1
00:00:03,440 --> 00:00:05,760
They say life is a competition.

2
00:00:08,320 --> 00:00:10,400
The fight to be first.

3
00:00:12,600 --> 00:00:15,120
The survival of the fittest.

4
00:00:16,960 --> 00:00:17,960
Winning.

5
00:00:19,800 --> 00:00:23,200
Nobody wants to be a loser.

6
00:00:23,200 --> 00:00:27,280
Now, winning basically means getting
more of what you want in life,

7
00:00:27,280 --> 00:00:29,960
and, as a mathematician, I can help
with that.

8
00:00:29,960 --> 00:00:33,240
Because there is a real science
to success,

9
00:00:33,240 --> 00:00:37,640
based on ideas from some of the last
century's most beautiful minds.

10
00:00:37,640 --> 00:00:42,320

That's the truth.

11
00:00:42,320 --> 00:00:47,080
From how you can analyse and take
advantage of your opponents...

12
00:00:47,080 --> 00:00:50,400
You need to find a way to be as
unpredictable as possible.

13
00:00:51,720 --> 00:00:55,160
..to deciding whether to cheat or
play fair...

14
00:00:55,160 --> 00:00:57,600
You can put a good apple in a barrel
of bad apples,

15
00:00:57,600 --> 00:00:59,600
it's always going to turn bad.

16
00:00:59,600 --> 00:01:03,160
..being more strategic will help you
get ahead.

17
00:01:04,240 --> 00:01:07,280
But if all of this sounds a bit
selfish,

18
00:01:07,280 --> 00:01:12,000
I'll also be showing you how
cooperation can trump conflict,

19
00:01:12,000 --> 00:01:14,920
and how everyone can be a winner.

20
00:01:14,920 --> 00:01:17,640
Hello, Sheriff.
Hi, Hannah. How are you?

21
00:01:17,640 --> 00:01:20,960
You are walking away with no money
because you are an idiot.
No. You're an idiot.

22
00:01:22,920 --> 00:01:27,080
Whether it is happier relationships
you are after, a bargain,

23
00:01:27,080 --> 00:01:29,040
or a better world,

24
00:01:29,040 --> 00:01:32,240
a look at the science of winning can
help

25
00:01:32,240 --> 00:01:34,880
and reveal some surprising truths.

26
00:01:36,280 --> 00:01:39,000
So join me.
You'll lose out if you don't.

27
00:01:52,120 --> 00:01:57,160
To get the best result in any
situation, maths can help.

28
00:01:57,240 --> 00:01:59,640
Let's take just one example.

29
00:01:59,640 --> 00:02:02,800
Something as simple as ordering
dinner.

30
00:02:02,800 --> 00:02:06,360
Should you choose burger or steak?

31
00:02:09,080 --> 00:02:11,880
Now, you absolutely cannot
afford it,

32
00:02:11,880 --> 00:02:15,000
but that steak does look incredibly
tempting.

33
00:02:15,000 --> 00:02:17,080
And it's almost your birthday,

34
00:02:17,080 --> 00:02:21,720
surely you deserve a £25 splurge
every now and then?

35
00:02:21,720 --> 00:02:26,760
First of all, ask yourself, what do
you really want -

36
00:02:26,880 --> 00:02:29,320
steak or savings?

37
00:02:30,360 --> 00:02:34,360
Well, you need to decide
what winning means for you.

38
00:02:34,360 --> 00:02:37,600
And that might change
when your mates arrive.

39
00:02:39,760 --> 00:02:43,440
Let's imagine that everybody has
exactly the same choice.

40
00:02:43,440 --> 00:02:45,600
Burger or steak.

41
00:02:45,600 --> 00:02:49,520
But you know that your friends are
notoriously frugal, so

42
00:02:49,520 --> 00:02:53,320
if they all go for the burger and
you split the bill evenly,

43
00:02:53,320 --> 00:02:57,880
you could end up getting
that tasty steak for a lot less.

44
00:02:58,800 --> 00:03:01,240
The £25 hit from your steak

45
00:03:01,240 --> 00:03:05,080
combined with the £5 for their
burgers means

46
00:03:05,080 --> 00:03:07,320
everyone pays £10 each,

47
00:03:07,320 --> 00:03:10,680
which frankly, for you,
is an amazing bargain.

48
00:03:12,040 --> 00:03:16,280
But hold on. The waiter is coming
over and is heading towards

49
00:03:16,280 --> 00:03:18,960
your friend, who you know loves
a rib eye.

50
00:03:18,960 --> 00:03:21,520
Is she thinking
what you're thinking?

51
00:03:21,520 --> 00:03:23,800
And what about the others?

52
00:03:23,800 --> 00:03:28,840
And so begins a frankly very British
game of silent guessing and

53
00:03:29,360 --> 00:03:32,480
calculating. If she goes for the
steak as well as you,

54
00:03:32,480 --> 00:03:35,600
then you're shelling out £15 each,

55
00:03:35,600 --> 00:03:40,040
but if he goes for the steak too,
then that's £20 a head,

56
00:03:40,040 --> 00:03:43,880
which is far more than you can
possibly afford.

57
00:03:43,880 --> 00:03:46,520
But is it really fair that your best
friend,

58
00:03:46,520 --> 00:03:49,280
who always does the generous thing,

59
00:03:49,280 --> 00:03:53,280
ends up spending £20
for a measly burger?

60
00:03:54,960 --> 00:03:57,600
So what is the real win here?

61
00:03:57,600 --> 00:03:59,840
Well, it's complicated.

62
00:03:59,840 --> 00:04:02,600
Do you prefer saving money,

63
00:04:02,600 --> 00:04:07,280
a delicious dinner,
or being fair to your friend?

64
00:04:07,280 --> 00:04:10,680
Having been fleeced, will she ever
want to meet up again?

65
00:04:12,080 --> 00:04:16,040
The key thing is that what you want
can be scored,

66
00:04:16,040 --> 00:04:20,680
and mathematicians like me give it a
fancy name - your expected utility.

67
00:04:20,680 --> 00:04:23,600
We express your preferences

68
00:04:23,600 --> 00:04:26,800
as values that capture just how much

69
00:04:26,800 --> 00:04:29,840
you instinctively want something,

70
00:04:29,840 --> 00:04:34,840
against the price, and then factor
in other stuff, like risk.

71
00:04:35,120 --> 00:04:37,560
In this case, that your dear friend

72
00:04:37,560 --> 00:04:40,520
might ditch you for busting
her budget.

73
00:04:42,600 --> 00:04:45,400
Looks like steak pipped friendship
this time.

74
00:04:47,600 --> 00:04:49,200
So there you go.

75
00:04:49,200 --> 00:04:51,920
If you can figure out everyone's
preferences

76
00:04:51,920 --> 00:04:53,720
and you know what YOU want,

77
00:04:53,720 --> 00:04:58,040
then the key question becomes,
at the lowest cost to yourself,

78
00:04:58,040 --> 00:05:01,960
without having any control over
other people's actions,

79
00:05:01,960 --> 00:05:04,480
how do you get what YOU want?

80
00:05:07,760 --> 00:05:10,160
Well, thankfully, mathematics can

81
00:05:10,160 --> 00:05:12,920
help you find a winning strategy,
too.

82
00:05:12,920 --> 00:05:15,920
Whatever decisions we make,

83
00:05:15,920 --> 00:05:18,600
having a good strategy means the
difference

84
00:05:18,600 --> 00:05:21,480
between knowing what you want
and getting it.

85
00:05:23,080 --> 00:05:27,320
It was the mid-20th-century maths
genius John von Neumann

86
00:05:27,320 --> 00:05:29,240
who, more than anyone else,

87
00:05:29,240 --> 00:05:32,360
turned our winning instincts into
a science.

88
00:05:35,040 --> 00:05:39,400
Von Neumann established a completely
new field of study,

89
00:05:39,400 --> 00:05:43,520
called Game Theory, which was
concerned with the mathematics

90
00:05:43,520 --> 00:05:46,440
of both cooperation and of conflict.

91
00:05:46,440 --> 00:05:50,520
How to avoid losing
and, more importantly, how to win.

92
00:05:50,520 --> 00:05:53,120
His efforts culminated in this here.

93
00:05:53,120 --> 00:05:57,720
A Theory Of Games And Economic
Behaviour, 641 pages,

94
00:05:57,720 --> 00:06:00,400
packed to the brim with formulas,

95
00:06:00,400 --> 00:06:04,560
which was once described as one of
the most-influential

96
00:06:04,560 --> 00:06:07,960
and least-read books
of the last century.

97
00:06:09,600 --> 00:06:14,600
Written with the economist Oskar
Morgenstern and published in 1944,

98
00:06:15,040 --> 00:06:17,800
it proved how we can all get ahead

99
00:06:17,800 --> 00:06:21,280
when one party's loss
is the other's gain.

100
00:06:23,880 --> 00:06:26,800
It's not for the mathematically
faint-hearted,

101
00:06:26,800 --> 00:06:30,320
but at the core of the book is a
strategic principle

102
00:06:30,320 --> 00:06:33,080
that is surprisingly intuitive.

103
00:06:33,080 --> 00:06:36,760
And easily explained with some
cold, hard cash.

104
00:06:38,200 --> 00:06:40,000
So, you and me are here.

105
00:06:40,000 --> 00:06:42,320
I'm going to have a go on those slot
machines

106
00:06:42,320 --> 00:06:44,800
and I have got a nice bag of
2p coins

107
00:06:44,800 --> 00:06:47,200
that you and I are going to share.

108
00:06:47,200 --> 00:06:51,320
Now, the polite way to do this is
just to let you help yourself.

109
00:06:51,320 --> 00:06:54,480
After all, I trust you to play fair.

110
00:06:55,680 --> 00:06:58,280
But actually, I really shouldn't.

111
00:06:58,280 --> 00:07:01,120
Von Neumann assumes, rightly so,

112
00:07:01,120 --> 00:07:04,360
that we both want
as many coins as we can get.

113
00:07:04,360 --> 00:07:07,040
So, what SHOULD I do?

114
00:07:07,040 --> 00:07:12,080
Well, I should split this into
two piles, as evenly as possible,

115
00:07:12,640 --> 00:07:15,680
and then let you decide
which half you want.

116
00:07:15,680 --> 00:07:19,640
Given that you're watching me like a
hawk, the only way I can guarantee

117
00:07:19,640 --> 00:07:23,440
I don't lose out to you is
by me splitting and you choosing.

