1
00:00:13,900 --> 00:00:18,790
Welcome back to the six events of the Oxford Mathematics Public Lectures Home Edition.

2
00:00:18,790 --> 00:00:25,400
My name is Alan Goretti and I'm in charge of external relations for the Mathematical Institute as usual.

3
00:00:25,400 --> 00:00:29,890
Well, thanks to a sponsor's execs, markets execs, market leading,

4
00:00:29,890 --> 00:00:36,430
quantitative driven electronic market maker with offices in London, Singapore and New York.

5
00:00:36,430 --> 00:00:42,100
Tonight's speaker is Professor David Sumpter from the University of Uppsala in Sweden.

6
00:00:42,100 --> 00:00:49,960
David is actually closely connected to Oxford as the Old Orile Society Fellowship with us before heading to Sweden,

7
00:00:49,960 --> 00:00:57,820
used and applied mathematician with a wide range of interests and has made seminal contribution to the problem of collective behaviour,

8
00:00:57,820 --> 00:01:04,150
which cover everything from the inner working of fish schools and colonies, segregation in society,

9
00:01:04,150 --> 00:01:11,360
machine learning and artificial intelligence to the analysis of passing networks for football team.

10
00:01:11,360 --> 00:01:12,320
A few years ago,

11
00:01:12,320 --> 00:01:20,180
you actually wrote an entire book on the mathematics of football called Soccer Metrics on which give us a brilliant lecture last year.

12
00:01:20,180 --> 00:01:28,550
Since then, David has been very prolific and has written yet another book, which is a general reflection about the role of mathematics in the world.

13
00:01:28,550 --> 00:01:34,370
All summarise two 10 equation that rule our lives and society.

14
00:01:34,370 --> 00:01:39,890
Typically, for such event, we would have a big box of book for the other to sign up to the lecture.

15
00:01:39,890 --> 00:01:45,350
Obviously, we will not be able to do this tonight, however, do not despair.

16
00:01:45,350 --> 00:01:49,670
You can still obtain a signed copy of his book from Blackwood's Bookshop.

17
00:01:49,670 --> 00:01:59,360
All you need to do is email Oxford at Blackwell's the Echo Dot UK by 15 November, and they will provide you with all the information that you need.

18
00:01:59,360 --> 00:02:06,560
I'm very keen to hear which a questioner should know so that I would be able to forget all the other ones that currently quote my brain.

19
00:02:06,560 --> 00:02:10,490
So please, David, start now and tell us all about learning.

20
00:02:10,490 --> 00:02:24,830
10 Equation can improve my life. Thank you. So it's not every day that you ask a mathematician for advice about how you should live your life.

21
00:02:24,830 --> 00:02:32,030
But that's what I'm going to try and do today. I'm going to talk about 10 equations, which I think can make you happier.

22
00:02:32,030 --> 00:02:38,060
Make you a better person might even make you richer and more successful.

23
00:02:38,060 --> 00:02:41,660
Now, when a mathematician starts trying to offer this advice,

24
00:02:41,660 --> 00:02:49,620
it's worth starting by looking at some of the alternatives that are already on the market and there's lots of them.

25
00:02:49,620 --> 00:02:58,770
So this is just a small sample. There's Marie Kondo, she tells you how you should tidy up and make things better.

26
00:02:58,770 --> 00:03:03,450
I was leaning in this Jordan Peterson and his 12 rules of life.

27
00:03:03,450 --> 00:03:07,590
You can learn like Ice Einstein, you can declutter your mind.

28
00:03:07,590 --> 00:03:15,960
A lot of them involve a few swear words where you just have to stop doing these extreme things you can solve to be more happier.

29
00:03:15,960 --> 00:03:22,530
There's lots and lots of different, I think, the seven, 12 and 15 laws for confidence, and this one's my favourite down in the corner here.

30
00:03:22,530 --> 00:03:29,950
This is gangster confidence. So the 16 laws for gangster confidence, apparently.

31
00:03:29,950 --> 00:03:32,080
So what can a mathematician add to this?

32
00:03:32,080 --> 00:03:43,960
Well, one kind of clue and maths is really an abstract well, but one kind of clue can be found in the book The DA Vinci Code.

33
00:03:43,960 --> 00:03:53,780
Dan Brown. The da Vinci Code is a very different type of books than these self-help books, and it's a work of fiction, of course.

34
00:03:53,780 --> 00:04:03,100
But in the first 60 to 100 pages of this, it brings out a lovely idea that mathematics can underlie everything.

35
00:04:03,100 --> 00:04:11,830
And it describes this using the number PHI the golden ratio one point six one eight and

36
00:04:11,830 --> 00:04:17,620
the idea of PHI comes up a lot in mathematics that comes up in the golden rectangle.

37
00:04:17,620 --> 00:04:22,270
It comes up in the Fibonacci sequence and in sunflowers.

38
00:04:22,270 --> 00:04:30,490
And the idea proposed in the book, and it comes up with a lot of fake examples as well as real examples.

39
00:04:30,490 --> 00:04:38,490
But the idea in the book is that there is a secret held in mathematics and it's held in a single number.

40
00:04:38,490 --> 00:04:46,230
So the problem or the question I'm going to pose in this talk is could mathematics hold the answer?

41
00:04:46,230 --> 00:04:55,720
We are all looking for? Could we just throw away or leave all of these other self-help books and concentrate on finding a mathematical answer?

42
00:04:55,720 --> 00:05:02,560
And I'm not going to tell you every one of the 10 equations today, I'm going to give you a feeling for a few of them.

43
00:05:02,560 --> 00:05:08,320
The first one that I'm going to talk about is called the judgement equation.

44
00:05:08,320 --> 00:05:14,290
Now. I'm going to introduce it by thinking about an example, and it's quite a scary example,

45
00:05:14,290 --> 00:05:21,390
but one we're all familiar with, or we were at least familiar with before the corona times.

46
00:05:21,390 --> 00:05:29,730
Imagine you are an experienced traveller, having flown 100 times before the flight you are now on is different.

47
00:05:29,730 --> 00:05:36,150
As you descend, the plane starts to rattle and shake in the way you've never experienced before.

48
00:05:36,150 --> 00:05:40,800
The woman next to you lets out a gasp. The man sitting across the aisle grips his knees.

49
00:05:40,800 --> 00:05:45,870
Everyone around you is visibly scared. Could this be it?

50
00:05:45,870 --> 00:05:53,050
Could the worst possible scenario be about to unfurl? So here is the nightmare scenario.

51
00:05:53,050 --> 00:05:57,930
You feel that the plane you're on is possibly going to crash.

52
00:05:57,930 --> 00:06:03,000
This is maybe a overdramatic representation of what you're currently experiencing.

53
00:06:03,000 --> 00:06:10,500
We want to know the probability of a crash, given the plane is shaking, we can feel the plane shaking.

54
00:06:10,500 --> 00:06:15,120
We're worried about the crash. The shake is the data, the crashes, the model.

55
00:06:15,120 --> 00:06:24,810
OK, so let's deal with that problem. You've been on a hundred flights before in this scenario and what's happening in this particular flight?

