:17,730 --> 00:00:20,199
different payoff D where a is greater
a较大时的不同收益D
8
00:00:20,399 --> 00:00:24,669
than D however if they miss coordinate
比D但是如果他们错过坐标
9
00:00:24,868 --> 00:00:26,980
and one of them cooperates and the other
其中一个合作,另一个
10
00:00:27,179 --> 00:00:30,460
defects then the cooperator gets the
缺陷,然后合作者得到
11
00:00:30,660 --> 00:00:32,768
lowest possible payoff and the defector
最低的回报和叛逃者
12
00:00:32,969 --> 00:00:34,899
gets the largest possible path and
得到最大可能的路径
13
00:00:35,100 --> 00:00:37,329
that's true symmetrically here as well
在这里对称也是如此
14
00:00:37,530 --> 00:00:41,320
and this a very well-known example that
这是一个非常著名的例子
15
00:00:41,520 --> 00:00:44,099
has a rather counterintuitive
有点违反直觉
16
00:00:44,299 --> 00:00:49,088
paradoxical properties most games are
大多数游戏的悖论性质
17
00:00:49,289 --> 00:00:55,689
not as conceptually confusing here's an
没有概念上的混乱,这是
18
00:00:55,890 --> 00:00:58,149
example that conceptually very clear and
这个例子在概念上非常清楚,
19
00:00:58,350 --> 00:01:00,788
these are gains of pure competition the
这些都是纯竞争的收益
20
00:01:00,988 --> 00:01:03,070
situation here is limited to two players
这种情况仅限于两名球员
21
00:01:03,270 --> 00:01:07,140
where one players payoff is exactly the
一个玩家的收益正好是
22
00:01:07,340 --> 00:01:10,869
complement of another players payoff so
补充另一个玩家的收益,所以
23
00:01:11,069 --> 00:01:13,988
they always sum to some constant C often
他们经常总和一些常数C
24
00:01:14,188 --> 00:01:18,009
that concept is that we use zero and we
这个概念是我们使用零,我们
25
00:01:18,209 --> 00:01:20,230
call them for that reason zero-sum games
称他们为零和游戏
26
00:01:20,430 --> 00:01:22,119
as opposed to constant sum games
与固定和博弈相反
27
00:01:22,319 --> 00:01:27,849
and since they do sum to zero or to
由于它们的总和为零或
28
00:01:28,049 --> 00:01:31,119
constant we only need to remember one
不变,我们只需要记住一个
29
00:01:31,319 --> 00:01:33,278
number the path to one of the players
编号玩家之一的路径
30
00:01:33,478 --> 00:01:36,668
and we can infer the payoff to the other
我们可以推断其他人的收益
31
00:01:36,868 --> 00:01:41,528
player from that here's the most simple
玩家从这里最简单
32
00:01:41,728 --> 00:01:43,808
version of it this is a games of
它的版本是
33
00:01:44,009 --> 00:01:46,599
matching pennies so you and I each need
匹配便士,所以您和我都需要
34
00:01:46,799 --> 00:01:49,840
to pick either heads or tail for the for
为之挑选头或尾
35
00:01:50,040 --> 00:01:54,189
the coin if we pick the same side either
如果我们选择同一面的话
36
00:01:54,390 --> 00:01:58,149
heads or tails I win which means that I
头或尾我赢了,这意味着我
37
00:01:58,349 --> 00:02:01,599
get a payoff of 1 and U of minus 1 if we
如果我们得到1的收益和-1的U
38
00:02:01,799 --> 00:02:04,569
miss coordinate and so I pick heads and
想念坐标,所以我挑了头
39
00:02:04,769 --> 00:02:06,878
you tailed or the other way around then
你拖尾或反过来
40
00:02:07,078 --> 00:02:10,000
you win a very straightforward game of
您会赢得一个非常简单的游戏
41
00:02:10,199 --> 00:02:12,080
pure competition
纯竞争
42
00:02:12,280 --> 00:02:14,210
here's another word our very well-known
这是我们非常知名的另一个词
43
00:02:14,409 --> 00:02:17,090
similar games with three actions both of
具有三个动作的相似游戏
44
00:02:17,289 --> 00:02:20,060
us and that's the game of rock papers
我们就是摇滚游戏
45
00:02:20,259 --> 00:02:21,130
and scissors
和剪刀
46
00:02:21,330 --> 00:02:25,670
also known as Rochambeau and so if we
也被称为Rochambeau,所以如果我们
47
00:02:25,870 --> 00:02:29,360
pick the same action then it's a draw
选择相同的动作然后平局
48
00:02:29,560 --> 00:02:33,500
and otherwise there are rules for who
否则有谁的规则
49
00:02:33,699 --> 00:02:36,020
wins for example if I pick Rock and you
例如,如果我选择Rock而您获胜
50
00:02:36,219 --> 00:02:38,210
paper the new one if I picked Rock and
如果我选择了Rock,
51
00:02:38,409 --> 00:02:42,800
use scissors then I win and so on again
用剪刀,然后我赢了,依此类推
52
00:02:43,000 --> 00:02:46,600
the payoffs in both cases sum to zero
两种情况下的收益总和为零
53
00:02:46,800 --> 00:02:52,490
this parenthetically this very simple
这很简单
54
00:02:52,689 --> 00:02:55,700
children game actually has an annual
儿童游戏实际上每年都有
55