118
00:07:28,400 --> 00:07:33,080
I don't get as many 2ps as I might
like, but I win by minimising the

119
00:07:33,080 --> 00:07:37,360
maximum amount of fun at the fair
that I can lose to you.

120
00:07:39,440 --> 00:07:41,720
This is known

121
00:07:41,720 --> 00:07:45,560
as the min-max theory,
and as strategies go,

122
00:07:45,560 --> 00:07:47,720
you can apply it to pretty much
anything.

123
00:07:47,720 --> 00:07:50,720
So, if you have got two kids that
are fighting over toys,

124
00:07:50,720 --> 00:07:52,920
just get one of them to make two
piles

125
00:07:52,920 --> 00:07:56,600
and the other one to pick which
pile they prefer,

126
00:07:56,600 --> 00:08:01,240
or if you've got boring household
chores, you write two lists,

127
00:08:01,240 --> 00:08:03,520
and your partner chooses which list.

128
00:08:06,960 --> 00:08:08,520
It sounds pretty simple,

129
00:08:08,520 --> 00:08:13,120
but you have to remember that your
split will take into account

130
00:08:13,120 --> 00:08:16,560
your preferences, so you get to
decide if you think

131
00:08:16,560 --> 00:08:19,640
the utility of a huge pile of
ironing and doing

132
00:08:19,640 --> 00:08:22,600
all of the hoovering is the
equivalent of

133
00:08:22,600 --> 00:08:26,840
the utility of unblocking the loo,
and sorting out all of the bins.

134
00:08:26,840 --> 00:08:31,480
Now, strangely, Von Neumann didn't
get into that particular conundrum

135
00:08:31,480 --> 00:08:35,480
in his book, but his proof that you
can always minimise the losses

136
00:08:35,480 --> 00:08:39,000
and maximise the gains, while taking
into account

137
00:08:39,000 --> 00:08:43,880
the preferences of two warring
parties was hugely influential.

138
00:08:46,120 --> 00:08:48,760
Such an analytical approach
to winning

139
00:08:48,760 --> 00:08:51,000
gripped people's imagination.

140
00:08:52,640 --> 00:08:55,920
And crucially, Von Neumann took
inspiration

141
00:08:55,920 --> 00:08:58,520
from one rather unpredictable game.

142
00:09:00,880 --> 00:09:04,440
Von Neumann was an exceptional
mathematician.

143
00:09:04,440 --> 00:09:07,000
And perhaps because of its strategic
complexity,

144
00:09:07,000 --> 00:09:10,320
he was particularly interested
in poker.

145
00:09:10,320 --> 00:09:13,840
He said that real life is about
bluffing,

146
00:09:13,840 --> 00:09:16,760
about little tactics of deception,

147
00:09:16,760 --> 00:09:21,760
about asking yourself, what does
the other man think I mean to do?

148
00:09:22,160 --> 00:09:26,200
"And that is what games are about
in my theory."

149
00:09:29,280 --> 00:09:32,040
When it comes to winning big at
poker,

150
00:09:32,040 --> 00:09:34,600
there's one woman you need to know.

151
00:09:34,600 --> 00:09:39,600
Wow! This is bold.
Four bet of 4,100.

152
00:09:40,200 --> 00:09:43,960
Sick move. Liv Boeree is the
European top-ranking

153
00:09:43,960 --> 00:09:45,680
female poker player.

154
00:09:45,680 --> 00:09:47,720
Really strong play by Liv Boeree,

155
00:09:47,720 --> 00:09:50,000
she's clearly been eating
her vitamins.

156
00:09:50,000 --> 00:09:52,320
Vitamin B - for bluff.

157
00:10:06,840 --> 00:10:10,960
I'm meeting Liv to find out how
poker strategy might help us

158
00:10:10,960 --> 00:10:13,920
get more of what we want in life.

159
00:10:15,320 --> 00:10:19,680
These days, the very best players
are very analytically minded folks

160
00:10:19,680 --> 00:10:22,240
who are comfortable working with
maths,

161
00:10:22,240 --> 00:10:25,040
to sort of really get a solid
understanding

162
00:10:25,040 --> 00:10:28,320
of the theory behind the game.
And then the psychology is,

163
00:10:28,320 --> 00:10:30,560
I like to think of it
as the cherry on top.

164
00:10:30,560 --> 00:10:32,480
OK, let's talk about the basics,
then.

165
00:10:32,480 --> 00:10:34,960
What calculations are you making?

166
00:10:36,560 --> 00:10:39,480
Um, if I'm going to bet chips, if
it's 30% of the time

167
00:10:39,480 --> 00:10:42,720
I'm going to be the best hand,
and 70% of the time I'm going to be

168
00:10:42,720 --> 00:10:44,600
the worst, then I can multiply those

169
00:10:44,600 --> 00:10:48,840
numbers against the chips that are
involved, and that'll be my expected

170
00:10:48,840 --> 00:10:50,840
value from a situation.

171
00:10:50,840 --> 00:10:54,480
And that applies in many real-life
situations, too.

172
00:10:54,480 --> 00:10:57,000
Say I am running late for a flight.

173
00:10:57,000 --> 00:10:58,480
I have two options. I can either...

174
00:10:59,600 --> 00:11:02,440
..go to the airport and try to make
my flight but run the risk

175
00:11:02,440 --> 00:11:04,880
of then missing it and losing the
value of the flight,

176
00:11:04,880 --> 00:11:06,520
and having to buy a new one,

177
00:11:06,520 --> 00:11:09,360
or I can stay home and do the sort
of safer bet,

178
00:11:09,360 --> 00:11:11,400
and call and pay the change fee.

179
00:11:11,400 --> 00:11:14,480
Thinking rationally and quantifying
things is the very first step to

180
00:11:14,480 --> 00:11:16,320
thinking strategically.
Exactly. Yes.

181
00:11:16,320 --> 00:11:19,360
It's very hard to build strategies

182
00:11:19,360 --> 00:11:23,800
if you don't have a numerical idea
of the value of the outcomes

183
00:11:23,800 --> 00:11:25,760
of different strategies.

184
00:11:30,280 --> 00:11:33,560
By evaluating both the chances of
something happening

185
00:11:33,560 --> 00:11:35,760
and the outcome when it does,

186
00:11:35,760 --> 00:11:38,840
you have a better view of your best
options.

187
00:11:41,200 --> 00:11:45,880
But at this level, everyone is just
as calculating.

188
00:11:45,880 --> 00:11:49,720
To beat the optimal strategy,
you have to find an edge.

189
00:11:53,040 --> 00:11:56,360
There something that's called
game theory optimal, where,

190
00:11:56,360 --> 00:12:00,560
if you're playing this style, then
it means that you are unexploitable,

191
00:12:00,560 --> 00:12:05,360
basically. The best case that an
opponent can do to counter it

192
00:12:05,360 --> 00:12:08,040
is to also play
game theory optimal,

193
00:12:08,040 --> 00:12:12,320
therefore you just sort of reach the
same very high standard of play,

194
00:12:12,320 --> 00:12:16,160
and you're trying to not deviate
from it at all.

195
00:12:16,160 --> 00:12:18,960
And that's where sort of the more,
like, creativity

196
00:12:18,960 --> 00:12:21,880
and, I guess, the artistic side of
the game comes in,

197
00:12:21,880 --> 00:12:25,120
where you can do these deviations
from the optimal play

198
00:12:25,120 --> 00:12:27,160
to exploit your opponent's
weaknesses.

199
00:12:27,160 --> 00:12:29,000
But they could be doing that to you
as well.

200
00:12:29,000 --> 00:12:34,040
Exactly. So you need to find a way
to be as unpredictable as possible.

201
00:12:35,560 --> 00:12:39,080
Von Neumann's maths proves that
poker players must bluff,

202
00:12:39,080 --> 00:12:40,760
unpredictably,

203
00:12:40,760 --> 00:12:44,440
to avoid exploitation
by a savvy opponent.

204
00:12:44,440 --> 00:12:49,400
It is an essential strategy and it
can be applied elsewhere.

205
00:12:49,560 --> 00:12:52,360
You know, if you're in a business
negotiation,

206
00:12:52,360 --> 00:12:54,320
what is the minimum amount
you will take?

207
00:12:54,320 --> 00:12:58,000
There's a great opportunity to bluff
and give a high number because

208
00:12:58,000 --> 00:13:00,040
you're giving yourself the option

209
00:13:00,040 --> 00:13:02,040
for them to pay you
that high number.

210
00:13:02,040 --> 00:13:05,080
It might seem obvious, but bluffing
like this

211
00:13:05,080 --> 00:13:08,200
is exactly what gives you the edge
over others.

212
00:13:08,200 --> 00:13:11,360
Particularly if their loss
is your gain.

213
00:13:12,920 --> 00:13:15,920
Anything that is a competition,
to make yourself unexploitable

214
00:13:15,920 --> 00:13:18,040
will just put you at a huge
advantage.

215
00:13:20,280 --> 00:13:22,600
So quantify everything, folks.

216
00:13:22,600 --> 00:13:26,280
And don't make yourself
too easy to read.

217
00:13:26,280 --> 00:13:29,920
In win-lose games your carefully
calculated strategies

218
00:13:29,920 --> 00:13:34,880
can only get you what you want
if your opponents are in the dark.

219
00:13:43,000 --> 00:13:47,480
But what if you're involved in a
conflict where everyone could lose?

220
00:13:51,200 --> 00:13:53,880
During the second half
of the 20th century,

221
00:13:53,880 --> 00:13:57,400
the biggest problem that game
theorists has had to contend with

222
00:13:57,400 --> 00:14:00,160
was the Cold War.
The US and the Soviet Union

223
00:14:00,160 --> 00:14:03,640
were facing off against each other,
and just as in poker,

224
00:14:03,640 --> 00:14:08,680
every strategy, every move and
countermove had to be analysed.

225
00:14:09,400 --> 00:14:13,880
Except, this time, the stakes could
not have been higher.

226
00:14:17,240 --> 00:14:21,640
As an ally to the US, Britain had
to prepare for the worst.

227
00:14:23,640 --> 00:14:28,640
In 1961, a top-secret facility was
set up here in rural Worcestershire,

228
00:14:29,400 --> 00:14:32,080
Regional Seat of Government 9.

229
00:14:33,640 --> 00:14:35,360
It is one of a network of bunkers

230
00:14:35,360 --> 00:14:37,560
from which the British Government
would

231
00:14:37,560 --> 00:14:40,760
have operated in the event
of a nuclear war.