56
00:06:24,810 --> 00:06:30,630
Is your flight shaking? So this is a 10 by 10 grid of flights.

57
00:06:30,630 --> 00:06:36,510
There's only one of them that's ever shaken as badly as this particular one that you've been on.

58
00:06:36,510 --> 00:06:43,170
And so what you know from this information is the probability of this much shaken given that

59
00:06:43,170 --> 00:06:47,220
you're not going to crash if you assume you're not going to crash is one in a hundred.

60
00:06:47,220 --> 00:06:51,960
This is the worst possible flight that you've ever been on. So it could be you're going to survive.

61
00:06:51,960 --> 00:06:58,600
And a good estimate for the probability of shaking, given you're not going to crash is one in a hundred.

62
00:06:58,600 --> 00:07:03,870
It's worth just thinking about what might happen if you went on two thousand flights like this.

63
00:07:03,870 --> 00:07:10,020
So if you have a probability of shaking this badly of one in one hundred,

64
00:07:10,020 --> 00:07:16,830
then if you go in 2000 flights, you'll have 20 of those very bad shaky rides.

65
00:07:16,830 --> 00:07:20,250
And then let's go one step further.

66
00:07:20,250 --> 00:07:29,280
The background rate of plane crashes, plane crashes are very unlikely, so you don't know exactly how many plane crashes occur,

67
00:07:29,280 --> 00:07:33,540
but the chance of being in a plane crash is something like one in 10 million.

68
00:07:33,540 --> 00:07:40,290
You actually have time to look this up on your phone and in your current situation, I can tell you it's around one in 10 million.

69
00:07:40,290 --> 00:07:49,050
It might be a bit less than that. It might be a bit more than that. So now start thinking about all of those possible plane journeys you can be on.

70
00:07:49,050 --> 00:07:54,820
What I've done here is I've made a thousand times, ten times 10000 thousand array of journey,

71
00:07:54,820 --> 00:08:03,360
so imagine a thousand times 10000 journeys and that is 10 million journeys.

72
00:08:03,360 --> 00:08:11,410
An inside all of those journeys, there's always going to be these shaky journeys and you happen to be on one of those shaky journeys right now.

73
00:08:11,410 --> 00:08:18,790
But the probability you're going to crash is only one in 10 million before you got onto the plane if you weren't shaking.

74
00:08:18,790 --> 00:08:25,660
So here's the crashing plane somewhere hidden somewhere in between all of those 10 million journeys.

75
00:08:25,660 --> 00:08:31,480
So that's horrible for the people on that flight, that train, that that plane.

76
00:08:31,480 --> 00:08:39,640
So now you actually have all the information you need in order to work out your probability of being in a plane crash.

77
00:08:39,640 --> 00:08:51,580
In fact, there's going to be 100000 shaky journeys in this in these 10 million journeys, and only one of them is going to end in the crash journey.

78
00:08:51,580 --> 00:08:58,600
That's essentially 10 million divided by 100 gives you a hundred thousand.

79
00:08:58,600 --> 00:09:07,880
And that means roughly. The probability your shaky journey is the crash journey is approximately one in 100000.

80
00:09:07,880 --> 00:09:14,720
You really aren't going to die. It's the worst thing you've experienced, but you don't actually need to be worried given the circumstances.

81
00:09:14,720 --> 00:09:21,890
Now what? What's the judgement equation? I've gone through this problem with which I've gone through it from a frequent his point of view.

82
00:09:21,890 --> 00:09:29,810
I've looked at the frequency of these different things occurring, but we can actually use an equation which I call the judgement equation.

83
00:09:29,810 --> 00:09:39,290
But it originally was found by Thomas Bayes. It's called Bayes rule came from the end or mid end of 18th century.

84
00:09:39,290 --> 00:09:46,370
And what it does is it gathers up all the information we have, so you notice that while I was doing my calculation,

85
00:09:46,370 --> 00:09:52,910
I find the probability of crashing the background crash rate for planes is one in 10 million.

86
00:09:52,910 --> 00:09:56,780
If your plane is going to crash, it will shake.

87
00:09:56,780 --> 00:10:02,360
So the probability of it shaking given a crash is one.

88
00:10:02,360 --> 00:10:06,890
And probability of Shay Given not crash, that's based on your experience.

89
00:10:06,890 --> 00:10:11,330
We estimate that to be around one in 100. It's the worst flight you've been on.

90
00:10:11,330 --> 00:10:15,980
But what we really want to know is the probability of crash given shake,

91
00:10:15,980 --> 00:10:21,860
and that is given by base equation, which looks like follows as follows the here.

92
00:10:21,860 --> 00:10:29,750
We just put in all of the numbers and we work out, we come up with this thing being approximately equal to one in 100000 now.

93
00:10:29,750 --> 00:10:38,120
Bayes rule and the judgement equation have come up a lot recently because they also come up with regard to the coronavirus.

94
00:10:38,120 --> 00:10:46,940
One of the problems that's come up during the COVID crisis is there are false positives in tests for the disease.

95
00:10:46,940 --> 00:10:58,850
So if you go and get yourself tested for the virus sometimes and the rate is about one in a thousand, you'll come up with a false positive example.

96
00:10:58,850 --> 00:11:04,940
Even though you haven't got the disease that you'll get a test result that comes back, that says you have got the disease.

97
00:11:04,940 --> 00:11:10,460
And we can also think about this using the judgement equation. So here we've got 100 people.

98
00:11:10,460 --> 00:11:16,430
I've just said that it's more like a one in a thousand, but imagine it was a one in 100 error rate for the disease.

99
00:11:16,430 --> 00:11:18,260
So we've got one in 100 error rate.

100
00:11:18,260 --> 00:11:27,050
Well, one in 100 people are going to get back a positive test even though they're healthy and don't have the disease.

101
00:11:27,050 --> 00:11:35,420
So that's one in a hundred there. But imagine also that we have one in 100 people who have got the disease.

102
00:11:35,420 --> 00:11:42,560
So here's 100 people. And one of them. Has got the disease.

103
00:11:42,560 --> 00:11:54,140
That's this one here. Now, what we can see here quite clearly, actually, is that in this case, it's just 50 50 if you get a test result.

104
00:11:54,140 --> 00:11:58,320
So if you get a test result and it tells you you could be that person who has

105
00:11:58,320 --> 00:12:05,700
the disease or you could just be this person who got a wrong test result back,

106
00:12:05,700 --> 00:12:11,120
in fact, there's a 50 50 chance that you're one of these two people.

107
00:12:11,120 --> 00:12:17,260
Now, the one in 100 the I've I've put up there.

108
00:12:17,260 --> 00:12:24,790
It really depends on a lot of different factors. Now one of them, for example, is you only go and get tested if you've got symptoms.

109
00:12:24,790 --> 00:12:30,080
If you look at these, these little guys here, all of the ones with face masks on, they don't go and get tested.

110
00:12:30,080 --> 00:12:36,850
But there's a few individuals here who have got got symptoms.

111
00:12:36,850 --> 00:12:44,200
So if only one in seven individuals have, even if only seven out of 100 individuals have symptoms,

112
00:12:44,200 --> 00:12:49,220
then it's much more likely that a test result is correct.