00:02:55,900 --> 00:02:58,430
competition that carries a non-trivial
不平凡的竞争
56
00:02:58,629 --> 00:03:01,580
prize of $10,000 and it's actually a
奖金$ 10,000,实际上是
57
00:03:01,780 --> 00:03:02,540
sobering thought
清醒的思想
58
00:03:02,740 --> 00:03:05,150
that when we look at this trivial game
当我们看这个琐碎的游戏时
59
00:03:05,349 --> 00:03:08,330
then perhaps check a little bit if we
然后也许检查一下
60
00:03:08,530 --> 00:03:10,759
actually participate in this competition
参加比赛
61
00:03:10,959 --> 00:03:12,230
we'd actually think hard about how to
我们实际上会认真思考如何
62
00:03:12,430 --> 00:03:18,890
play it is the other extreme of games of
玩是游戏的另一极端
63
00:03:19,090 --> 00:03:22,670
pure coordination or pure cooperation in
纯粹的协调或纯粹的合作
64
00:03:22,870 --> 00:03:25,939
this case all agents have exactly the
在这种情况下,所有代理商都有
65
00:03:26,139 --> 00:03:28,039
same interest in other words their
换句话说,他们的兴趣相同
66
00:03:28,239 --> 00:03:32,569
payoffs for every action vector that
每个动作向量的收益
67
00:03:32,769 --> 00:03:37,610
they take is the same and so the utility
他们采取的是相同的,所以实用程序
68
00:03:37,810 --> 00:03:40,039
for play I is always the same of utility
对于玩游戏,我永远都是相同的工具
69
00:03:40,239 --> 00:03:44,539
for pair J for every action sequence
对于J对每个动作序列
70
00:03:44,739 --> 00:03:47,990
that they choose and so again we here -
他们选择,所以我们再次在这里-
71
00:03:48,189 --> 00:03:50,450
we only need to write in each cell of
我们只需要在
72
00:03:50,650 --> 00:03:52,160
the matrix only one number because it's
矩阵只有一个数字,因为
73
00:03:52,360 --> 00:03:55,099
common to all the players it drives home
它带回家的所有球员都是共同的
74
00:03:55,299 --> 00:03:58,240
the perhaps the unfortunate term
也许是不幸的术语
75
00:03:58,439 --> 00:04:01,250
noncorporeal game theory that describes
非物质博弈论
76
00:04:01,449 --> 00:04:03,259
this dominant strand of game theory
博弈论的主导
77
00:04:03,459 --> 00:04:06,439
we're discussing for now it's the name
我们现在正在讨论的是名字
78
00:04:06,639 --> 00:04:08,710
would suggest that these are games for
会建议这些是
79
00:04:08,909 --> 00:04:10,910
that describe situations that are
描述情况
80
00:04:11,110 --> 00:04:14,180
inherent inherently conflictual but as
内在固有的冲突,但作为
81
00:04:14,379 --> 00:04:17,598
we see they apply also to games in which
我们看到它们也适用于其中
82
00:04:17,798 --> 00:04:20,750
the interests of the players coincide so
玩家的利益是一致的
83
00:04:20,949 --> 00:04:23,350
here's a here's a game that
这是一个游戏
84
00:04:23,550 --> 00:04:28,600
describes the purely court cooperative
描述纯粹的法院合作社
85
00:04:28,800 --> 00:04:31,259
situation you and I walk each other
你和我互相走动的情况
86
00:04:31,459 --> 00:04:34,120
towards each other on the sidewalk we
在人行道上彼此接近
87
00:04:34,319 --> 00:04:36,129
can each decide whether to go to our
每个人都可以决定是否去我们的
88
00:04:36,329 --> 00:04:37,840
respective lay lifts our respective
各自的升降机
89
00:04:38,040 --> 00:04:40,900
right and if we pick the same side then
对,如果我们选择同一边,那么
90
00:04:41,100 --> 00:04:44,079
all is good we avoid a collision if we
一切都很好,如果我们避免碰撞
91
00:04:44,279 --> 00:04:48,550
don't then the then we do collide and
然后,那么我们就发生了碰撞,
92
00:04:48,750 --> 00:04:52,810
that's equally bad for both of us of
这对我们两个人都同样有害
93
00:04:53,009 --> 00:04:55,110
course in general gains will be neither
当然,总的收益不会
94
00:04:55,310 --> 00:04:58,020
purely cooperative nor purely
纯粹合作也不是纯粹
95
00:04:58,220 --> 00:05:02,460
conflictual and here's a game that
冲突,这是一个游戏
96
00:05:02,660 --> 00:05:05,050
exemplifies that this is a game that
举例说明这是一款
97
00:05:05,250 --> 00:05:07,900
called battle of the sexes imagine a
所谓的性别之战想象
98
00:05:08,100 --> 00:05:10,629
husband and a wife we want to go out to
丈夫和妻子,我们想出去
99
00:05:10,829 --> 00:05:13,300
a movie there are two movies they could
电影可以有两部电影
100
00:05:13,500 --> 00:05:16,870
choose from let's say Battle of
从说“战役”中选择
101
00:05:17,069 --> 00:05:21,759
Armageddon and flower child the one a
大决战和花童一个
102
00:05:21,959 --> 00:05:25,900
violent war movie and the other is a is
暴力战争电影,另一个是
103
00:05:26,100 --> 00:05:30,100