232
00:14:45,880 --> 00:14:48,560
The moment that the Second World War

233
00:14:48,560 --> 00:14:52,600
ended with such a devastating
show of nuclear force,

234
00:14:52,600 --> 00:14:56,160
the talk in the West began to focus
on trying to prevent

235
00:14:56,160 --> 00:15:00,280
a military face-off that would mean
that nuclear bunkers like this one

236
00:15:00,280 --> 00:15:02,280
would become a necessity.

237
00:15:02,280 --> 00:15:06,920
Now, the strategy that was on the
table at the time was referred to,

238
00:15:06,920 --> 00:15:10,960
somewhat euphemistically,
as preventative war.

239
00:15:10,960 --> 00:15:15,240
The idea being that the Americans
would launch an unprovoked attack

240
00:15:15,240 --> 00:15:20,320
on the Soviet Union before they had
the chance to acquire the bomb.

241
00:15:24,880 --> 00:15:27,720
But it would at least avoid a slow,

242
00:15:27,720 --> 00:15:31,520
expensive and far more dangerous
cold one.

243
00:15:34,040 --> 00:15:38,800
Preventative war is an argument
still used today over North Korea

244
00:15:38,800 --> 00:15:43,040
and at the time, its supporters
included Winston Churchill

245
00:15:43,040 --> 00:15:44,920
and von Neumann himself.

246
00:15:48,120 --> 00:15:52,760
In 1950, von Neumann remarked,
"With the Russians,

247
00:15:52,760 --> 00:15:56,480
"it's not a question of whether,
but of when.

248
00:15:56,480 --> 00:16:00,480
"If you say, 'Why not bomb them
tomorrow?' I say, 'Why not today?'

249
00:16:00,480 --> 00:16:03,240
"And if you say,
'Today at five o'clock,'

250
00:16:03,240 --> 00:16:05,480
"I say, 'Why not one o'clock?'"

251
00:16:05,480 --> 00:16:08,560
You can see where this kind of
reasoning ends,

252
00:16:08,560 --> 00:16:10,640
and it is perfectly rational.

253
00:16:10,640 --> 00:16:14,640
If someone has to win, better
that it be us.

254
00:16:18,040 --> 00:16:23,040
Despite the impeccable logic,
America didn't attack Russia.

255
00:16:23,600 --> 00:16:26,640
And then it was too late
because Russia got the bomb, too.

256
00:16:27,920 --> 00:16:32,920
By 1953, there was no outcome in
which only one side could win.

257
00:16:35,360 --> 00:16:37,960
This demanded a new strategy.

258
00:16:37,960 --> 00:16:40,120
Something which became known as

259
00:16:40,120 --> 00:16:44,680
mutually assured destruction,
or MAD for short. Basically,

260
00:16:44,680 --> 00:16:49,360
the threat of you using massively
destructive weapons on your enemy

261
00:16:49,360 --> 00:16:53,440
prevents your enemy from using
those same weapons on you.

262
00:16:53,440 --> 00:16:55,280
Once everybody is armed,

263
00:16:55,280 --> 00:17:00,320
nobody has an incentive to either
initiate a conflict or to disarm.

264
00:17:01,720 --> 00:17:06,600
And so, nuclear war was avoided,
but places like this

265
00:17:06,600 --> 00:17:11,520
and a hugely expensive, vast global
nuclear arsenal, was the result.

266
00:17:16,880 --> 00:17:19,120
Compared to von Neumann's games,

267
00:17:19,120 --> 00:17:22,120
in which one person's loss is the
other's gain,

268
00:17:22,120 --> 00:17:25,080
mutually assured destruction,
strangely,

269
00:17:25,080 --> 00:17:28,480
is something we come across far more
in real life.

270
00:17:29,960 --> 00:17:33,600
Because most conflicts, whether
at work, with friends or family

271
00:17:33,600 --> 00:17:37,320
have the potential to end up
in a stalemate.

272
00:17:40,840 --> 00:17:45,880
So often we choose strategies
that mean no-one wins.

273
00:17:51,920 --> 00:17:53,280
To explain why,

274
00:17:53,280 --> 00:17:57,000
I'd like to begin with a rather
unusual detective story.

275
00:18:00,480 --> 00:18:05,200
I'm looking at the Instagram page
of the rapper Ludacris.

276
00:18:05,200 --> 00:18:10,200
In 2015, he shared this picture with
his millions of followers.

277
00:18:10,200 --> 00:18:13,200
It had originally been posted on the
Facebook page of the police

278
00:18:13,200 --> 00:18:16,920
department in Franklin County,
Kentucky.

279
00:18:16,920 --> 00:18:19,600
And it reads...
"Attention, drug dealers.

280
00:18:20,800 --> 00:18:25,840
"We offer a free service to help you
eliminate your drug competition."

281
00:18:26,920 --> 00:18:30,360
And then under a quite large
marijuana leaf,

282
00:18:30,360 --> 00:18:33,360
there are a series of sections
where dealers

283
00:18:33,360 --> 00:18:35,800
can identify their competitors.

284
00:18:35,800 --> 00:18:40,320
You can fill out who your
competition is, where they live,

285
00:18:40,320 --> 00:18:43,680
their phone numbers,
even their hours of operation.

286
00:18:47,080 --> 00:18:49,880
It might seem like a gimmick, but
the fact is,

287
00:18:49,880 --> 00:18:53,120
Franklin County PD
were onto something.

288
00:18:54,320 --> 00:18:56,440
I think I need to find out a bit
more about this,

289
00:18:56,440 --> 00:18:57,920
so I'm going to call the sheriff.

290
00:18:57,920 --> 00:18:59,960
RINGING TONE

291
00:18:59,960 --> 00:19:01,680
Hello, Sheriff.

292
00:19:01,680 --> 00:19:04,280
Hi, Hannah. How are you?
I'm good, thank you.

293
00:19:04,280 --> 00:19:09,080
Why would a dealer call in and
tip off one of their competitors?

294
00:19:10,320 --> 00:19:12,160
Obviously, if they eliminate it,

295
00:19:12,160 --> 00:19:14,560
the more they get,
the more business they do,

296
00:19:14,560 --> 00:19:17,560
the more money, the potential
earning that they have,

297
00:19:17,560 --> 00:19:20,160
that they will make. And it helps
you get rid of your competition,

298
00:19:20,160 --> 00:19:21,600
you're the only game in town.

299
00:19:21,600 --> 00:19:23,880
Especially by having law enforcement
do it.

300
00:19:23,880 --> 00:19:26,280
Does this really work for you,
then, as a tactic,

301
00:19:26,280 --> 00:19:29,280
getting the criminals to do
the dirty on each other?

302
00:19:29,280 --> 00:19:32,280
Yes. It's something that was very
simple to do

303
00:19:32,280 --> 00:19:35,600
and, at the end of the day, if we
have got drug dealers

304
00:19:35,600 --> 00:19:38,000
turning in drug dealers,
that is a win.

305
00:19:38,000 --> 00:19:40,640
Police 1, drug dealers 0.

306
00:19:40,640 --> 00:19:43,480
Wow. This was obviously a pretty
good score

307
00:19:43,480 --> 00:19:47,360
for Sheriff Melton because what he
did was force

308
00:19:47,360 --> 00:19:51,840
the local drug dealers into the most
famous conundrum in

309
00:19:51,840 --> 00:19:54,040
the history of winning.

310
00:19:54,040 --> 00:19:57,680
It is called, appropriately enough,
the prisoner's dilemma,

311
00:19:57,680 --> 00:20:01,240
and it demonstrates very clearly the
dangers

312
00:20:01,240 --> 00:20:04,240
of falling into a lose-lose
situation

313
00:20:04,240 --> 00:20:07,200
when acting in your own
best interests.

314
00:20:07,200 --> 00:20:09,200
Think of it like this.

315
00:20:09,200 --> 00:20:13,280
Imagine Sheriff Melton has chucked
a couple of shady looking characters

316
00:20:13,280 --> 00:20:15,480
from his local beat into jail.

317
00:20:16,880 --> 00:20:20,200
He has got evidence to charge them
with possession,

318
00:20:20,200 --> 00:20:24,920
but can only get them for dealing
if they rat each other out.

319
00:20:24,920 --> 00:20:28,200
So, he puts them
in two separate cells

320
00:20:28,200 --> 00:20:31,520
and gives them both his
clever flyer.

321
00:20:33,240 --> 00:20:36,680
Our suspects must now consider
whether they should

322
00:20:36,680 --> 00:20:39,120
both keep their traps shut and only

323
00:20:39,120 --> 00:20:41,960
go away for one year each for
possession...

324
00:20:43,200 --> 00:20:45,800
..both rat the other out for dealing

325
00:20:45,800 --> 00:20:48,520
and each get
a three-year sentence...

326
00:20:50,720 --> 00:20:54,000
..or hope that they alone grass
the other up,

327
00:20:54,000 --> 00:20:57,800
getting off scot-free while their
competition gets

328
00:20:57,800 --> 00:21:02,360
a whopping five years - otherwise
known as the sucker's payoff.

329
00:21:06,360 --> 00:21:07,720
So, what to do?

330
00:21:08,880 --> 00:21:11,240
Well, this option is the best.

331
00:21:12,800 --> 00:21:17,640
But there's a risk it could quickly
turn sour if one person decides

332
00:21:17,640 --> 00:21:19,840
to basically betray the other.

333
00:21:22,400 --> 00:21:25,720
Good old Sheriff Melton knows these
two are toerags,

334
00:21:25,720 --> 00:21:27,640
and they know it too.

335
00:21:27,640 --> 00:21:30,120
So, to avoid being screwed over,

336
00:21:30,120 --> 00:21:33,200
the only stable solution for them
is this.

337
00:21:34,960 --> 00:21:37,640
They both must implicate the other.

338
00:21:37,640 --> 00:21:41,600
A win for Sheriff Melton,
but lose-lose for them.

339
00:21:43,360 --> 00:21:47,240
This unhappy scenario is called the
Nash equilibrium.

340
00:21:49,960 --> 00:21:54,200
It's a beautifully simple proof
that was published in 1950 by the

341
00:21:54,200 --> 00:21:58,160
mathematician John Nash, when he was
still in his early 20s.

342
00:22:00,200 --> 00:22:05,240
It won him a Nobel Prize, and it
transformed the analysis of winning.