113
00:12:49,220 --> 00:12:55,170
There's all sorts of parameters that you have to think about when you're thinking about the coronavirus and the testing of that.

114
00:12:55,170 --> 00:13:02,600
And so I don't want to give I don't want to give a conclusive discussion of that just now that's raging over Twitter and in the news,

115
00:13:02,600 --> 00:13:10,790
these types of discussions. But it's the same mathematical problem based on the judgement equation, which underlies both of those things.

116
00:13:10,790 --> 00:13:18,140
There's a general way which I think mathematicians think about the world, which is very useful or applied.

117
00:13:18,140 --> 00:13:27,400
Mathematicians think about the world, which is very useful. And I like to think about how we break the world down into model and data.

118
00:13:27,400 --> 00:13:35,350
I often see I personally see the future or I have a sort of model in my head, which is a little bit like a film.

119
00:13:35,350 --> 00:13:37,330
I see different things happening.

120
00:13:37,330 --> 00:13:45,640
So when I'm in an aeroplane and it starts to shake and I start to get worried, I start to imagine the bad things that could happen to me.

121
00:13:45,640 --> 00:13:57,100
And that's a model of the future of the world. The data of the world is the shaking experience that something bad is happening, and all of the time,

122
00:13:57,100 --> 00:14:04,330
what we how we should be thinking is updating our model on the basis of the data we get in, and that applies to everything you can think about.

123
00:14:04,330 --> 00:14:12,820
If you're swimming in the sea where you know there might be a shark attack and you see something moving, you need to your model is the shark attack.

124
00:14:12,820 --> 00:14:18,580
Your data is the movement in the water. It's the same when you get an interview for a new job.

125
00:14:18,580 --> 00:14:22,570
Your model is that you'll be successful and you'll get the new job.

126
00:14:22,570 --> 00:14:27,370
The data is the outcome of the interview and how the interview went and the same.

127
00:14:27,370 --> 00:14:32,890
When you get a medical test, you've got a model that you'll be infected with the disease and you get the test and

128
00:14:32,890 --> 00:14:38,380
that improves your understanding of if you will actually have the disease or not.

129
00:14:38,380 --> 00:14:45,970
And all of these things can be solved by putting the collecting up the data that you have and putting them into Bayes theorem,

130
00:14:45,970 --> 00:14:54,620
putting them into the judgement equation. And. The judgement equation isn't just about.

131
00:14:54,620 --> 00:14:59,930
Objective things that happen in time in our lives, it's also about our feelings.

132
00:14:59,930 --> 00:15:09,170
You can also use the judgement equation in order to update and have a better understanding of the personal things that happen in your life.

133
00:15:09,170 --> 00:15:18,050
So I'm going to give an example of this. So you probably can't help but notice the change of scene for this particular segment of the talk.

134
00:15:18,050 --> 00:15:23,270
And that's because when Darrell, who organises the talk talks at the institute, and Alan,

135
00:15:23,270 --> 00:15:29,810
who presented me and myself when we reviewed the original lecture that I did in a kind of cold,

136
00:15:29,810 --> 00:15:38,030
empty lecture theatre, we felt that the setting wasn't quite right for the message that I'm going to try and bring out now.

137
00:15:38,030 --> 00:15:43,190
And I think it works very well when I'm presenting things to do with aeroplanes and coronavirus,

138
00:15:43,190 --> 00:15:50,900
when we're thinking about factual situations, when we're handling hard data and trying to understand that.

139
00:15:50,900 --> 00:15:56,540
But what I'm going to talk about now isn't so much objective. Data is much more subjective data.

140
00:15:56,540 --> 00:16:00,710
It's things about feelings and it's things about emotions.

141
00:16:00,710 --> 00:16:09,530
Now, I think all of us have a lot of feelings and emotions generated by our interactions with the people around us.

142
00:16:09,530 --> 00:16:18,170
And I think that the judgement equation based rule actually suggests that we should be less judgemental and more forgiving of other people.

143
00:16:18,170 --> 00:16:25,400
And I'm going to illustrate that through an example. So if I think about my own life, I mean, I have lots of frustrations.

144
00:16:25,400 --> 00:16:30,770
For example, I'm teaching a course just now with students emailing me about different things to do with the lectures,

145
00:16:30,770 --> 00:16:37,040
and I just feel I've answered these questions 100 times. And of course, I know it's different students, but I get frustrated.

146
00:16:37,040 --> 00:16:46,490
And then you get frustrated. I also work in a football club giving data analytics and doing analysis, and I get frustrated by the demands on me.

147
00:16:46,490 --> 00:16:54,380
And the idea is that there's a lot there's just a lot of passion inside football, for example, and it's all about winning.

148
00:16:54,380 --> 00:17:00,530
And things do get very heated and people do get angry with each other.

149
00:17:00,530 --> 00:17:10,250
And also, if I think about in my personal life, I mean, you know, everybody I have a family wife and two kids, a teenage kids.

150
00:17:10,250 --> 00:17:11,030
And there's, you know,

151
00:17:11,030 --> 00:17:21,950
there's a lot of debate and discussions that go on and how can we actually use mathematics to be better at dealing with those types of situations?

152
00:17:21,950 --> 00:17:27,270
Now, I think when I started studying mathematics, I had a very kind of hard idea about it.

153
00:17:27,270 --> 00:17:33,680
I saw it as like a very precise logic. And if you go back to the idea of the da Vinci Code, I talked about to the star.

154
00:17:33,680 --> 00:17:39,800
It's this idea that there's a code, a secret which gives you the correct right answer.

155
00:17:39,800 --> 00:17:49,070
But it's not really like this because this is a list of some of the bad things that can happen to you on an everyday basis.

156
00:17:49,070 --> 00:17:55,400
I made the list here of when friends or apparent friends let you down.

157
00:17:55,400 --> 00:18:02,210
And this is something which all of us have experience, sometimes we overhear a friend say something nasty about us.

158
00:18:02,210 --> 00:18:08,510
Sometimes your friend steals an idea or a joke that you've made up and they Petraeus is your own.

159
00:18:08,510 --> 00:18:16,370
Sometimes your friends have parties and they don't invite you. And these things make us all sad and they make us all well.

160
00:18:16,370 --> 00:18:22,250
They nobody can have these types of things without feeling a little bit upset.

161
00:18:22,250 --> 00:18:24,560
And the question is, Well, what should you think?

162
00:18:24,560 --> 00:18:30,950
And one thing that you can think in these types of situations is that the people who do these things to us,

163
00:18:30,950 --> 00:18:35,390
the apparent friends, well, actually, they're idiots.

164
00:18:35,390 --> 00:18:43,070
And you question you have to think about is, is this friend who's let you down in this way, an idiot?

165
00:18:43,070 --> 00:18:46,850
Well, I do want to say that I don't go around.

166
00:18:46,850 --> 00:18:49,640
I don't really go around thinking that people are idiots.

167
00:18:49,640 --> 00:18:55,580
You know, if you meet me for the first time, it's not that I have this percentage probability that you're an idiot,

168
00:18:55,580 --> 00:19:01,730
but it's a useful way of thinking about the problem. And the numbers here aren't that important.