a romantic comedy above all they want to
他们想要的浪漫喜剧
104
00:05:30,300 --> 00:05:31,660
go together to the movie if they go to
一起去看电影,如果他们去
105
00:05:31,860 --> 00:05:34,840
different movies then they are equally
不同的电影,他们都是一样的
106
00:05:35,040 --> 00:05:36,910
unhappy so they want to go to the same
不开心,所以他们想去同一个
107
00:05:37,110 --> 00:05:38,530
movie but they have conflicting
电影,但他们有冲突
108
00:05:38,730 --> 00:05:41,980
preferences the wife clearly would
妻子的喜好显然会
109
00:05:42,180 --> 00:05:47,250
prefer to go to battle of Armageddon and
宁愿去参加世界末日之战,
110
00:05:47,449 --> 00:05:50,379
husband romantic as he is would like to
丈夫浪漫,因为他想
111
00:05:50,579 --> 00:05:51,460
go to flower child
去花童
112
00:05:51,660 --> 00:05:56,860
so both cooperation and competition in
因此,在
113
00:05:57,060 --> 00:06:02,060
this game
这个游戏
1
00:00:00,000 --> 00:00:02,889
hi folks its Matt again and now we're
嗨,大家好,Matt,现在我们
2
00:00:03,089 --> 00:00:04,509
ready to start solving games and making
准备开始解决游戏和制作
3
00:00:04,710 --> 00:00:06,519
some predictions of how people will play
关于人们如何玩的一些预测
4
00:00:06,719 --> 00:00:08,500
in different settings and so we're
在不同的设置中,所以我们
5
00:00:08,699 --> 00:00:11,859
talking right now about Nash equilibrium
现在谈论纳什均衡
6
00:00:12,058 --> 00:00:14,950
which is one probably the most basic and
这可能是最基本的
7
00:00:15,150 --> 00:00:17,260
standard solution concept of all in all
全部的标准解决方案概念
8
00:00:17,460 --> 00:00:19,600
of game theory it's named after John
博弈论的名字,以约翰命名
9
00:00:19,800 --> 00:00:21,310
Nash who was a mathematician at
纳什(Nash)是数学家
10
00:00:21,510 --> 00:00:23,470
Princeton and actually some years back
普林斯顿,实际上几年前
11
00:00:23,670 --> 00:00:25,329
won the Nobel Prize for his work on this
他在这方面的工作获得了诺贝尔奖
12
00:00:25,528 --> 00:00:28,510
subject and it's a very basic and
主题,这是一个非常基本的
13
00:00:28,710 --> 00:00:31,329
fundamental concept and in order to sort
基本概念和排序
14
00:00:31,528 --> 00:00:33,070
of motivate it let's let's start by just
激励它,让我们从一开始就开始
15
00:00:33,270 --> 00:00:38,349
talking through some a particular game
通过特定游戏进行交谈
16
00:00:38,549 --> 00:00:41,320
that was described and invented by
被描述和发明的
17
00:00:41,520 --> 00:00:45,189
another famous person so this is John
另一个名人,所以这是约翰
18
00:00:45,390 --> 00:00:48,219
Kane John Maynard Keynes beauty contest
凯恩·约翰·梅纳德·凯恩斯选美大赛
19
00:00:48,420 --> 00:00:51,998
game so what's the idea here so let's
游戏,这是什么主意,让我们
20
00:00:52,198 --> 00:00:53,919
let's think of a basic situation that
让我们考虑一个基本情况
21
00:00:54,119 --> 00:00:57,038
you might be interested in and this was
您可能对此感兴趣,这是
22
00:00:57,238 --> 00:00:58,658
one that Keynes described in some detail
凯恩斯对此进行了详细描述
23
00:00:58,859 --> 00:01:01,419
so the idea was you have a stock you're
所以想法是你有一只股票
24
00:01:01,619 --> 00:01:04,388
holding on to it and the stock price is
坚持下去,股价就是
25
00:01:04,588 --> 00:01:07,090
rising that's great you're an investor
上升那太好了,您是投资者
26
00:01:07,290 --> 00:01:08,769
you're trying to make profits off of
你试图从中获利
27
00:01:08,969 --> 00:01:12,278
your your stock holdings and you begin
您的库存量,然后开始
28
00:01:12,478 --> 00:01:13,869
to believe that maybe the stock is too
相信也许股票也是如此
29
00:01:14,069 --> 00:01:15,849
high to be justified by the value of the
高要由价值来证明
30
00:01:16,049 --> 00:01:17,799
company so you're thinking that it's
公司,所以您认为这是
31
00:01:18,000 --> 00:01:19,599
possible that this stock is overvalued
这只股票可能被高估了
32
00:01:19,799 --> 00:01:22,209
maybe there's a bubble in the market and
也许市场上有泡沫,
33
00:01:22,409 --> 00:01:24,629
you're starting to think about selling
您开始考虑出售
34
00:01:24,829 --> 00:01:27,308
okay well you'd like to sell it but
好吧,你想卖掉它,但是
35
00:01:27,509 --> 00:01:28,840
you'd like to wait until the price is at
您想等到价格在
36
00:01:29,040 --> 00:01:31,090
its peak right so you'd want to wait
它的峰值合适,所以您想等待
37
00:01:31,290 --> 00:01:33,009
until the price was just where it's
直到价格刚好在
38
00:01:33,209 --> 00:01:35,909
going to hit its maximum before you sold