343
00:22:06,640 --> 00:22:08,840
What Nash proved was that in a
situation

344
00:22:08,840 --> 00:22:11,720
with several possible outcomes,
like this one,

345
00:22:11,720 --> 00:22:15,760
where people either can't or
won't cooperate with each other,

346
00:22:15,760 --> 00:22:18,480
there is always a strategy

347
00:22:18,480 --> 00:22:23,160
which you and everyone else is best
off opting for.

348
00:22:23,160 --> 00:22:26,000
Now, in the case of
the prisoner's dilemma,

349
00:22:26,000 --> 00:22:30,360
paradoxically that strategy ends up
with lose-lose.

350
00:22:30,360 --> 00:22:34,080
Everybody wants to win by ratting
out their competition,

351
00:22:34,080 --> 00:22:37,840
and everyone wants to avoid
the sucker's payoff.

352
00:22:37,840 --> 00:22:42,400
And as a result,
everyone ends up going to jail.

353
00:22:44,400 --> 00:22:48,560
Being uncooperative can cost you
more than you bargained for.

354
00:22:50,840 --> 00:22:53,360
The reason Nash's proof matters so
much

355
00:22:53,360 --> 00:22:55,520
is because these kind of dilemmas

356
00:22:55,520 --> 00:23:00,520
constantly appear in real life
because of the way our individual

357
00:23:00,840 --> 00:23:05,360
interests often clash with those of
others and of society.

358
00:23:05,360 --> 00:23:10,360
Whether you are a drug dealer, or
just out for dinner with your mates,

359
00:23:10,720 --> 00:23:13,880
no-one wants to be the sucker,
right?

360
00:23:18,760 --> 00:23:23,680
I remember the day I was arrested,
then I was locked up for 48 hours.

361
00:23:23,760 --> 00:23:25,640
I remember thinking...

362
00:23:25,640 --> 00:23:29,360
"Oh, THIS is what happens."

363
00:23:29,360 --> 00:23:33,280
If ever there was an example of
Nash's theory in action,

364
00:23:33,280 --> 00:23:37,000
it is in the controversial world
of professional cycling.

365
00:23:38,160 --> 00:23:40,360
The global hero at the time was
Lance Armstrong.

366
00:23:40,360 --> 00:23:42,360
We were really good friends.

367
00:23:42,360 --> 00:23:45,760
I remember him saying, "I got caught
with my hand in the cookie jar".

368
00:23:45,760 --> 00:23:47,320
I knew what Lance was doing.

369
00:23:47,320 --> 00:23:50,600
It was like... Hang on a second.

370
00:24:07,360 --> 00:24:11,080
David is one of the highest profile
riders to talk candidly

371
00:24:11,080 --> 00:24:14,560
about the use of
performance-enhancing drugs
within cycling.

372
00:24:16,360 --> 00:24:19,480
Tell me about the first time that
you realised that cycling

373
00:24:19,480 --> 00:24:21,360
perhaps wasn't squeaky-clean.

374
00:24:21,360 --> 00:24:26,400
Um, my very first pro race in
February, 1997.

375
00:24:27,440 --> 00:24:30,200
My roommate was offered
a cortisone pill.

376
00:24:30,200 --> 00:24:33,680
We all carried around medical bags
and it was omnipresent.

377
00:24:33,680 --> 00:24:37,240
It was white noise, is what I've
always described it as.

378
00:24:37,240 --> 00:24:40,160
For four years, David trained and
raced clean,

379
00:24:40,160 --> 00:24:44,760
but the pressure to remain
competitive took its toll.

380
00:24:44,760 --> 00:24:48,120
Eventually, despite your initial
resistance,

381
00:24:48,120 --> 00:24:49,920
you did decide to give in.

382
00:24:49,920 --> 00:24:53,480
Yeah. I think that is the term -
I gave in.

383
00:24:53,480 --> 00:24:58,440
I didn't decide to join them, I gave
in and thought, "Well, you know,

384
00:24:58,440 --> 00:25:01,560
"I can't fight this any more".

385
00:25:01,560 --> 00:25:05,000
I was an ethical person,
but over time I just went...

386
00:25:05,000 --> 00:25:06,880
It wears you down?

387
00:25:06,880 --> 00:25:10,080
Well, just...the environment
you are in wears you down.

388
00:25:10,080 --> 00:25:12,800
You can put a good apple in a barrel
of bad apples,

389
00:25:12,800 --> 00:25:14,720
it's always going to turn bad.

390
00:25:20,040 --> 00:25:22,360
It's a classic prisoner's dilemma.

391
00:25:24,520 --> 00:25:28,040
In professional cycling,
riders compete to win.

392
00:25:29,720 --> 00:25:32,400
If doping makes winning more likely,

393
00:25:32,400 --> 00:25:35,600
then abiding by the rules risks
the sucker's payoff...

394
00:25:37,720 --> 00:25:41,520
..and so the rational thing to do
is cheat.

395
00:25:41,520 --> 00:25:46,520
Because a faster field and a doping
arms race has become the new norm.

396
00:25:48,520 --> 00:25:52,000
This dilemma unites all
competitive sport.

397
00:25:52,000 --> 00:25:55,840
Though it is a lose-lose scenario
if everyone cheats, nonetheless,

398
00:25:55,840 --> 00:25:59,840
for individuals,
cheating can give you the edge,

399
00:25:59,840 --> 00:26:02,080
IF you can get away with it.

400
00:26:02,080 --> 00:26:06,040
The fundamental reason I was there
was because I wanted to win.

401
00:26:06,040 --> 00:26:08,440
I was very ambitious. If the goals
you want to achieve

402
00:26:08,440 --> 00:26:12,040
are being achieved by the people who
are cheating, why don't I do it?

403
00:26:12,040 --> 00:26:15,240
David was the reigning world
time-trial champion

404
00:26:15,240 --> 00:26:19,560
when he was caught in 2004,
receiving a two-year ban.

405
00:26:19,560 --> 00:26:24,000
But after his punishment,
he returned to racing clean.

406
00:26:24,000 --> 00:26:28,680
David Millar comes up, hits the
line. Best time, 31:15.

407
00:26:28,680 --> 00:26:30,080
Millar is back.

408
00:26:33,040 --> 00:26:35,800
Every time I won, from then on, it
was to prove a point,

409
00:26:35,800 --> 00:26:37,560
that it was possible to win clean.

410
00:26:37,560 --> 00:26:39,960
And you were appealing because you
were clean? Because we were clean.

411
00:26:39,960 --> 00:26:44,160
It actually became a valuable asset
to be clean and win races.

412
00:26:44,160 --> 00:26:46,560
It goes back to that
whole barrel of apples,

413
00:26:46,560 --> 00:26:49,120
make sure it is a barrel of good
apples with one bad apple.

414
00:26:49,120 --> 00:26:50,920
That's easier to control.

415
00:26:50,920 --> 00:26:53,280
These guys will never have to
encounter it because we have

416
00:26:53,280 --> 00:26:55,480
an anti-doping culture
in professional cycling now.

417
00:26:57,920 --> 00:27:02,240
David's experience illustrates the
profound tension between individual

418
00:27:02,240 --> 00:27:06,280
drive to succeed
and the greater good.

419
00:27:06,280 --> 00:27:09,880
Our winning instincts can be both
our greatest asset

420
00:27:09,880 --> 00:27:11,720
and deeply destructive.

421
00:27:16,240 --> 00:27:19,240
This is something that is relevant
to all of us.

422
00:27:20,280 --> 00:27:23,760
If you think about it, how many
times have you been in a situation

423
00:27:23,760 --> 00:27:26,400
where you didn't want to upset the
balance of things

424
00:27:26,400 --> 00:27:29,040
for fear that you might end
up missing out?

425
00:27:29,040 --> 00:27:32,560
Maybe you have had a job
where you ended up staying later

426
00:27:32,560 --> 00:27:36,480
and later at work just because
everybody else was doing it,

427
00:27:36,480 --> 00:27:38,800
and you didn't want to look like
you were lazy.

428
00:27:38,800 --> 00:27:43,760
Or maybe you have failed to call
someone out for a racist or sexist

429
00:27:44,280 --> 00:27:46,800
comment just because
you didn't want to look bad,

430
00:27:46,800 --> 00:27:48,920
you didn't want to make the fuss.

431
00:27:48,920 --> 00:27:54,000
Or maybe, you have waited and waited
to tell someone that you love them,

432
00:27:54,440 --> 00:27:58,600
just because if you said it first,
and it wasn't reciprocated,

433
00:27:58,600 --> 00:28:00,920
you'd end up feeling heartbroken.

434
00:28:00,920 --> 00:28:04,880
Now, all of these situations are
real-life examples

435
00:28:04,880 --> 00:28:07,080
of Nash equilibrium.

436
00:28:08,240 --> 00:28:12,440
So, what can be done about these
traps which are, admittedly,

437
00:28:12,440 --> 00:28:13,880
very logical?

438
00:28:13,880 --> 00:28:17,520
How can we break out of them and
secure that elusive win-win?

439
00:28:22,320 --> 00:28:25,000
Well, I would like to show you one
option,

440
00:28:25,000 --> 00:28:27,600
with the help of my favourite ever
game show.

441
00:28:28,920 --> 00:28:32,240
So I have invited some people along
to a screening

442
00:28:32,240 --> 00:28:34,000
of its dramatic climax.

443
00:28:38,760 --> 00:28:42,080
You each have a golden ball
with the word "split"

444
00:28:42,080 --> 00:28:46,200
written inside. You each have a
golden ball with the word "steal".

445
00:28:46,200 --> 00:28:49,840
If you both choose the split ball,

446
00:28:49,840 --> 00:28:53,600
you split today's jackpot
of £100,150.

447
00:28:53,600 --> 00:28:56,560
If one of you splits and one of
you steals,

448
00:28:56,560 --> 00:29:01,560
whoever chooses the steal ball will
go home with today's jackpot.

449
00:29:02,120 --> 00:29:07,160
If you both choose the steal ball,
you go home with nothing.

450
00:29:07,720 --> 00:29:09,680
I am going to split this.

451
00:29:09,680 --> 00:29:13,080
I am very, very happy
to go home with 50,000.

452
00:29:13,080 --> 00:29:15,400
You're genuinely going to split
with me? If I stole off you,

453
00:29:15,400 --> 00:29:18,240
every single person there would
run over and lynch me.

454
00:29:18,240 --> 00:29:19,920
There's no way I could...

455
00:29:19,920 --> 00:29:23,600
Everyone who knew me would just be
disgusted if I stole.