169
00:19:01,730 --> 00:19:10,070
So I've said here that I have a sort of baseline rate of people that I meet are likely to be idiots one in 20.

170
00:19:10,070 --> 00:19:14,300
So five percent of people are just unhelpful idiots.

171
00:19:14,300 --> 00:19:19,820
And then 95 percent of people are nice and easy to get along with.

172
00:19:19,820 --> 00:19:28,430
Now again, these numbers are subjective. I don't really think people are actually, maybe I do think some people are idiots.

173
00:19:28,430 --> 00:19:33,890
Anyway, the point is that we're going to work through this problem and we're going to find out the

174
00:19:33,890 --> 00:19:41,540
probability that somebody who said something not very nice to you is actually a total idiot.

175
00:19:41,540 --> 00:19:46,160
And this is a a representation of this problem.

176
00:19:46,160 --> 00:19:54,140
We've got PM, which is the probability of our model, and we've got PMC, which has a probability of the complement of our model.

177
00:19:54,140 --> 00:20:03,080
And there's a five percent chance this little grey box here is the five percent chance that the person is an idiot and the 90 percent,

178
00:20:03,080 --> 00:20:09,920
95 percent chance here that they're not in India. That's before they do anything bad to you at all.

179
00:20:09,920 --> 00:20:14,720
So now comes the data now comes with we've set up the model.

180
00:20:14,720 --> 00:20:19,490
Now comes the data. The data is the bad thing that this person has done to you.

181
00:20:19,490 --> 00:20:25,580
They've let you down by not inviting you to a party or by saying something bad behind your back.

182
00:20:25,580 --> 00:20:30,290
What the data tells us. Is that?

183
00:20:30,290 --> 00:20:33,590
We can what we can say is that, well.

184
00:20:33,590 --> 00:20:42,410
Idiots spend about 50 percent of their time doing nasty things, talking about people behind their backs, letting other people down.

185
00:20:42,410 --> 00:20:48,560
But what we have to remember is that even known idiots can have bad days, in fact, 10 percent of the time.

186
00:20:48,560 --> 00:20:56,450
I've estimated that perfectly nice people do something rather regrettable and not very nice.

187
00:20:56,450 --> 00:21:04,700
So these idiot type people in the well, they spend 50 percent of their time making nasty comments and non idiot spend 10 per cent of the time.

188
00:21:04,700 --> 00:21:10,160
And now we can break this down. Also in this square, so this this has area 100 percent.

189
00:21:10,160 --> 00:21:12,350
And these are the ideas we've divided.

190
00:21:12,350 --> 00:21:21,470
Those five percent of people are idiots into 50 percent of the time saying nasty things and 50 percent of the time just being generally OK.

191
00:21:21,470 --> 00:21:27,890
Now, for the non-media as well, we can also divide them and we've got 90 percent of the time.

192
00:21:27,890 --> 00:21:32,660
They're being nice and just doing nice things, being really great friends.

193
00:21:32,660 --> 00:21:38,720
But 10 percent of the time, they're also making mistakes. They're not idiots, but they've made a mistake.

194
00:21:38,720 --> 00:21:47,720
And when something nasty happens to us, when we've collected the data, we know we're in either this situation or this situation.

195
00:21:47,720 --> 00:21:55,640
So the other situations disappear because something bad has happened to us and we just left with these two situations.

196
00:21:55,640 --> 00:22:02,330
Now, what's interesting? Is that if you look at the relative size of that box?

197
00:22:02,330 --> 00:22:09,370
So if I go all the way back here, this was the idea of being nasty. This was the nice person making a mistake.

198
00:22:09,370 --> 00:22:11,920
We've just got these two boxes left.

199
00:22:11,920 --> 00:22:19,570
Those are the two possibilities of where we lie after we've had this nasty comment or we've not been invited to the party.

200
00:22:19,570 --> 00:22:23,830
We put them together, and I want you to look at the relative sizes of these.

201
00:22:23,830 --> 00:22:27,280
So the idiot being nasty is just this size here.

202
00:22:27,280 --> 00:22:36,910
The not idiot making a mistake is a much larger box, and this is the key to what Bayes rule tells us, Bayes rule says.

203
00:22:36,910 --> 00:22:42,220
What are the comparative areas of these two particular boxes?

204
00:22:42,220 --> 00:22:50,560
And we can see that this one is about actually one fourth of the size of this one here.

205
00:22:50,560 --> 00:22:54,250
And when we put the numbers that I told you into Bayes equation,

206
00:22:54,250 --> 00:23:02,170
the nought point five times the Note .05 is calculating the area of this box, which is the idea of being nasty.

207
00:23:02,170 --> 00:23:06,070
And that's also put down here because we add up the two areas together and the

208
00:23:06,070 --> 00:23:12,490
nought point one times nought point nine five is adding up the area of this box.

209
00:23:12,490 --> 00:23:23,590
And to find out the probability that your friend is an idiot, we need to find out the area of this divided by the area of this, plus the area of this.

210
00:23:23,590 --> 00:23:27,670
That's what Bayes base rule tells us. That's for the judgement equation tells us.

211
00:23:27,670 --> 00:23:32,950
And the conclusion is that whatever this person did to you,

212
00:23:32,950 --> 00:23:40,420
the probability that they are an idiot and you should stop talking to them immediately is that only 20 percent?

213
00:23:40,420 --> 00:23:45,490
And this tells you that you should be a much more forgiving person.

214
00:23:45,490 --> 00:23:51,400
Most of the bad things that happen to you are just because nice people are making mistakes.

215
00:23:51,400 --> 00:23:55,450
Now, the numbers I've given in this example are not really the important thing.

216
00:23:55,450 --> 00:24:02,020
The important thing is thinking about the alternative hypotheses in all of these different situations, and I think we don't have it.

217
00:24:02,020 --> 00:24:08,770
We have a tendency not to do that immediately. We think we take problems personally.

218
00:24:08,770 --> 00:24:12,820
We immediately start to get this idea that people are idiots.

219
00:24:12,820 --> 00:24:21,190
But when we just calm down for a minute, we think through the alternative that actually even nice people make mistakes now and again,

220
00:24:21,190 --> 00:24:32,380
we come to a very, very different conclusion. A key in this example is that it started with me allowing myself to have a wrong model of the world.

221
00:24:32,380 --> 00:24:39,400
I allowed myself to have some emotional reaction where when my friend let me down, I concluded that he was an idiot.

222
00:24:39,400 --> 00:24:45,370
But then after that initial response, this initial model that I had created,

223
00:24:45,370 --> 00:24:52,600
I worked through the judgement equation and was able to see that there was another explanation behind all of this.

224
00:24:52,600 --> 00:25:00,400
So what's important to emphasise is this imagining wrong models or feeling frustrations and being

225
00:25:00,400 --> 00:25:08,740
disappointed with other people is actually a good part is actually a key part of making a good judgement.

226
00:25:08,740 --> 00:25:11,860
OK, I think I've got the point of that one across now.