在您卖出之前将达到最高
39
00:01:36,109 --> 00:01:37,959
so you wanted to get out of the market
所以你想退出市场
40
00:01:38,159 --> 00:01:40,539
just before the other investors do so
就在其他投资者这样做之前
41
00:01:40,739 --> 00:01:41,950
this is a game where now you have to
这是一个游戏,现在您必须
42
00:01:42,150 --> 00:01:43,390
predict what other people think about
预测别人的想法
43
00:01:43,590 --> 00:01:45,189
the stock price and and what they're
股票价格以及它们的含义
44
00:01:45,390 --> 00:01:46,778
going to do and when they want to get
去做,什么时候想得到
45
00:01:46,978 --> 00:01:50,168
out so how will they act how should you
出来,他们将如何行动,你应该如何
46
00:01:50,368 --> 00:01:53,049
respond to that so this is the basic
对此做出回应,这是最基本的
47
00:01:53,250 --> 00:01:55,149
ingredients of Nash equilibrium are
纳什均衡的成分是
48
00:01:55,349 --> 00:01:57,009
going to be having some prediction of
会有一些预测
49
00:01:57,209 --> 00:01:58,778
what other players are doing and then
其他玩家在做什么,然后
50
00:01:58,978 --> 00:02:00,579
choosing the optimal strategy in
选择最佳策略
51
00:02:00,780 --> 00:02:01,959
response to that so these are they're
回应,所以这些是
52
00:02:02,159 --> 00:02:03,579
going to be two key ingredients that we
将成为我们的两个关键要素
53
00:02:03,780 --> 00:02:07,000
have so there's a very stylized version
有一个非常程式化的版本
54
00:02:07,200 --> 00:02:09,880
of this which is known as the the Keynes
这就是凯恩斯
55
00:02:10,080 --> 00:02:12,850
beauty contest game where did it come
选美比赛是哪里来的
56
00:02:13,050 --> 00:02:13,600
from
57
00:02:13,800 --> 00:02:17,140
actually Keynes described the there was
实际上凯恩斯描述了
58
00:02:17,340 --> 00:02:22,950
a newspaper in the in England that had a
英格兰的一家报纸
59
00:02:23,150 --> 00:02:26,500
contest where players had to guess which
比赛中玩家必须猜测哪个
60
00:02:26,699 --> 00:02:32,290
picture of several women other readers
其他读者的几个女人的照片
61
00:02:32,490 --> 00:02:36,460
would think was the the the most
会认为是最
62
00:02:36,659 --> 00:02:38,290
attractive one so it wasn't to guess
有吸引力的一个,所以不用猜测
63
00:02:38,490 --> 00:02:39,910
what you thought but what you thought
你在想什么,但你在想什么
64
00:02:40,110 --> 00:02:43,620
other people were thinking so Keynes
其他人在想,所以凯恩斯
65
00:02:43,819 --> 00:02:47,290
likened investing to this you it's not
比作投资,这不是
66
00:02:47,490 --> 00:02:49,900
only what you think of this the stock
只有您对此的看法
67
00:02:50,099 --> 00:02:51,340
but what you think other people are
但是你认为别人是什么
68
00:02:51,539 --> 00:02:52,540
thinking about the stock that's
考虑那只股票
69
00:02:52,740 --> 00:02:54,130
important and driving your decisions
重要并推动您的决策
70
00:02:54,330 --> 00:02:57,280
okay so this now is represented by a
好吧,现在以
71
00:02:57,479 --> 00:02:59,530
very simple game that is is played by
玩的非常简单的游戏
72
00:02:59,729 --> 00:03:02,230
many people so what's this game look
很多人,所以这个游戏看起来像什么
73
00:03:02,430 --> 00:03:06,400
like each person gets to name an integer
就像每个人都用一个整数命名
74
00:03:06,599 --> 00:03:08,890
between 1 and 100 ok so you get to pick
在1到100 ok之间,所以您可以选择
75
00:03:09,090 --> 00:03:11,230
a number between 1 and 100 s to be an
1到100秒之间的数字是一个
76
00:03:11,430 --> 00:03:15,340
integer so 1 2 3 etc players are going
整数,所以1 2 3等玩家进行
77
00:03:15,539 --> 00:03:17,800
to move simultaneously and the player
与玩家同时移动
78
00:03:18,000 --> 00:03:19,900
who names the integer that's closest to
谁命名最接近的整数
79
00:03:20,099 --> 00:03:23,020
two-thirds of the average integer wins a
平均整数的三分之二赢得
80
00:03:23,219 --> 00:03:25,900
prize and the other players get nothing
奖,其他玩家一无所获
81
00:03:26,099 --> 00:03:28,509
so to win this game you have to get you
所以要赢得这场比赛,你必须让你
82
00:03:28,709 --> 00:03:32,050
have to guess the average and then 2/3
必须猜测平均值,然后再2/3
83
00:03:32,250 --> 00:03:34,120
of it right so you want to be right at
它是正确的,所以你想在正确
84
00:03:34,319 --> 00:03:36,280
2/3 of whatever the average guess is so
平均猜测的2/3
85
00:03:36,479 --> 00:03:39,520
a little bit below the average guess if
比平均猜测低一点
86
00:03:39,719 --> 00:03:41,289
there's two people that happen to hit
有两个人碰巧
87