456
00:29:23,600 --> 00:29:26,440
Sarah, I can look you straight
in the eye and tell you,

457
00:29:26,440 --> 00:29:29,480
I am going to split.
Promise? I swear down to you.

458
00:29:29,480 --> 00:29:30,920
I AM going to split.

459
00:29:30,920 --> 00:29:35,720
Sarah, Steve, choose either
the split or the steal ball now.

460
00:29:37,200 --> 00:29:39,240
Hold it up.

461
00:29:39,240 --> 00:29:40,600
We're going home with 50 grand each.

462
00:29:40,600 --> 00:29:42,040
I promise you that.

463
00:29:42,040 --> 00:29:43,160
Split or steal?

464
00:29:46,360 --> 00:29:48,720
SHOUTING AND GROANING

465
00:29:50,720 --> 00:29:53,240
You never know what's coming
in this game.

466
00:29:54,760 --> 00:29:57,880
Trust me, that never gets
any less painful to watch!

467
00:29:59,320 --> 00:30:02,360
I think we can all agree
that was pretty incredible.

468
00:30:02,360 --> 00:30:05,080
I'm just really shocked that she
would do such a thing.

469
00:30:05,080 --> 00:30:07,360
What would you have done in that
situation, though?

470
00:30:07,360 --> 00:30:09,560
To be honest, I would actually have
stolen as well.

471
00:30:09,560 --> 00:30:12,120
There you are, slight contradiction!

472
00:30:13,440 --> 00:30:16,440
She seemed to be reeling him in
really well

473
00:30:16,440 --> 00:30:18,840
to say, "Are you sure you're
going to split?

474
00:30:18,840 --> 00:30:21,080
"Do you promise me you are going to
split?"

475
00:30:21,080 --> 00:30:23,760
As soon as she knew that,
she could've equally thought,

476
00:30:23,760 --> 00:30:26,720
I'm going to split as well and we're
both going to leave with some money.

477
00:30:26,720 --> 00:30:28,800
Well, if you thought that was
interesting.

478
00:30:28,800 --> 00:30:31,800
I want to show you
another clip which might be

479
00:30:31,800 --> 00:30:34,440
one of my favourite things
of all time.

480
00:30:34,440 --> 00:30:36,960
That's the only introduction
I'm going to give you.

481
00:30:39,200 --> 00:30:40,720
Nick, Ibrahim.

482
00:30:42,360 --> 00:30:44,840
Ibrahim, I want you to trust me.

483
00:30:44,840 --> 00:30:48,200
100%, I'm going to pick
the steal ball.

484
00:30:48,200 --> 00:30:51,440
Sorry, you are going... I'm going
to choose the steal ball.

485
00:30:51,440 --> 00:30:52,760
I want you to do split,

486
00:30:52,760 --> 00:30:55,200
and I promise you that I will split
the money with you.

487
00:30:55,200 --> 00:30:56,840
After you've took the steal?

488
00:30:56,840 --> 00:31:00,240
Yes. You're going to take steal,
I'm going to take split.

489
00:31:00,240 --> 00:31:02,040
Yes. So you take the money...

490
00:31:02,040 --> 00:31:04,600
And I will split it with you.
After the show. Yes.

491
00:31:04,600 --> 00:31:06,360
CHUCKLING

492
00:31:06,360 --> 00:31:09,280
APPLAUSE

493
00:31:09,280 --> 00:31:11,320
Why don't we just both pick split?

494
00:31:11,320 --> 00:31:13,840
I'm not going to pick split,
I'm going to steal, Ibrahim.

495
00:31:13,840 --> 00:31:15,320
Honestly, 100% I'm going to steal.

496
00:31:15,320 --> 00:31:17,840
If you do split,
then I will split the money.

497
00:31:17,840 --> 00:31:21,280
I can't see myself doing that.
OK, well, I'm going to steal, so
we are going to leave with nothing.

498
00:31:21,280 --> 00:31:24,040
You are walking away with no money
because you're an idiot.

499
00:31:24,040 --> 00:31:26,680
That's not true. You're an idiot.
That's what you are.

500
00:31:26,680 --> 00:31:30,360
You're an idiot.
Nick, choose split or steal.

501
00:31:30,360 --> 00:31:33,240
Ibrahim, choose split or steal.
Now, please.

502
00:31:36,680 --> 00:31:38,840
Split or steal?

503
00:31:42,240 --> 00:31:44,040
Yes, congratulations!

504
00:31:44,040 --> 00:31:47,760
You have both split
and each receive £6,800.

505
00:31:47,760 --> 00:31:50,160
I would have split it.
Why did you put me through that?!

506
00:31:50,160 --> 00:31:52,640
Why would you do that to me?!

507
00:31:55,600 --> 00:31:59,400
Let's take a moment to recap
what just happened.

508
00:31:59,400 --> 00:32:02,080
By promising he will steal upfront,

509
00:32:02,080 --> 00:32:05,240
Nick removes two possible outcomes
to the game.

510
00:32:06,520 --> 00:32:11,280
This means Ibrahim can only steal
and be guaranteed nothing,

511
00:32:11,280 --> 00:32:15,920
or split, handing the cash to Nick
in the hope that he will share it.

512
00:32:17,000 --> 00:32:20,000
Nick knows Ibrahim wants the cash,

513
00:32:20,000 --> 00:32:23,360
so despite his ultimatum,
he also splits.

514
00:32:24,960 --> 00:32:26,280
It's a win-win,

515
00:32:26,280 --> 00:32:30,120
thanks to a brilliant way of forcing
cooperation.

516
00:32:34,080 --> 00:32:38,400
I am delighted that we are joined
today by Nick from that video.

517
00:32:43,240 --> 00:32:46,120
Wow! Thank you.
Nick, that was pretty amazing.

518
00:32:46,120 --> 00:32:48,920
How did it come about,
you appearing on Golden Balls?

519
00:32:48,920 --> 00:32:51,000
So, when I initially saw it,

520
00:32:51,000 --> 00:32:53,720
I thought it was a bit horrendous as
a game show

521
00:32:53,720 --> 00:32:57,600
because what happened basically
is you are rewarded for shafting

522
00:32:57,600 --> 00:33:01,920
people. And I just wanted to prove
that if you trust someone, actually,

523
00:33:01,920 --> 00:33:04,960
more often than not,
people will come through for you.

524
00:33:04,960 --> 00:33:06,520
That was the motivation behind it.

525
00:33:06,520 --> 00:33:09,320
Was there any point when you were
worried

526
00:33:09,320 --> 00:33:13,120
that he wouldn't go with you,
that he wouldn't go for split?

527
00:33:13,120 --> 00:33:18,000
If he didn't split, then he'd lost
everything,

528
00:33:18,000 --> 00:33:21,400
so you can either be stupid or
REALLY stupid -

529
00:33:21,400 --> 00:33:23,440
and when I say really stupid,

530
00:33:23,440 --> 00:33:25,880
you'd rather have nothing than
see somebody get something.

531
00:33:25,880 --> 00:33:28,320
You just had one clear strategy
you just needed to stick to.

532
00:33:28,320 --> 00:33:30,240
I tried to sort of remove the
dilemma.

533
00:33:30,240 --> 00:33:31,680
Forced him into a position?

534
00:33:31,680 --> 00:33:33,640
Yes. If you had won the money,

535
00:33:33,640 --> 00:33:36,240
would you have split it with him
outside of the game?

536
00:33:36,240 --> 00:33:38,520
Yes. I REALLY would have done.

537
00:33:38,520 --> 00:33:40,840
It was about proving that if you
trust people,

538
00:33:40,840 --> 00:33:42,760
sometimes they don't always
shaft you,

539
00:33:42,760 --> 00:33:45,200
and that was the philosophy
of why I wanted to do it.

540
00:33:45,200 --> 00:33:47,600
Were you expecting it to get
that much attention?

541
00:33:47,600 --> 00:33:49,640
No, nobody could have foreseen that.

542
00:33:49,640 --> 00:33:53,560
Harvard University teach it
in economics.

543
00:33:53,560 --> 00:33:57,360
UCL teach it... As do I!
Yes! Bath University do.

544
00:33:57,360 --> 00:33:59,680
I've been asked to go to New York
to talk about it.

545
00:33:59,680 --> 00:34:02,520
President Obama's economist has
included me in his book,

546
00:34:02,520 --> 00:34:04,600
it's just a bit insane, really.

547
00:34:04,600 --> 00:34:07,720
Yeah, wow. You did break
the prisoner's dilemma.

548
00:34:10,200 --> 00:34:13,680
It's safe to conclude that having a
strategic genius

549
00:34:13,680 --> 00:34:17,560
and all-round lovely man like Nick
take control

550
00:34:17,560 --> 00:34:19,600
can be a real win for all.

551
00:34:20,880 --> 00:34:25,200
So, like him, look for strategies
to force cooperation.

552
00:34:28,880 --> 00:34:33,080
Strong leadership, incentives
or shared morals and values

553
00:34:33,080 --> 00:34:36,960
underpin equilibria that
work in everyone's favour.

554
00:34:38,640 --> 00:34:40,720
But they are hard won

555
00:34:40,720 --> 00:34:44,040
and we don't need to look far to see
the terrible consequences

556
00:34:44,040 --> 00:34:46,520
of our failure to cooperate.

557
00:34:48,480 --> 00:34:50,840
There are many situations in life,

558
00:34:50,840 --> 00:34:55,800
egotistical behaviour is not only
morally problematic but it is also

559
00:34:56,320 --> 00:34:58,440
strategically unwise.

560
00:35:12,520 --> 00:35:16,080
I think a nice domestic example is
washing dishes.

561
00:35:16,080 --> 00:35:17,920
No-one loves it, OK.

562
00:35:17,920 --> 00:35:21,280
It is a little bit annoying,
you know.

563
00:35:21,280 --> 00:35:24,720
And we all know that it is better
for someone else to do it.

564
00:35:24,720 --> 00:35:27,240
But if you're living
in a shared house,

565
00:35:27,240 --> 00:35:32,240
the problem begins when all the
players decide to adopt my strategy

566
00:35:33,160 --> 00:35:36,560
and then NO-ONE is washing
the dishes.

567
00:35:36,560 --> 00:35:41,560
The outcome is a tower of dirty
plates, and it is not good.

568
00:35:42,480 --> 00:35:47,240
When you are taking into account
only your personal good

569
00:35:47,240 --> 00:35:50,440
and all your thinking is
self-centred,

570
00:35:50,440 --> 00:35:53,720
you don't care at all
about other players,

571
00:35:53,720 --> 00:35:57,520
the result may be a total disaster.