227
00:25:11,860 --> 00:25:20,220
I think we'll go back to the lecture theatre and listen to what more I've got to say in this case.

228
00:25:20,220 --> 00:25:24,030
One thing I worry about personally is screen time with my kids.

229
00:25:24,030 --> 00:25:27,840
My kids use their mobile for their teenagers 15 17.

230
00:25:27,840 --> 00:25:35,610
They use their mobile phones a lot and to be honest, their parents also use their mobile phones and their screens a lot.

231
00:25:35,610 --> 00:25:39,840
But what we are interested in knowing is, what's the effect of that?

232
00:25:39,840 --> 00:25:53,270
Is there a big negative effect of that? And this article I came across, which was from Berkeley, is published in a Berkeley website.

233
00:25:53,270 --> 00:26:04,430
Outlined and it gave this graph about how mental wellbeing of of teenagers decreases with the amount of time they spend looking at their screens.

234
00:26:04,430 --> 00:26:11,050
Now this looks quite worrying because you've got a negative trend here, but what I think you need to remember?

235
00:26:11,050 --> 00:26:19,690
As well, what's the effect size? So what's quite unclear when you first look at the picture is that this is a scale of zero to 100.

236
00:26:19,690 --> 00:26:26,920
They've asked people zero to 100, how do you feel? And the scale is only 40 to 50 here.

237
00:26:26,920 --> 00:26:32,740
In fact, from the average answer to the most extreme answer.

238
00:26:32,740 --> 00:26:35,170
There's just a one percent effect size.

239
00:26:35,170 --> 00:26:46,400
So you become one percent less happy because you've because you've you're using your screen and this or because of teenagers using the screen.

240
00:26:46,400 --> 00:26:52,560
And you can think of this in terms of the. Of the judgement equation.

241
00:26:52,560 --> 00:27:01,110
In fact, also in this same in this same study, they found a three percent effect size for if you got a good night's sleep,

242
00:27:01,110 --> 00:27:06,120
if teenagers got a good night's sleep of teenagers, eating breakfast had a three percent effect size.

243
00:27:06,120 --> 00:27:14,780
So not eating breakfast reduces your happiness by three percent. Not having a good night's sleep reduces your happiness by three percent.

244
00:27:14,780 --> 00:27:19,700
And that in the context of the phone usage shows that it's quite a small effect.

245
00:27:19,700 --> 00:27:24,380
So when I talk to my teenagers about using their phone too much,

246
00:27:24,380 --> 00:27:32,840
I also talk to them or I spend three times more time talking to them about having a good night's sleep and eating breakfast than I do,

247
00:27:32,840 --> 00:27:37,830
nagging them about how much they use their mobile phones.

248
00:27:37,830 --> 00:27:45,870
It's a very nice article by Candice Ortigas in Nature, where she balances all sides of the debate, and that's how,

249
00:27:45,870 --> 00:27:53,220
you know, when you're reading something that's proper science, when you actually hear all sides of the debate.

250
00:27:53,220 --> 00:27:59,680
And she comes to the conclusion that smartphones are bad for some teams, but not for all of them.

251
00:27:59,680 --> 00:28:05,080
So that's the judgement equation, and I've gone through some examples of how it can be used,

252
00:28:05,080 --> 00:28:10,000
it can be used for thinking about disasters, it can be used for thinking about diseases.

253
00:28:10,000 --> 00:28:13,120
It can be used for thinking about your personal life.

254
00:28:13,120 --> 00:28:19,720
And it can also be thinking used for thinking about the things that you read online and evaluating evidence for them.

255
00:28:19,720 --> 00:28:26,790
One of the nine equations now what the other nine? Well.

256
00:28:26,790 --> 00:28:31,980
These are the names of them. I've got the betting equation, the advertising equation, the confidence equation,

257
00:28:31,980 --> 00:28:38,550
learning equation, influencer equation, skill equation, market equation and the reward equation.

258
00:28:38,550 --> 00:28:43,710
Since we're at the mathematics lecture, I'll tell you the actual equations themselves.

259
00:28:43,710 --> 00:28:53,120
These are what they look like. And my argument is that if you have these equations in your life,

260
00:28:53,120 --> 00:28:59,630
you will both be able to understand the situations around you better on a more personal, on an everyday level.

261
00:28:59,630 --> 00:29:07,610
And also, if you learn the mathematics behind these equations, they lie behind a lot of the financial industry,

262
00:29:07,610 --> 00:29:16,350
the social media industry, the betting industry and lots of other work inside public service as well.

263
00:29:16,350 --> 00:29:20,730
So I'm going to give one more example, I think I'm yeah, I'm going to give one more example of of this.

264
00:29:20,730 --> 00:29:26,080
I'm going to take out the reward equation and tell you a little bit about that.

265
00:29:26,080 --> 00:29:30,280
The reward equation is widely used by social media companies,

266
00:29:30,280 --> 00:29:37,960
and it's used to decide what types of rewards and information it should give us if it finds that we like something a lot.

267
00:29:37,960 --> 00:29:41,920
It gives us more of that stuff so that we keep clicking on it.

268
00:29:41,920 --> 00:29:48,340
My idea is actually, you can reverse engineer the reward equation both to monitor what's going on in your

269
00:29:48,340 --> 00:29:55,750
life and also and also to decide what you should do and what you shouldn't do.

270
00:29:55,750 --> 00:30:05,450
So here's my scenario. Imagine you lying on the couch, rewarding yourself by bingeing on TV and you start watching a Netflix series,

271
00:30:05,450 --> 00:30:12,710
the first episode is brilliant, as they always are. The second is OK and the third is slightly better again.

272
00:30:12,710 --> 00:30:16,730
The question is how long should you keep watching before you give up?

273
00:30:16,730 --> 00:30:20,490
I think this is a problem that's familiar to all of us.

274
00:30:20,490 --> 00:30:26,660
We're stuck with so many TV series out there always feels you could be watching something else that might be better.

275
00:30:26,660 --> 00:30:32,270
But how long should you keep watching the series before you decide now? I'm just going to give up.

276
00:30:32,270 --> 00:30:39,960
So here here's how it works. First, you should rank the first episode you watch out of 10.

277
00:30:39,960 --> 00:30:47,600
So, yeah, give it a score zero to 10. This is your tracking variable.

278
00:30:47,600 --> 00:30:52,550
Keep this in your head. And what I'm going to propose is for all the activities you do.

279
00:30:52,550 --> 00:30:56,720
Maybe it's going to the gym, meeting friends, everything you think you're doing,

280
00:30:56,720 --> 00:31:01,940
you can have this track invariable one number, which describes how much you enjoy this, this activity.

281
00:31:01,940 --> 00:31:09,870
But just now we're watching Netflix. And the rule is following, if you're tracking variable is seven or below, then stop watching.

282
00:31:09,870 --> 00:31:13,670
So this is Netflix, you. There's lots of series available.

283
00:31:13,670 --> 00:31:22,120
I'm going to be quite strict here. If it's below a seven, then stop. If the first episode is below a seven, then you stop.

284
00:31:22,120 --> 00:31:31,150
Then watch the next episode and you give it a score out of 10 at the school for this episode to the quality score and divide by two.