00:03:41,489 --> 00:03:43,240
the same integer that it's the the right
正确的整数
88
00:03:43,439 --> 00:03:46,090
one then ties are going to be broken
一个然后领带将被打破
89
00:03:46,289 --> 00:03:47,890
uniformly at random so we'll just flip a
均匀地随机分布,所以我们只需要翻转一个
90
00:03:48,090 --> 00:03:49,710
coin or if there's three people will
硬币,或者如果有三个人
91
00:03:49,909 --> 00:03:54,520
roll a dice a three-sided die etc okay
掷骰子,三面骰子等
92
00:03:54,719 --> 00:03:57,789
so how would you play this game you have
所以你会怎么玩这个游戏
93
00:03:57,989 --> 00:03:58,930
to think about what other players are
考虑其他玩家
94
00:03:59,129 --> 00:04:01,600
going to do and then forecast what you
去做,然后预测你
95
00:04:01,800 --> 00:04:04,120
think the best integer is in a response
认为最好的整数在响应中
96
00:04:04,319 --> 00:04:09,319
to that
那个
1
00:00:00,000 --> 00:00:03,099
I will now speak about how we predict
我现在将谈论我们如何预测
2
00:00:03,299 --> 00:00:05,139
that people will play in a game and
人们会玩游戏,
3
00:00:05,339 --> 00:00:07,780
specifically about the concepts of best
特别是关于最佳概念
4
00:00:07,980 --> 00:00:12,100
response and Nash equilibrium now let's
反应和纳什均衡现在让我们
5
00:00:12,300 --> 00:00:14,319
go home were to assume that you are one
回家假设你是一个
6
00:00:14,519 --> 00:00:17,530
of the players and the other players are
的球员和其他球员是
7
00:00:17,730 --> 00:00:19,109
going to play in a way that you know
会以您知道的方式演奏
8
00:00:19,309 --> 00:00:22,239
then we can speak about what is your
那我们可以说说你是什么
9
00:00:22,439 --> 00:00:24,039
best thing to do and we'll call that the
最好的事情,我们称之为
10
00:00:24,239 --> 00:00:27,550
best response and technically speaking
最佳回应,从技术上讲
11
00:00:27,750 --> 00:00:31,149
let's use the following notation let's
让我们使用以下表示法
12
00:00:31,349 --> 00:00:36,099
call AI the action sequence of everybody
把AI称为每个人的行动顺序
13
00:00:36,299 --> 00:00:39,878
except you player I so it's player 1 and
除了你我是玩家1
14
00:00:40,079 --> 00:00:42,909
so on internal player I minus 1 I plus 1
所以内部玩家我减去1我加1
15
00:00:43,109 --> 00:00:46,619
and all the rest so the entire action
其余的一切,整个动作
16
00:00:46,820 --> 00:00:49,649
vector we call it the action profile a
向量,我们将其称为行动概况
17
00:00:49,850 --> 00:00:53,320
is really made up of all those a minus I
真的是由所有减去我的人组成
18
00:00:53,520 --> 00:00:58,329
and your own action with that notation
以及您使用该符号的操作
19
00:00:58,530 --> 00:01:01,149
here's the definition we will say that
这是我们要说的定义
20
00:01:01,350 --> 00:01:03,849
your best response which we'll call AI
您最好的回应,我们称之为AI
21
00:01:04,049 --> 00:01:07,689
star it may not be unique but it's got
明星可能不是唯一的但有
22
00:01:07,890 --> 00:01:10,599
to be part of the set it will call BR
作为集合的一部分,它将调用BR
23
00:01:10,799 --> 00:01:13,269
the best response assuming that all the
假设所有
24
00:01:13,469 --> 00:01:17,709
other agents play is a a - I if it's the
其他特工打的是aa-如果是
25
00:01:17,909 --> 00:01:20,890
case that for any anything you might be
不管你做什么,
26
00:01:21,090 --> 00:01:24,250
thinking of doing a sub I for all a sub
想为所有子项目做一个子项目
27
00:01:24,450 --> 00:01:27,399
I it's got to be the case that your
我一定要这样
28
00:01:27,599 --> 00:01:31,028
utility for playing that in other words
换句话说,实用程序
29
00:01:31,228 --> 00:01:34,899
when you playing AI star and the others
当你玩AI之星和其他人
30
00:01:35,099 --> 00:01:38,500
are playing AI a - I for whatever it is
在玩AI a-我不管
31
00:01:38,700 --> 00:01:40,778
has got to be at least as great as
必须至少与
32
00:01:40,978 --> 00:01:44,278
anything else you might choose a sub I
还有其他你可以选择的子
33
00:01:44,478 --> 00:01:47,439
that's the case we'll say that a AI star
就是这样,我们会说一个AI明星
34
00:01:47,640 --> 00:01:51,069
is a best response very intuitive notion
最好的回应是非常直观的概念
35
00:01:51,269 --> 00:01:55,119
and that would bring us all almost all
那将带给我们几乎所有人
36
00:01:55,319 --> 00:01:57,369
the way there the problem of course is
那里的问题当然是
37
00:01:57,569 --> 00:01:59,319
that you don't know what the others will
你不知道别人会怎么做
38
00:01:59,519 --> 00:02:02,378
do but that's ok because when we use the