572
00:35:57,520 --> 00:36:01,920
It's so easy to disregard the small
impact we might make

573
00:36:01,920 --> 00:36:05,240
just by ditching stuff
for someone else to deal with.

574
00:36:06,600 --> 00:36:11,640
But many small acts of selfishness
can have huge consequences.

575
00:36:11,880 --> 00:36:16,920
It's a phenomenon known
as the tragedy of the commons.

576
00:36:17,160 --> 00:36:20,920
For example,
the damage to the atmosphere.

577
00:36:20,920 --> 00:36:25,800
The destruction of ocean ecosystems,
rainforest loggings,

578
00:36:25,800 --> 00:36:29,200
these issues are far more important.

579
00:36:29,200 --> 00:36:32,760
And it is really hard to tackle
with them,

580
00:36:32,760 --> 00:36:35,720
even with rules and regulations.

581
00:36:36,960 --> 00:36:39,520
Arguably, the challenge for
government

582
00:36:39,520 --> 00:36:44,560
is to create rules that align the
interests of individuals and society

583
00:36:44,960 --> 00:36:47,160
to find those win-wins.

584
00:36:49,240 --> 00:36:51,440
But like the rest of us,

585
00:36:51,440 --> 00:36:56,480
our leaders have their own political
and economic goals to pursue,

586
00:36:56,600 --> 00:37:01,200
meaning there will always be the
temptation for some to put their own

587
00:37:01,200 --> 00:37:03,800
short-term interests first.

588
00:37:09,960 --> 00:37:13,400
So, in the real conflicts we all
face,

589
00:37:13,400 --> 00:37:16,040
can we ever justify
being uncooperative?

590
00:37:21,240 --> 00:37:24,160
Definitely, I would argue -
on occasion.

591
00:37:25,560 --> 00:37:29,000
Because some people really don't
know what's good for them.

592
00:37:33,720 --> 00:37:38,760
I'd like to suggest we take a
potentially very unstable situation,

593
00:37:39,320 --> 00:37:44,320
taking two small argumentative
children on holiday,

594
00:37:44,760 --> 00:37:48,720
to demonstrate some strategies that
can ensure you win

595
00:37:48,720 --> 00:37:50,800
in a domestic stand-off.

596
00:37:54,320 --> 00:37:59,040
The first thing absolutely NOT to do
is to make a non-credible threat.

597
00:38:00,520 --> 00:38:03,080
Right, that's it.
Holiday's cancelled.

598
00:38:03,080 --> 00:38:05,360
I'm turning the car around,
we're going home.

599
00:38:07,960 --> 00:38:11,520
The kids quickly work out that you
won't do this,

600
00:38:11,520 --> 00:38:16,560
as it harms you just as much as
them. Everybody wants a holiday.

601
00:38:20,640 --> 00:38:24,160
So try a far more credible threat
instead.

602
00:38:25,440 --> 00:38:28,880
If you don't shut up, we're going to
be spending the first day of the

603
00:38:28,880 --> 00:38:31,200
holiday visiting an art gallery.

604
00:38:35,880 --> 00:38:39,160
As long as they know you REALLY like
art galleries

605
00:38:39,160 --> 00:38:43,600
and you know they really hate them,
this one is a much better strategy.

606
00:38:51,960 --> 00:38:56,600
Even better, try a grand gesture
called pre-commitment.

607
00:38:56,600 --> 00:39:00,240
If one of you starts arguing again,
one of you is going to go to bed

608
00:39:00,240 --> 00:39:01,680
at six o'clock all week

609
00:39:01,680 --> 00:39:04,440
and the other one can stay up
as late as they like.

610
00:39:07,560 --> 00:39:09,720
This one is quite a serious
pre-commitment

611
00:39:09,720 --> 00:39:12,680
because you are risking not
being able to pack them off to bed

612
00:39:12,680 --> 00:39:16,280
like normal, but you would just have
to hope that the prospect

613
00:39:16,280 --> 00:39:20,440
of them missing out is enough for
them to be smart and stay quiet.

614
00:39:25,040 --> 00:39:27,000
In order to get what you want,

615
00:39:27,000 --> 00:39:32,040
any adversary has to believe you
have the credibility and commitment

616
00:39:32,320 --> 00:39:34,400
to go through with your threats.

617
00:39:38,840 --> 00:39:42,720
But, in the end,
no matter how winning your tactics,

618
00:39:42,720 --> 00:39:46,520
there is one thing that is
incredibly difficult to deal with.

619
00:39:46,520 --> 00:39:50,480
KIDS: It's not fair!

620
00:40:01,600 --> 00:40:03,480
In any kind of interaction,

621
00:40:03,480 --> 00:40:07,800
our sense of what is and isn't fair
is incredibly powerful.

622
00:40:10,080 --> 00:40:14,800
And in some circumstances, it can
actually make us deliberately lose.

623
00:40:17,000 --> 00:40:20,720
To explore this apparently quite
irrational way of getting what

624
00:40:20,720 --> 00:40:23,320
we want, Haim and I are going to
play something

625
00:40:23,320 --> 00:40:27,000
called the ultimatum game,
and the rules are very simple.

626
00:40:28,840 --> 00:40:31,400
Two players, one is called the
proposer

627
00:40:31,400 --> 00:40:34,800
and the other is called the
respondent.

628
00:40:34,800 --> 00:40:38,160
The proposer gets a sum of money,
for example £100.

629
00:40:38,160 --> 00:40:40,800
Is that what you've got in there?
Yes, yes, yes.

630
00:40:40,800 --> 00:40:44,880
And the proposer has to decide how
to split the money.

631
00:40:44,880 --> 00:40:46,520
And the responder,

632
00:40:46,520 --> 00:40:51,520
the responder must decide
whether to accept or to reject.

633
00:40:51,520 --> 00:40:54,280
If the responder rejects the split,

634
00:40:54,280 --> 00:40:57,280
both players will end up
with nothing.

635
00:40:57,280 --> 00:41:00,520
And it is one short game.
Take it or leave it offer.

636
00:41:00,520 --> 00:41:03,920
OK. The fairest thing to do
obviously is to say 50-50.

637
00:41:03,920 --> 00:41:05,800
50-50, of course.

638
00:41:05,800 --> 00:41:09,040
I could get an extra £10
by saying 60-40.

639
00:41:09,040 --> 00:41:11,880
Easily, I think.
However I decide...

640
00:41:11,880 --> 00:41:14,800
I mean, I do genuinely get
some of the money.

641
00:41:14,800 --> 00:41:18,120
Yes, yes, yes, real money.
A real game. Everything is real.

642
00:41:19,360 --> 00:41:23,360
Now to make some offers to a series
of unsuspecting volunteers.

643
00:41:24,440 --> 00:41:28,720
I think I'm going to split it 70-30.

644
00:41:28,720 --> 00:41:30,440
I get the 70.

645
00:41:30,440 --> 00:41:33,720
Yes. Well, obviously it is really
unfair, but I'm going to accept it.

646
00:41:33,720 --> 00:41:36,440
You have to decide.
I'm going to accept it.

647
00:41:36,440 --> 00:41:38,000
OK. £30.

648
00:41:38,000 --> 00:41:39,680
OK. Are you going to check it?

649
00:41:39,680 --> 00:41:41,640
I'll check it, I'm a mathematician.

650
00:41:41,640 --> 00:41:43,560
Thank you. Thank you.

651
00:41:43,560 --> 00:41:47,400
So, unfair offer number one
works in my favour.

652
00:41:47,400 --> 00:41:51,680
What is your decision? OK, I'm going
to go for

653
00:41:51,680 --> 00:41:54,080
£80 to me and £20 to you.

654
00:41:54,080 --> 00:41:56,360
Yes or no,
take it or leave it offer.

655
00:41:56,360 --> 00:41:59,160
Leave it. No. But hang on,
you came here with nothing.

656
00:41:59,160 --> 00:42:01,840
I did, but I feel like
if you were like, "£50 for you,

657
00:42:01,840 --> 00:42:03,760
"£50 for me,"
I'd be fine with that.

658
00:42:03,760 --> 00:42:06,840
I would've said deal. We both came
here with nothing so we both should

659
00:42:06,840 --> 00:42:09,960
have left with something, and an
equal amount would've been better

660
00:42:09,960 --> 00:42:11,400
than an unequal amount.

661
00:42:11,400 --> 00:42:13,200
Wow. Goodness. OK.

662
00:42:13,200 --> 00:42:17,680
You see, it contradicts all the
economic predictions, yes?

663
00:42:17,680 --> 00:42:20,360
Economics says one is better than
none -

664
00:42:20,360 --> 00:42:22,600
20 is for sure better than none.

665
00:42:22,600 --> 00:42:26,600
Economics says
Sayed here should accept,

666
00:42:26,600 --> 00:42:28,720
but I am denied. Next?

667
00:42:29,840 --> 00:42:32,400
Hi. Hi. Hello.

668
00:42:32,400 --> 00:42:34,040
I quite like money.

669
00:42:34,040 --> 00:42:38,440
So £90 to me and £10 to you.

670
00:42:40,840 --> 00:42:43,480
You know what?
Why not, I'll take £10.

671
00:42:43,480 --> 00:42:45,040
You have £10.

672
00:42:45,040 --> 00:42:47,600
It is yours. Thank you very much.

673
00:42:47,600 --> 00:42:50,080
Hannah, you are now such a rich
person!

674
00:42:50,080 --> 00:42:52,640
Thank you very much.
Are you happy with your decision?

675
00:42:52,640 --> 00:42:54,880
It's pretty unfair, but...
Pretty unfair.

676
00:42:54,880 --> 00:42:58,200
I am a realist and I like £10
at least in my pocket,

677
00:42:58,200 --> 00:43:01,600
so I am at least £10 richer, so...
£10 is better than none.

678
00:43:01,600 --> 00:43:05,120
Yes.
A man after my own logical heart.

679
00:43:07,560 --> 00:43:12,560
After three offers, I've bagged £160
of a possible 240.

680
00:43:14,520 --> 00:43:16,960
So why did I not clean up?

681
00:43:16,960 --> 00:43:20,000
The 20-80 split,

682
00:43:20,000 --> 00:43:23,720
not taking the £20
seems bonkers to me.

683
00:43:23,720 --> 00:43:28,200
The 20, it is approximately the
world average of the split

684
00:43:28,200 --> 00:43:30,440
that people refuse to take.