285
00:31:31,150 --> 00:31:37,510
And then you round this number up to get your new tracking variable Kutty +1.

286
00:31:37,510 --> 00:31:45,280
Then you go back to step two. OK, you start. Here's how it works in practise and you start with the first episode.

287
00:31:45,280 --> 00:31:48,880
You watch it. You give it nine out of 10. Kutty is equal to nine.

288
00:31:48,880 --> 00:31:55,300
This is one of my favourite favourite of series Big Little Lies and enjoy the first episode.

289
00:31:55,300 --> 00:32:00,080
Nine out of 10. The second episode isn't quite as good.

290
00:32:00,080 --> 00:32:08,770
So I give it a six out of 10. Now, what I do to update my track, invariably is I take the original nine, I add the six to it.

291
00:32:08,770 --> 00:32:12,970
So I've got 15 and I divide by two and I get seven point five.

292
00:32:12,970 --> 00:32:16,090
And remember, I said I round up, so I round up to eight.

293
00:32:16,090 --> 00:32:22,060
I think it's good to round up because then you just need to keep one hole number in your in your head.

294
00:32:22,060 --> 00:32:29,870
So it's still an eight. So even if it got six, it survives at six and goes on to the next next round, you're going to watch the next one.

295
00:32:29,870 --> 00:32:34,600
And that's lucky because the next one you watch is a nine out of 10.

296
00:32:34,600 --> 00:32:42,050
So you're tracking variable is updated. You had eight from your previous time, plus nine divided by two is an eight point five.

297
00:32:42,050 --> 00:32:51,190
Now round it up to nine. So the main advantage of this method is a good show will survive a six whenever you have a six out of 10.

298
00:32:51,190 --> 00:32:55,760
It pops back up to the eight and you keep watching the show.

299
00:32:55,760 --> 00:33:01,750
Of course, you can change the parameters in this model. You can make things decay more slowly.

300
00:33:01,750 --> 00:33:10,810
You can change this divided by two, or you can change your threshold, but it's a good model for keeping track of things.

301
00:33:10,810 --> 00:33:15,220
But it won't survive a five, if you have a nine, then you have nine plus six.

302
00:33:15,220 --> 00:33:21,730
Divided by two is seven point five. But if you have an eight series plus a five, you're just not going to survive it.

303
00:33:21,730 --> 00:33:26,200
This was another series which I watched, which wasn't quite so good.

304
00:33:26,200 --> 00:33:31,740
Very quickly, even though it started very well, I stopped watching it quite quickly.

305
00:33:31,740 --> 00:33:37,170
Now, the general form of the reward equation is as follows. You have Kutty +1.

306
00:33:37,170 --> 00:33:46,590
This is your tracking variable in the next time stat is equal to one minus alpha, we set alpha equal to two in the original.

307
00:33:46,590 --> 00:33:53,700
Sorry, we set out for equal to half in the original set up, plus a half times the reward that we got.

308
00:33:53,700 --> 00:33:56,970
The reward is that is the score we have for the show.

309
00:33:56,970 --> 00:34:05,610
The cute is our tracking variable and what's very nice in this general form of the reward equation, which has been around since the 1950s.

310
00:34:05,610 --> 00:34:10,350
We know that the tracking variable is going to converge to the average reward.

311
00:34:10,350 --> 00:34:13,500
So this is a relative series watching.

312
00:34:13,500 --> 00:34:25,450
We can actually see that whatever we start off with our tracking variable at, it will eventually converge to the real world that we have.

313
00:34:25,450 --> 00:34:30,520
And this is something that we can use to monitor all parts of our life online.

314
00:34:30,520 --> 00:34:37,330
We can give each thing that we use on social media, not just Netflix series a score.

315
00:34:37,330 --> 00:34:40,420
And we can keep track of those scores in order to see,

316
00:34:40,420 --> 00:34:47,280
is this something that I want to continue doing or is this something that I should get rid of and get out of my life?

317
00:34:47,280 --> 00:34:53,090
There is a problem here, and it's that these scores actually keep changing over time.

318
00:34:53,090 --> 00:35:00,770
How do we keep track of the changes in these scores when we don't know the underlying value of something?

319
00:35:00,770 --> 00:35:07,320
And here actually there's an answer we can look to the answer. So it's.

320
00:35:07,320 --> 00:35:14,530
Use firemen trail. And what The Ferryman trial does is whenever the ants find something good, they find some food,

321
00:35:14,530 --> 00:35:19,930
they leave FireMon Trail going to that food, and it recruits other aunts to that foods.

322
00:35:19,930 --> 00:35:27,180
So the ants pheromone trail is like a tracking variable, which follows after where the food is.

323
00:35:27,180 --> 00:35:32,810
And you can do experiments with ants. Where you can offer them two alternatives.

324
00:35:32,810 --> 00:35:38,530
This is where they've been offered some red food and some blue food.

325
00:35:38,530 --> 00:35:41,590
And there's a y shape junction here.

326
00:35:41,590 --> 00:35:50,480
This is a close up of the junction where the ants have to choose if they will go to the red food or to the blue food.

327
00:35:50,480 --> 00:35:50,810
Now,

328
00:35:50,810 --> 00:35:58,520
how do the ants manage to get the balance right because they're trying to work in this dynamic environment where things are changing all of the time?

329
00:35:58,520 --> 00:36:07,790
And a good way to think about this from a mathematical point of view is to think of the ants as balancing on a seesaw.

330
00:36:07,790 --> 00:36:19,550
So here is my first aunt. She's on this seesaw and this seesaw is like the left right choice where the food to look for the food.

331
00:36:19,550 --> 00:36:24,170
And she maybe makes the left choice, she goes in that direction.

332
00:36:24,170 --> 00:36:29,450
Now what she does is when she makes a left choice, she leaves Ferryman going towards that choice.

333
00:36:29,450 --> 00:36:33,890
And so that bias is the decision for the next stamps which come along.

334
00:36:33,890 --> 00:36:39,950
The ants don't actually walk on seafood on a seesaw. They're on this Y-shaped maze.

335
00:36:39,950 --> 00:36:46,580
But the idea is to say that once they start going in one direction, they bias the choice towards that direction.

336
00:36:46,580 --> 00:36:51,440
So the next time comes along and she follows the previous one down next one,

337
00:36:51,440 --> 00:36:58,760
she does the same thing, and they keep following one after the other down towards the food.

338
00:36:58,760 --> 00:37:04,340
Now this is good. They've managed to find this lollipop, which was the food that they were looking for.

339
00:37:04,340 --> 00:37:09,380
But the problem is, once they've tipped the seesaw in this direction,

340
00:37:09,380 --> 00:37:17,120
if the food moves up here, how do they get back to the to the food, to the new food source?

341
00:37:17,120 --> 00:37:28,010
And they can't. They're sort of stuck in the nuance which will come along will follow the other end, so follow the trail down in the wrong direction.

342
00:37:28,010 --> 00:37:35,520
So. What you can do if you're an end is you can tip a bit less, so here we'll have the first and.