可以,但是没关系,因为当我们使用
39
00:02:02,578 --> 00:02:04,778
notion of best response as a building
最佳响应的概念
40
00:02:04,978 --> 00:02:07,028
block to define what we call the Nash
定义我们称为Nash的块
41
00:02:07,228 --> 00:02:09,850
equilibrium and the National Caribbean
均衡与加勒比国家联盟
42
00:02:10,050 --> 00:02:12,520
briefly is a set of acts
简单来说是一套行为
43
00:02:12,719 --> 00:02:14,950
one for each of the agents such that
每个代理一个
44
00:02:15,150 --> 00:02:17,730
each is the best response to the others
每个都是对其他人的最佳回应
45
00:02:17,930 --> 00:02:20,469
specifically we'll look at an action
具体来说,我们来看看一个动作
46
00:02:20,669 --> 00:02:24,850
profile here a a one through a n and
在这里通过一个和一个配置文件
47
00:02:25,050 --> 00:02:26,710
we'll say that it's a Nash equilibrium
我们会说那是纳什均衡
48
00:02:26,909 --> 00:02:29,560
and later on we'll tell you why we call
稍后我们会告诉您为什么我们打电话
49
00:02:29,759 --> 00:02:32,140
it specifically a pure strategy Nash
它专门是一种纯粹的策略纳什
50
00:02:32,340 --> 00:02:35,020
equilibrium if it's the case that for
如果是这样的话
51
00:02:35,219 --> 00:02:38,650
every agent that action a sub I
每个行动一个子我
52
00:02:38,849 --> 00:02:40,120
associate with agent is your best
与代理商合作是您最好的选择
53
00:02:40,319 --> 00:02:45,319
response all the rest
其余所有回应
1
00:00:00,149 --> 00:00:03,188
let us now look at some examples and of
现在让我们来看一些例子
2
00:00:03,388 --> 00:00:07,609
games and Nash equilibria in those games
游戏和那些游戏中的纳什均衡
3
00:00:08,000 --> 00:00:10,390
so here's the first game a familiar game
所以这是第一个熟悉的游戏
4
00:00:10,589 --> 00:00:14,189
this is of course the prisoner's dilemma
这当然是囚犯的困境
5
00:00:14,388 --> 00:00:18,190
the if both prisoners cooperate and
如果两个囚犯合作,
6
00:00:18,390 --> 00:00:20,859
content and then they get a light
内容,然后他们就亮了
7
00:00:21,059 --> 00:00:25,568
punishment and if they do not cooperate
惩罚,如果他们不配合
8
00:00:25,768 --> 00:00:28,839
they get a more severe punishment if the
他们会受到更严厉的惩罚
9
00:00:29,039 --> 00:00:30,550
one cooperates and the others does not
一个合作而其他不合作
10
00:00:30,750 --> 00:00:33,309
then the co-op irrigated terrible
然后合作社灌溉很糟糕
11
00:00:33,509 --> 00:00:36,608
punishment and the one that does not
惩罚与不惩罚
12
00:00:36,808 --> 00:00:38,500
cooperate gets off scot it gets off
合作下车苏格兰人下车
13
00:00:38,700 --> 00:00:41,378
scot-free and of course this game has a
无苏格兰人,当然这个游戏有
14
00:00:41,579 --> 00:00:45,038
dominant strategy to defect no matter
主导缺陷策略无论如何
15
00:00:45,238 --> 00:00:47,649
what the other agent does you're better
另一个代理商你做得更好
16
00:00:47,850 --> 00:00:52,570
off not cooperating and so of course the
不合作,所以当然
17
00:00:52,770 --> 00:00:56,018
only dominant strategy outcome is this
唯一的主要策略结果是
18
00:00:56,219 --> 00:00:58,689
one of both defecting and indeed that is
既有缺陷,又有缺陷的是
19
00:00:58,890 --> 00:01:00,788
the only Nash equilibrium in this game
这场比赛中唯一的纳什均衡
20
00:01:00,988 --> 00:01:03,729
so it's the Nash equilibrium it's the
所以这是纳什均衡
21
00:01:03,929 --> 00:01:05,560
best response if the other person
如果对方最佳反应
22
00:01:05,760 --> 00:01:07,659
defects then it's the best respond to
缺陷,那么它是最好的应对方法
23
00:01:07,859 --> 00:01:09,969
defect but in fact it's much stronger
缺陷,但实际上要强大得多
24
00:01:10,170 --> 00:01:11,709
than that it's best to defect no matter
最好不要有任何缺点
25
00:01:11,909 --> 00:01:14,890
what the other the other agent does so
另一个代理人这样做
26
00:01:15,090 --> 00:01:17,439
this is an example of one unique Nash
这是一个独特的纳什的例子
27
00:01:17,640 --> 00:01:19,090
equilibrium that happened to be very
恰好是非常均衡
28
00:01:19,290 --> 00:01:21,278
strong one a dominant strategy Nash
强势策略Nash
29
00:01:21,478 --> 00:01:25,659
equilibrium so so here's another game
平衡,所以这是另一个游戏
30
00:01:25,859 --> 00:01:27,549
this is the game of pure coordination I
这是纯粹协调的游戏
31
00:01:27,750 --> 00:01:31,299
think of it as walking towards each
认为它是走向每个
32
00:01:31,500 --> 00:01:33,789
other on the sidewalk and you both can
其他人行道上,你们俩都可以
33
00:01:33,989 --> 00:01:35,950
decide whether to go to your respective
决定是否去你各自的
34