685
00:43:30,440 --> 00:43:35,440
People seem to show an unwillingness
to accept unfairness and they are

686
00:43:35,680 --> 00:43:37,520
prepared to pay. Even at 80-20?

687
00:43:37,520 --> 00:43:40,160
They are basically being irrational,
right?

688
00:43:40,160 --> 00:43:45,120
People usually behave rationally
from their point of view,

689
00:43:45,120 --> 00:43:48,720
so it is really hard to define
what is rational, you know,

690
00:43:48,720 --> 00:43:51,040
what is not. For example, Sayed,

691
00:43:51,040 --> 00:43:55,600
yes, he paid £20, but to teach you a
lesson,

692
00:43:55,600 --> 00:44:00,600
£20, it seems to me a low price,
you see? Nothing, almost.

693
00:44:00,680 --> 00:44:03,880
Well, it was an expensive lesson
for me.

694
00:44:03,880 --> 00:44:05,560
For you, but not for him.

695
00:44:07,360 --> 00:44:11,600
It might seem like an irrational way
to get what we want, but what we're

696
00:44:11,600 --> 00:44:13,760
looking at here is the long game.

697
00:44:16,120 --> 00:44:20,720
The act of sacrificing a small win
to punish someone means

698
00:44:20,720 --> 00:44:25,480
they may well be less selfish with
us and others in the future.

699
00:44:26,840 --> 00:44:31,880
It's a perfectly rational way to
secure greater collective success.

700
00:44:42,080 --> 00:44:45,680
But most creatures on earth
AREN'T rational

701
00:44:45,680 --> 00:44:48,280
so why do we see cooperative,

702
00:44:48,280 --> 00:44:51,920
even altruistic behaviour
in the animal kingdom,

703
00:44:51,920 --> 00:44:55,400
that seems to benefit others more
than the individual?

704
00:44:57,240 --> 00:45:01,240
Surely this contradicts
the theory of evolution,

705
00:45:01,240 --> 00:45:04,160
based as it is on the struggle for
life

706
00:45:04,160 --> 00:45:06,720
and the survival of the fittest?

707
00:45:09,320 --> 00:45:11,560
When I learned about evolution
at school,

708
00:45:11,560 --> 00:45:15,160
I was under the impression that it
was all about competition.

709
00:45:15,160 --> 00:45:18,880
That every animal was necessarily
out for itself

710
00:45:18,880 --> 00:45:20,720
and the fitter you are,

711
00:45:20,720 --> 00:45:23,920
the more likely you are to
win fights for mates,

712
00:45:23,920 --> 00:45:26,360
for food and for territory.

713
00:45:26,360 --> 00:45:30,840
If you can ace THOSE conflicts, the
more likely you are to pass on your

714
00:45:30,840 --> 00:45:33,800
genes and win at the game of life.

715
00:45:35,320 --> 00:45:39,080
It has been said that nothing in
biology makes sense

716
00:45:39,080 --> 00:45:41,600
except in the light of evolution,

717
00:45:41,600 --> 00:45:44,640
and evolution has been based on
conflict.

718
00:45:44,640 --> 00:45:47,680
It is a struggle for existence.

719
00:46:04,560 --> 00:46:08,000
In animal behaviour,
there exists many examples

720
00:46:08,000 --> 00:46:10,680
where animals fight between

721
00:46:10,680 --> 00:46:15,160
the species, but where they
fight, according almost to rules,

722
00:46:15,160 --> 00:46:17,800
where they seem to restrain
themselves

723
00:46:17,800 --> 00:46:21,200
and don't escalate the conflict
too much.

724
00:46:22,200 --> 00:46:26,840
You could try to explain individuals
avoiding violence as being for the

725
00:46:26,840 --> 00:46:30,880
benefit of the group, but that's not
quite how evolution works.

726
00:46:32,160 --> 00:46:36,480
If there is a single individual
ready to escalate the conflict,

727
00:46:36,480 --> 00:46:38,880
and all others are going
to run away,

728
00:46:38,880 --> 00:46:41,640
when they see that
things become serious,

729
00:46:41,640 --> 00:46:43,920
then this individual will win
everything,

730
00:46:43,920 --> 00:46:45,400
will have many offsprings.

731
00:46:45,400 --> 00:46:49,080
The offspring will win his trait
of escalating a conflict

732
00:46:49,080 --> 00:46:52,360
and this trait
will become more and more frequent.

733
00:46:54,080 --> 00:46:59,000
In this scenario, aggressive
behaviour will always win out.

734
00:46:59,000 --> 00:47:01,520
So there has to be an alternative
explanation.

735
00:47:02,560 --> 00:47:07,600
It only began to emerge in the early
1970s, when British biologist John

736
00:47:08,160 --> 00:47:12,880
Maynard Smith was given a manuscript
to review, written by an unknown

737
00:47:12,880 --> 00:47:15,000
American called George Price.

738
00:47:16,280 --> 00:47:20,920
Now, Price wasn't a mathematician or
a biologist or a game theorist,

739
00:47:20,920 --> 00:47:23,080
but he had read von Neumann

740
00:47:23,080 --> 00:47:27,120
and he had written about
Cold War games and the uses

741
00:47:27,120 --> 00:47:32,120
of deterrence. His paper, titled
Antlers, Intraspecific Combat,

742
00:47:33,160 --> 00:47:38,120
And Altruism, suggested that things
like giant deer antlers weren't

743
00:47:42,520 --> 00:47:45,720
they were clever strategic
accessories

744
00:47:45,720 --> 00:47:48,400
that would help you limit conflict

745
00:47:48,400 --> 00:47:53,040
and avoid the destructive effects
of fighting.

746
00:47:53,040 --> 00:47:55,640
It's just like nuclear missiles,

747
00:47:55,640 --> 00:47:58,960
antlers are the kind of weapon that
you could

748
00:47:58,960 --> 00:48:01,400
parade in front of your enemy,

749
00:48:01,400 --> 00:48:04,000
hopefully without ever
having to actually use them.

750
00:48:05,880 --> 00:48:08,000
Maynard Smith was struck by this

751
00:48:08,000 --> 00:48:10,520
and, together with Price,
created a simple

752
00:48:10,520 --> 00:48:15,600
game to explore how such limited
war strategies could evolve.

753
00:48:17,080 --> 00:48:19,440
And he called it the Hawk-Dove Game.

754
00:48:20,480 --> 00:48:24,720
Picture an imaginary species of bird
that only has two inherited

755
00:48:24,720 --> 00:48:27,160
behavioural strategies.

756
00:48:27,160 --> 00:48:30,440
One is aggressive, or hawkish,

757
00:48:30,440 --> 00:48:34,200
and the other more cooperative,
or dovish.

758
00:48:34,200 --> 00:48:37,600
If two hawks clash over something
like food,

759
00:48:37,600 --> 00:48:40,200
their strategy means total war.

760
00:48:40,200 --> 00:48:45,040
They will fight, risking injury and
no guarantee of food either.

761
00:48:46,760 --> 00:48:48,920
Now, if a hawk and a dove meet,

762
00:48:48,920 --> 00:48:53,080
the dove initially stands up for
itself, but then scarpers.

763
00:48:54,800 --> 00:48:58,040
Finally, if doves meet,
they'll share.

764
00:48:59,520 --> 00:49:03,320
In each interaction,
both birds receive a score.

765
00:49:04,960 --> 00:49:09,080
And if these are added up over time,
it reveals which behaviours are most

766
00:49:09,080 --> 00:49:14,080
likely to benefit individuals, and
so survive down the generations.

767
00:49:17,040 --> 00:49:18,680
On the face of it,

768
00:49:18,680 --> 00:49:20,800
if everyone was displaying dovelike
behaviour,

769
00:49:20,800 --> 00:49:22,160
it looks like it could work,

770
00:49:22,160 --> 00:49:25,120
as well as being genuinely nice
for everyone.

771
00:49:25,120 --> 00:49:26,920
But, unfortunately,

772
00:49:26,920 --> 00:49:30,840
that's unstable because just one
meanie hawk could come along and

773
00:49:30,840 --> 00:49:33,400
immediately have the upper hand
on everybody.

774
00:49:33,400 --> 00:49:35,800
And if that hawkish behaviour really
took off,

775
00:49:35,800 --> 00:49:37,400
that too would be unstable,

776
00:49:37,400 --> 00:49:41,840
because of the constant risk and
cost of violence.

777
00:49:41,840 --> 00:49:46,840
Now, what the maths shows is that a
stable population is possible,

778
00:49:47,440 --> 00:49:52,480
but only if you have one-third hawks
and two-thirds doves.

779
00:49:53,200 --> 00:49:55,560
Now, those exact numbers might
change slightly

780
00:49:55,560 --> 00:49:57,160
if you tweak the payoff,

781
00:49:57,160 --> 00:50:02,200
but the key point here is that this
was mathematical proof that there

782
00:50:02,640 --> 00:50:07,000
was a strategic advantage to
avoiding conflict.

783
00:50:07,000 --> 00:50:09,160
Evolution, it turned out,

784
00:50:09,160 --> 00:50:12,880
is not simply a winner-takes-all
kind of game.

785
00:50:14,240 --> 00:50:19,200
And so, scientists took to this new
evolutionary form of game theory.

786
00:50:20,200 --> 00:50:24,560
The reason? To find the ultimate
long-term winning strategy

787
00:50:24,560 --> 00:50:26,440
for us all.

788
00:50:36,080 --> 00:50:39,440
It was an ingenious American
political scientist

789
00:50:39,440 --> 00:50:41,120
called Robert Axelrod

790
00:50:41,120 --> 00:50:43,880
who, around 1980,
took the next big step.

791
00:50:45,760 --> 00:50:49,560
Axelrod invited economists,
mathematicians,

792
00:50:49,560 --> 00:50:51,560
political scientists,

793
00:50:51,560 --> 00:50:54,440
psychologists and sociologists,

794
00:50:54,440 --> 00:50:58,800
all of whom had written theoretical
papers on co-operation and the

795
00:50:58,800 --> 00:51:02,880
prisoner's dilemma, to compete
in a computer tournament.

796
00:51:05,000 --> 00:51:07,560
The challenge was to design a
computer program

797
00:51:07,560 --> 00:51:09,880
that would play the best combination

798
00:51:09,880 --> 00:51:12,720
of attacking and more cooperative
behaviour.