343
00:37:35,520 --> 00:37:40,420
And so this is equivalent to having less reaction to the ceremony.

344
00:37:40,420 --> 00:37:47,470
She's making a decision. She goes that way. And you'll see that the seesaw is balanced slightly less so that she's left less ferryman.

345
00:37:47,470 --> 00:37:54,460
Or she's reacted less strongly to the ferryman than the next and maybe goes the other way and balances a seesaw up again.

346
00:37:54,460 --> 00:37:56,620
And then you start to get off in that direction.

347
00:37:56,620 --> 00:38:07,540
But when it starts to get it, it never really gets too unbalanced because they're making decisions to go left or right, basically in a random way.

348
00:38:07,540 --> 00:38:14,710
But the problem now is if you don't tip enough, then you never really get focussed on the best food,

349
00:38:14,710 --> 00:38:19,230
so the food is moving around in different directions.

350
00:38:19,230 --> 00:38:28,400
And the answer following the food, but they never really they never really tip over strongly to One Direction or the other.

351
00:38:28,400 --> 00:38:35,960
And it turns out the answer for the ends is to kind of get their seesaw on a balance like this.

352
00:38:35,960 --> 00:38:42,380
So by getting that see saw it is sort of tipping point, they just enough, but not too much.

353
00:38:42,380 --> 00:38:48,350
What they can do is that they can track where the lollipop or the food is moving to.

354
00:38:48,350 --> 00:38:53,770
And this means getting exactly the right responsiveness to the food.

355
00:38:53,770 --> 00:39:01,300
The mathematical model behind this is is consists of two different equations which are based on the reward equation.

356
00:39:01,300 --> 00:39:07,780
But what we've done is we have this feedback term here as well. This is a feedback and this is a feedback here.

357
00:39:07,780 --> 00:39:11,710
And this model's both the the seesaw idea that I've talked about.

358
00:39:11,710 --> 00:39:19,900
It models the Y. May's decision, and it can also model how we make our choices about social media.

359
00:39:19,900 --> 00:39:23,080
I'm not going to go into the details of solving these equations,

360
00:39:23,080 --> 00:39:32,590
but when you solve them using a theory called bifurcation theory, you can actually show that if you have two weaker response,

361
00:39:32,590 --> 00:39:42,010
if you are not very sensitive to the Ferryman trail, you get the situation like this that they just sort of randomly distributed.

362
00:39:42,010 --> 00:39:46,390
If you get too strong a response, they can tip and go in the wrong direction.

363
00:39:46,390 --> 00:39:53,950
But if you get this just right response, then the answer can be very flexible when the food sources change position.

364
00:39:53,950 --> 00:40:01,480
And this is something you find in all animals, including ourselves. So.

365
00:40:01,480 --> 00:40:06,300
This is something you find in all animals, including ourselves, this is just a few examples here.

366
00:40:06,300 --> 00:40:12,120
This is humans applauding after a talk. This is fish escaping when we frighten them.

367
00:40:12,120 --> 00:40:18,540
And over here we've got bird flocks flying around. And what animals have is a collective FOMO,

368
00:40:18,540 --> 00:40:27,540
a collective fear of missing out that always balanced at this point where they can switch direction to to move towards food,

369
00:40:27,540 --> 00:40:35,820
to move towards shelter, to move towards a mate. And that's kind of what we are in the internet, on the internet as well.

370
00:40:35,820 --> 00:40:39,120
If you want to collect all the rewards that are available to you,

371
00:40:39,120 --> 00:40:44,820
then you have to be there all of the time, clicking and searching and trying to find out what's going on.

372
00:40:44,820 --> 00:40:47,010
And that's just incredibly stressful.

373
00:40:47,010 --> 00:40:55,680
It kind of lies in our collective nature to do this, but it's a very stressful thing to be doing on a daily basis.

374
00:40:55,680 --> 00:41:03,210
So I have a recommendation that I've also learnt from the hands. Another thing the ants are very good at is just taking it easy.

375
00:41:03,210 --> 00:41:11,010
The ants do this running around like crazy. They put themselves at tipping point when they're working for ants also like to rest.

376
00:41:11,010 --> 00:41:12,480
They like to sleep.

377
00:41:12,480 --> 00:41:25,530
And so I have this picture of my wife plays a lot of Pokemon Go and have this picture of Snorlax resting because as well as you can at times,

378
00:41:25,530 --> 00:41:34,560
go crazy and try and follow all the information. But other times you want to take it easy and have a sleep like Snorlax does.

379
00:41:34,560 --> 00:41:42,810
OK. I've told you about two of the equations today, I've told you about the judgement equation and the reward equation.

380
00:41:42,810 --> 00:41:48,450
There are if there's 10 equations, of course, eight other equations in the book,

381
00:41:48,450 --> 00:41:53,670
if you've been looking carefully, you'll see that I've listed seven of them here.

382
00:41:53,670 --> 00:41:58,260
These equations are used throughout our lives,

383
00:41:58,260 --> 00:42:03,510
the learning equation was used by YouTube when they wanted to learn about what videos we like to

384
00:42:03,510 --> 00:42:11,190
increased massively our use of YouTube by finding out what we liked and showing us more of it.

385
00:42:11,190 --> 00:42:18,300
The advertising equation is used by Facebook to find out what we're interested in and target adverts at us.

386
00:42:18,300 --> 00:42:24,540
It's the correlation equation as well. The learning equation is a derivative of maximisation.

387
00:42:24,540 --> 00:42:31,170
The advertising equation is the correlation equation. The confidence equation is based on the confidence interval.

388
00:42:31,170 --> 00:42:37,770
It's used both in gambling to find out if you really have got a winning strategy,

389
00:42:37,770 --> 00:42:43,290
and it's also used widely in research in order to find out if our results are really true.

390
00:42:43,290 --> 00:42:49,350
The influencer equation dice was first invented by Google or used by Google.

391
00:42:49,350 --> 00:42:52,620
It's actually based on mathematics, which is 100 years old,

392
00:42:52,620 --> 00:42:59,580
and it's used to measure who are the most important people in the social network or on on the internet.

393
00:42:59,580 --> 00:43:10,380
The market equation is used in finance. It's a diffusion equation which describes as I break it down into signal noise and feedback

394
00:43:10,380 --> 00:43:15,870
there the way ways the market work and how you can use that to make a lot of money.

395
00:43:15,870 --> 00:43:23,880
The skill equation is used in sports to evaluate the quality of different football players, for example,

396
00:43:23,880 --> 00:43:31,620
and the betting equation is well, it's used in betting, but also about how how we can use probability in our own lives.

397
00:43:31,620 --> 00:43:38,700
So there's all these equations or these equations, which allow you to get a better understanding of the world.

398
00:43:38,700 --> 00:43:49,080
And they all combine models where you describe things as a mathematical equation and data that you take in the data can be things from social media,

399
00:43:49,080 --> 00:43:58,740
from finance, from banks and so on. Combining models and data is the secret to successfully using these equations.

400
00:43:58,740 --> 00:44:02,130
There's a tense equation as well. And so what's that?

401
00:44:02,130 --> 00:44:10,140
Well, all of these equations, they involve combining different types of models with data.