00:01:36,150 --> 00:01:40,769
lifts or respective rights in both cases
两种情况下的升降机或各自的权利
35
00:01:40,969 --> 00:01:43,659
you will do fine and you will not
你会做得很好,你不会
36
00:01:43,859 --> 00:01:45,549
collide and of course if you miss
碰撞,当然,如果您想念
37
00:01:45,750 --> 00:01:47,619
coordinate if you one goes to the left
协调,如果你去左边
38
00:01:47,819 --> 00:01:48,609
and the other to the right you will
另一个在右边
39
00:01:48,810 --> 00:01:52,628
collide so this is a natural game and in
碰撞,所以这是自然的游戏
40
00:01:52,828 --> 00:01:55,359
fact you see that you have two Nash
事实上你看到你有两个纳什
41
00:01:55,560 --> 00:01:58,028
equilibria the one that I wrote down
平衡我写下的那一个
42
00:01:58,228 --> 00:02:01,869
here if one one of the players goes to
如果一位玩家去
43
00:02:02,069 --> 00:02:03,730
the left it's the best respond to go to
左边是最好的回应
44
00:02:03,930 --> 00:02:06,609
the left and conversely if the other
左边,反之亦然
45
00:02:06,810 --> 00:02:08,469
player goes to the right you're best off
玩家向右走,你最好
46
00:02:08,669 --> 00:02:10,838
going to the right as well and the
也会向右走,
47
00:02:11,038 --> 00:02:12,550
others are not Nash equilibria so here's
其他人不是纳什均衡,所以这里的
48
00:02:12,750 --> 00:02:13,090
an exam
考试
49
00:02:13,289 --> 00:02:17,649
of a game where there are two Nash
一个游戏中有两个纳什
50
00:02:17,848 --> 00:02:20,560
equilibria or to specifically pure
均衡或特纯
51
00:02:20,759 --> 00:02:23,200
strategy Nash equilibria again we'll
策略纳什均衡再一次
52
00:02:23,400 --> 00:02:25,929
discuss why we call these pure strategy
讨论为什么我们称这些为纯策略
53
00:02:26,128 --> 00:02:31,110
later on here's a very different game
稍后这是一个非常不同的游戏
54
00:02:31,310 --> 00:02:33,879
this is often called the game of battle
这通常被称为战斗游戏
55
00:02:34,079 --> 00:02:37,480
of the sexes imagine a couple and they
的性别想象一对夫妇,他们
56
00:02:37,680 --> 00:02:39,039
want to go together to a movie and
想一起去看电影,
57
00:02:39,239 --> 00:02:41,379
they're considering two movies one of
他们正在考虑两部电影之一
58
00:02:41,579 --> 00:02:45,399
them a very violent movie Battle of the
他们是一部非常暴力的电影
59
00:02:45,598 --> 00:02:49,569
Titans and the other very relaxed movie
泰坦和另一部非常轻松的电影
60
00:02:49,769 --> 00:02:53,640
about flower growing called Lee's B&F
关于花的成长,叫做李的B&F
61
00:02:53,840 --> 00:02:56,409
the wife of course would prefer to go to
妻子当然更愿意去
62
00:02:56,609 --> 00:02:58,899
battle of the Titans and the the husband
泰坦与丈夫之战
63
00:02:59,098 --> 00:03:01,599
would prefer to watch flower growing but
宁愿看花开,但
64
00:03:01,799 --> 00:03:03,219
more than anything else they would want
比他们想要的更重要
65
00:03:03,419 --> 00:03:05,080
to go together and so here are the
一起去,所以这里是
66
00:03:05,280 --> 00:03:09,520
payoffs if they both go to battle of the
如果他们都参加战斗的话
67
00:03:09,719 --> 00:03:12,849
Titans then they're both positively
泰坦然后他们都是积极的
68
00:03:13,049 --> 00:03:15,580
happy their wife more than the husband
妻子比丈夫更快乐
69
00:03:15,780 --> 00:03:18,670
if the go both go to the flower growing
如果两者都去花开
70
00:03:18,870 --> 00:03:22,599
movie then the husband is slightly happy
电影,然后丈夫有点高兴
71
00:03:22,799 --> 00:03:24,340
and the wife but if they go to different
和妻子,但如果他们去不同的地方
72
00:03:24,539 --> 00:03:26,409
movies neither of them is happy that's
电影他们俩都不开心
73
00:03:26,609 --> 00:03:30,340
that's that's the that's that's the the
那就是那那就是那
74
00:03:30,539 --> 00:03:36,999
game how many how many equilibria we
游戏我们多少平衡
75
00:03:37,199 --> 00:03:40,629
have here well again we have two pure
再来一遍,我们有两个纯
76
00:03:40,829 --> 00:03:43,390
strategy Nash equilibria why is that
策略纳什均衡为什么
77
00:03:43,590 --> 00:03:49,030
well if either them goes to the Battle
好吧,如果他们去参加战斗
78
00:03:49,229 --> 00:03:50,439
of the Titans then the other one would
的泰坦,然后另一个
79
00:03:50,639 --> 00:03:52,659
want to go there too because if they go
也想去那里,因为如果他们去
80
00:03:52,859 --> 00:03:54,969
to a different one they would get zero
换成另一种,他们将得到零
81
00:03:55,169 --> 00:03:57,969
rather than whatever they get here one
而不是他们来到这里
82
00:03:58,169 --> 00:03:59,499