799
00:51:15,560 --> 00:51:19,480
It was a bold move, considering the
golden age of popular computer games

800
00:51:19,480 --> 00:51:21,640
and coding was still in its infancy.

801
00:51:24,640 --> 00:51:26,760
Researchers from around the world

802
00:51:26,760 --> 00:51:29,680
mailed in their computer programs
via post.

803
00:51:29,680 --> 00:51:32,840
It was all to see which of
their electronic beasts,

804
00:51:32,840 --> 00:51:35,320
some of which were a lot more
cooperative than others,

805
00:51:35,320 --> 00:51:37,000
would emerge the winner.

806
00:51:39,440 --> 00:51:43,160
14 programs were to play 200
rounds of the prisoner's dilemma

807
00:51:43,160 --> 00:51:45,880
against themselves and one another.

808
00:51:48,320 --> 00:51:52,040
As you would expect, cooperating
with your opponent scores well.

809
00:51:55,840 --> 00:51:58,280
But you could get a much higher
payoff

810
00:51:58,280 --> 00:52:01,480
by attacking a co-operator
all of the time.

811
00:52:02,720 --> 00:52:05,840
But attack an attacker,
and your score will suffer.

812
00:52:05,840 --> 00:52:08,360
BEEPING

813
00:52:10,560 --> 00:52:12,560
And so, the tournament ran,

814
00:52:12,560 --> 00:52:15,600
with some players doing better
than others.

815
00:52:16,960 --> 00:52:20,680
One strategy, for example,
was nicknamed "the grudger".

816
00:52:20,680 --> 00:52:23,080
It would co-operate
until it was attacked,

817
00:52:23,080 --> 00:52:24,960
at which point it would get the hump

818
00:52:24,960 --> 00:52:27,840
and never co-operate again
for the rest of the game.

819
00:52:29,920 --> 00:52:32,080
After competing against
everyone else,

820
00:52:32,080 --> 00:52:34,920
the grudger crawled in
in seventh place.

821
00:52:36,360 --> 00:52:39,120
The surprise winner, with 504
points,

822
00:52:39,120 --> 00:52:41,440
was written
in just four lines of code,

823
00:52:41,440 --> 00:52:45,320
and was easily the simplest
program that had been submitted.

824
00:52:45,320 --> 00:52:47,640
It was called tit-for-tat.

825
00:52:47,640 --> 00:52:49,720
It would start off by cooperating

826
00:52:49,720 --> 00:52:52,960
and then would just copy whatever
the opponent did

827
00:52:52,960 --> 00:52:56,720
in the previous round - so if the
opponent either attacked or

828
00:52:56,720 --> 00:53:00,520
cooperated, tit-for-tat
would respond in kind.

829
00:53:03,560 --> 00:53:06,640
Tit-for-tat's behaviour over the
course of the competition

830
00:53:06,640 --> 00:53:10,080
is basically, "You scratch my back,
I'll scratch yours".

831
00:53:12,480 --> 00:53:15,720
So was this the winning strategy
everyone was looking for?

832
00:53:16,960 --> 00:53:18,360
Well, no.

833
00:53:19,480 --> 00:53:22,880
Because, unlike in real life,
these computer programs,

834
00:53:22,880 --> 00:53:26,360
including tit-for-tat,
could play endlessly,

835
00:53:26,360 --> 00:53:28,680
without ever making mistakes.

836
00:53:32,800 --> 00:53:36,400
Imagine if I decided to use
tit-for-tat as my new strategy

837
00:53:36,400 --> 00:53:41,440
for winning at life. So I would
co-operate by default - lovely -

838
00:53:41,920 --> 00:53:44,720
and only retaliate if someone
was really mean to me,

839
00:53:44,720 --> 00:53:46,560
just to teach them a lesson.

840
00:53:46,560 --> 00:53:49,360
But what if, idiotically,

841
00:53:49,360 --> 00:53:53,360
I sort of accidentally was
a bit mean to someone?

842
00:53:53,360 --> 00:53:56,200
Say I just bumped into them.

843
00:53:56,200 --> 00:53:57,880
Oh, sorry.

844
00:53:57,880 --> 00:54:00,520
If they are also playing
tit-for-tat,

845
00:54:00,520 --> 00:54:04,360
they'll have to shove me back -
formerly known as defecting.

846
00:54:06,400 --> 00:54:09,800
And then I'd have to do the same,
as would he.

847
00:54:09,800 --> 00:54:12,080
We'll defect for ever.

848
00:54:14,880 --> 00:54:16,880
Hmm. Not great, is it?

849
00:54:16,880 --> 00:54:20,880
But some scientists weren't quite
ready to give up on tit-for-tat

850
00:54:20,880 --> 00:54:24,160
just yet. Karl Sigmund and his
student Martin Nowak

851
00:54:24,160 --> 00:54:25,880
started experimenting

852
00:54:25,880 --> 00:54:30,320
with a new kind of evolutionary
tournament, but this time,

853
00:54:30,320 --> 00:54:31,960
as in the natural world,

854
00:54:31,960 --> 00:54:35,720
competing strategies could evolve
and make mistakes.

855
00:54:37,520 --> 00:54:42,400
Using computer simulations, they
watched strategies like tit-for-tat
emerge

856
00:54:42,400 --> 00:54:45,680
and compete over thousands of
generations.

857
00:54:48,120 --> 00:54:51,240
After the first 50 or 100 rounds,

858
00:54:51,240 --> 00:54:54,920
it seemed as if everyone was
a defector.

859
00:54:54,920 --> 00:54:56,680
Everyone? Not quite.

860
00:54:56,680 --> 00:55:00,920
There was one little minority
playing something like tit-for-tat,

861
00:55:00,920 --> 00:55:05,280
and this little minority
very slowly increased

862
00:55:05,280 --> 00:55:07,520
and became more and more frequent,

863
00:55:07,520 --> 00:55:11,400
and was actually defeating the
defectors.

864
00:55:11,400 --> 00:55:16,440
The key was to let natural
selection find the winning strategy.

865
00:55:16,680 --> 00:55:21,080
If you wait still longer, then you
will see that another,

866
00:55:21,080 --> 00:55:26,080
more generous form of tit-for-tat is
going to evolve and to take over.

867
00:55:26,720 --> 00:55:30,000
And here it comes,
generous tit-for-tat.

868
00:55:31,320 --> 00:55:34,800
This strategy will always co-operate
first.

869
00:55:37,240 --> 00:55:39,600
But when facing defection...

870
00:55:45,120 --> 00:55:48,960
..around one out of every three
times, it will just ignore it

871
00:55:48,960 --> 00:55:51,080
and co-operate, regardless.

872
00:55:53,840 --> 00:55:58,560
Of all the possible strategies,
and despite its forgiving ways,

873
00:55:58,560 --> 00:56:02,200
generous tit-for-tat consistently
came out on top.

874
00:56:03,840 --> 00:56:05,560
For me it is still...

875
00:56:06,840 --> 00:56:11,920
..one of the fondest remembrances
of my scientific life.

876
00:56:12,640 --> 00:56:15,800
There was no master program
behind it,

877
00:56:15,800 --> 00:56:19,320
no design for forgiveness,
so to speak.

878
00:56:19,320 --> 00:56:22,120
It came out all by itself.

879
00:56:22,120 --> 00:56:27,120
A purely mathematical simulation had
revealed that winning strategies,

880
00:56:27,600 --> 00:56:32,480
in the long run, tend to be
generous, hopeful and forgiving.

881
00:56:32,480 --> 00:56:37,320
Now, this basic moral code
just emerged.

882
00:56:37,320 --> 00:56:42,360
Perhaps you could even take this as
proof of the existence and advantage

883
00:56:42,440 --> 00:56:44,080
of goodness.

884
00:56:44,080 --> 00:56:46,320
So why, then, don't we live

885
00:56:46,320 --> 00:56:50,200
in a wonderful, generous and
forgiving utopia?

886
00:56:51,520 --> 00:56:55,400
The simulations
pointed to an answer.

887
00:56:55,400 --> 00:56:57,680
Over generations of stability,

888
00:56:57,680 --> 00:57:01,000
the main strategy becomes total
cooperation.

889
00:57:02,200 --> 00:57:07,160
But this is fragile because hawkish
defectors can rapidly take over...

890
00:57:08,760 --> 00:57:10,600
..and do their worst.

891
00:57:13,000 --> 00:57:17,040
And yet slowly, thanks to strategies
like tit-for-tat,

892
00:57:17,040 --> 00:57:20,160
cooperation emerges again and again.

893
00:57:22,640 --> 00:57:25,720
Despite its fragility,
there is hope.

894
00:57:25,720 --> 00:57:29,360
Because ultimately cooperation
can't be suppressed.

895
00:57:29,360 --> 00:57:33,560
Thanks to things like altruism,
kindness and reciprocity,

896
00:57:33,560 --> 00:57:36,640
it re-emerges time after time.

897
00:57:36,640 --> 00:57:41,640
And these aren't just codes that are
divined by priests and philosophers.

898
00:57:41,920 --> 00:57:46,360
They have pure mathematics
and evolution itself behind them.

899
00:57:58,800 --> 00:58:02,200
So where does that leave
you and me?

900
00:58:03,520 --> 00:58:06,080
Well, I wanted to try something out.

901
00:58:06,080 --> 00:58:08,120
Starting tonight.

902
00:58:08,120 --> 00:58:11,240
Now, this could work for any kind of
relationship, but let's imagine

903
00:58:11,240 --> 00:58:15,360
that it is your partner. What
I want you to do is to co-operate,

904
00:58:15,360 --> 00:58:18,800
but copy their previous move,
so if they come home one day

905
00:58:18,800 --> 00:58:20,640
with a big bunch of flowers,

906
00:58:20,640 --> 00:58:23,160
you should get to work
on your own romantic gesture.

907
00:58:23,160 --> 00:58:27,040
But if they come home one night
much drunker and later

908
00:58:27,040 --> 00:58:30,840
than they promised,
you get to do something equivalent.

909
00:58:30,840 --> 00:58:34,840
But every now and then, you should
forgive one of their slip-ups

910
00:58:34,840 --> 00:58:36,480
because nobody's perfect, and

911
00:58:36,480 --> 00:58:38,680
you are going to mess up, too,
at some point.

912
00:58:40,920 --> 00:58:42,800
So, there you have it.

913
00:58:42,800 --> 00:58:47,840
Strategies for a happier life and a
better world, all thanks to maths.

914
00:58:48,920 --> 00:58:51,160
I call that a win.