or two depending on whether the husband
一两个取决于丈夫
83
00:03:59,699 --> 00:04:02,140
of the wife and conversely on the on the
的妻子,相反在
84
00:04:02,340 --> 00:04:04,300
flower watching movie flower grow movie
花看电影花种电影
85
00:04:04,500 --> 00:04:07,090
and so in both cases they were the best
所以在两种情况下它们都是最好的
86
00:04:07,289 --> 00:04:09,999
response is to go to the movie selected
回应是去看电影
87
00:04:10,199 --> 00:04:13,360
by the other party so on the face of it
对方如此面对
88
00:04:13,560 --> 00:04:15,159
it looks very similar to the game of
它看起来非常类似于
89
00:04:15,359 --> 00:04:17,499
pure coordination that we have here but
我们在这里有纯粹的协调,但是
90
00:04:17,699 --> 00:04:19,180
we do see a slight difference here and
我们在这里确实看到了细微的差别,
91
00:04:19,379 --> 00:04:21,310
it will revisit that later on when we
稍后,当我们
92
00:04:21,509 --> 00:04:23,860
speak about not pure strategies but
谈论的不是纯粹的策略,而是
93
00:04:24,060 --> 00:04:25,670
mixed strategies
混合策略
94
00:04:25,870 --> 00:04:28,259
here's a here's another example the last
这是一个这是另一个例子
95
00:04:28,459 --> 00:04:30,090
one we'll look at and this is the game
我们来看一看,这就是游戏
96
00:04:30,290 --> 00:04:32,939
called matching pennies imagine each of
所谓的匹配便士想象每个
97
00:04:33,139 --> 00:04:35,218
us two players needing to just need to
我们两个玩家只需要
98
00:04:35,418 --> 00:04:38,430
decide on some side of a of a coin heads
在硬币头的a的某一侧决定
99
00:04:38,629 --> 00:04:41,879
or tail if we decide on the same size
还是尾巴,如果我们决定相同的尺寸
100
00:04:42,079 --> 00:04:44,730
heads or tail but we decide on the same
头或尾,但我们决定相同
101
00:04:44,930 --> 00:04:48,960
one then then I win if we decide on
一个然后我赢了,如果我们决定
102
00:04:49,160 --> 00:04:51,090
different sides you heads and me tails
你的头和我的尾巴的不同侧面
103
00:04:51,290 --> 00:04:54,449
or vice versa the new win and so we see
反之亦然,新的胜利,所以我们看到
104
00:04:54,649 --> 00:04:59,999
this here if we both decides on heads or
如果我们俩都决定要这样做,还是在这里
105
00:05:00,199 --> 00:05:01,949
we both decides on tails I win and
我们都决定我赢的尾巴,
106
00:05:02,149 --> 00:05:03,900
otherwise you win by winning I mean I
不然你赢了就是我
107
00:05:04,100 --> 00:05:05,759
get 1 you get minus 1 so this is a
得到1你得到负1所以这是一个
108
00:05:05,959 --> 00:05:10,338
zero-sum game the sum of our payoff is 0
零和游戏,我们的收益之和为0
109
00:05:10,538 --> 00:05:13,439
what is a pure strategy Nash equilibrium
什么是纯策略纳什均衡
110
00:05:13,639 --> 00:05:16,710
here let's think about it suppose I pick
让我们考虑一下,假设我选择了
111
00:05:16,910 --> 00:05:20,400
head what is your best response well
领导什么是你最好的反应
112
00:05:20,600 --> 00:05:22,740
your best response then is to pick tails
你最好的选择就是挑尾巴
113
00:05:22,939 --> 00:05:26,100
because you get one rather than minus
因为你得到一个而不是减
114
00:05:26,300 --> 00:05:31,350
one but if you pick I tails then my best
一个,但是如果你选择我尾巴,那我最好
115
00:05:31,550 --> 00:05:33,540
response is not to play tail because I
回应不是打尾巴,因为我
116
00:05:33,740 --> 00:05:36,000
want to coordinate with you because then
想和你协调,因为
117
00:05:36,199 --> 00:05:39,660
I will get one rather than minus one
我会得到一个而不是减一
118
00:05:39,860 --> 00:05:42,360
that I would be getting here but now if
我会到这里,但现在如果
119
00:05:42,560 --> 00:05:47,009
I play tails you'd rather play heads
我打尾巴你宁可打头
120
00:05:47,209 --> 00:05:50,189
because you'd get one rather than the
因为你会得到一个而不是
121
00:05:50,389 --> 00:05:53,009
minus one you're getting here but again
减一,你要来这里,但是再来一次
122
00:05:53,209 --> 00:05:55,528
if you're playing tails I want to if you
如果你在打尾巴,我想
123
00:05:55,728 --> 00:05:57,300
playing heads I want to play heads to
我要打头
124
00:05:57,500 --> 00:06:01,528
match so we have this cycle where the
匹配,所以我们有这个周期
125
00:06:01,728 --> 00:06:07,230
best responses are leading us in the
最好的回应正在引领我们
126
00:06:07,430 --> 00:06:10,800
cycle and so there is no pure strategy
周期,所以没有纯粹的策略
127
00:06:11,000 --> 00:06:12,629
Nash equilibrium in this game of mashing
这场捣蛋游戏中的纳什均衡
128
00:06:12,829 --> 00:06:17,829
pennies
便士
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册