402
00:44:10,140 --> 00:44:13,290
Now, the tenth equation is different.

403
00:44:13,290 --> 00:44:23,520
I call it the universal equation, and I like to say it as if then, because eventually this is going to end up as a logical statement,

404
00:44:23,520 --> 00:44:28,350
but it goes back all the way to what I mentioned at the start of the talk.

405
00:44:28,350 --> 00:44:30,830
The DA Vinci Code.

406
00:44:30,830 --> 00:44:38,990
Now, the da Vinci Code is a fascinating idea that there could be a single equation that unifies everything and brings the world together.

407
00:44:38,990 --> 00:44:47,500
The single theory of everything which all revolves around, in this case, phi a single number.

408
00:44:47,500 --> 00:44:51,640
But I believe that this can't be the case.

409
00:44:51,640 --> 00:45:02,010
And I'm just going to give you a brief idea about why we can't have a single number, a single mathematical idea that unifies the entire universe.

410
00:45:02,010 --> 00:45:09,960
If we take, for example, fine, where does it come from? Well, fi, I mentioned at the start, it comes from the golden rectangle.

411
00:45:09,960 --> 00:45:16,320
One side of the rectangle has lent a the other side of the rectangle as a a plus b.

412
00:45:16,320 --> 00:45:18,420
And this is a golden rectangle.

413
00:45:18,420 --> 00:45:30,540
If the rectangle that's left over when you put in the square, the ratio of those sides B to A is the same as the ratio of A-plus beat A.

414
00:45:30,540 --> 00:45:41,630
And that's given by this equation here. A-plus B divided by a equals to a of a b and is equal to PHI, which is the golden ratio.

415
00:45:41,630 --> 00:45:47,900
So that's one place in which fire comes up and what Dan Brown emphasised in his book,

416
00:45:47,900 --> 00:45:52,430
he came up with quite a few examples which weren't true, but he came out with quite a few that are true.

417
00:45:52,430 --> 00:45:55,910
Is that fire causing quite a lot of different settings? And here's one of them.

418
00:45:55,910 --> 00:46:02,870
Here is the Fibonacci sequence. This is this the Fibonacci sequence itself,

419
00:46:02,870 --> 00:46:09,230
and the rule is as follows you have one plus one is equal to two one plus two is equal to three to plus three is equal to five,

420
00:46:09,230 --> 00:46:13,310
three plus five is equal to eight and so on, up and up and up.

421
00:46:13,310 --> 00:46:18,500
And where does that golden ratio occur? Well, it occurs here.

422
00:46:18,500 --> 00:46:21,290
So 13 divided by eight is equal to one point six two five.

423
00:46:21,290 --> 00:46:33,350
So we take consecutive numbers and we divide them by each other and we get closer and closer as we go off into infinitely large numbers.

424
00:46:33,350 --> 00:46:38,950
That ratio gets closer to one point six on eight.

425
00:46:38,950 --> 00:46:47,200
And this is one of a number of different mathematical results where the same number appears in lots of different settings,

426
00:46:47,200 --> 00:46:56,390
and it's this that gives the kind of semi mystical feeling about mathematics that there's some sort of hidden truth behind those numbers.

427
00:46:56,390 --> 00:47:00,620
And it also it's always confused mathematicians right through history.

428
00:47:00,620 --> 00:47:09,350
So it led Henri Poincaré to reflect on this, and he said, and this is a rhetorical question really by him.

429
00:47:09,350 --> 00:47:15,050
He said that if all the assertions of mathematics put forward can be derived from one another by formal logic,

430
00:47:15,050 --> 00:47:20,190
just like we've we've looked at with or I'm going to look at now with fine.

431
00:47:20,190 --> 00:47:28,550
The mathematics can't be anything than some sort of immense tautology, so that just means everything in mathematics is true.

432
00:47:28,550 --> 00:47:37,020
Logical inference can teach us essentially nothing new. But can it really be the case that all of these books that we look at in mathematics

433
00:47:37,020 --> 00:47:42,540
have no other purpose than in a roundabout way to say that A is equal to a?

434
00:47:42,540 --> 00:47:47,100
Now this is a rhetorical question on the part of CAR.

435
00:47:47,100 --> 00:47:56,940
This is. He can't really believe, I think when he states this, that all of mathematics just says that a is equal to a.

436
00:47:56,940 --> 00:48:04,440
So if we take the golden rectangle here and and the equality we've set up for the different sides of it.

437
00:48:04,440 --> 00:48:14,530
What we can do is we can rewrite this by we have a divided by a plus b divided by A is equal to a as a basis.

438
00:48:14,530 --> 00:48:19,750
That's fine. But we've also said Overby, well, that's equal to the golden ratio PHI.

439
00:48:19,750 --> 00:48:26,140
So let's just put Phi in here and be over is one over a baby.

440
00:48:26,140 --> 00:48:31,500
So we have one over five there and we have one.

441
00:48:31,500 --> 00:48:43,220
Yeah. And then if we multiply through by five, we have five here, plus one is equal to five squared.

442
00:48:43,220 --> 00:48:51,740
And we always like to have everything over on one side, so we have a nice quadratic equation and the quadratic equation for this,

443
00:48:51,740 --> 00:49:02,690
if we put up zero on this side and we've got phi squared minus five minus one.

444
00:49:02,690 --> 00:49:11,330
And if you remember the roots of the quadratic equation, your views, the quadratic formula,

445
00:49:11,330 --> 00:49:23,760
and that will be if we've got it right here because we have it here, but it's a one plus the square root of.

446
00:49:23,760 --> 00:49:36,450
One plus four. Divided by two, which is one plus or plus a minus plus or minus the square root of five, divided by two.

447
00:49:36,450 --> 00:49:42,590
So that's how we come up with the value of PHI. Now.

448
00:49:42,590 --> 00:49:49,430
That's the problem from the point of view of the golden rectangle.

449
00:49:49,430 --> 00:49:52,920
If we look at the other problem.

450
00:49:52,920 --> 00:50:08,730
So then if we look at the Fibonacci sequence, what we know about the Fibonacci sequence is we have x t plus one is equal to x t plus x T minus one.

451
00:50:08,730 --> 00:50:15,000
So what I mean by that? Well, if you remember the sequence one plus one was equal to two, one plus two is equal to three.

452
00:50:15,000 --> 00:50:21,170
And I'm just writing that same thing out in an algebraic form.

453
00:50:21,170 --> 00:50:26,510
Now what we can do is we can say, well, what happens when T gets very large?

454
00:50:26,510 --> 00:50:36,320
And first of all, if I just write out x t plus one divided by x t.

455
00:50:36,320 --> 00:50:46,870
Then that. Is equal to x t plus x T minus one.

456
00:50:46,870 --> 00:51:00,330
Divide it by x t. And this is also equal to one plus x T minus one divided by x t.

457
00:51:00,330 --> 00:51:11,530
Now, if I take the limits of tea goes to infinity in this, and that means I just keep going in lots and lots of steps.

458
00:51:11,530 --> 00:51:15,910
Exactly as I've done here, I keep going forward and forward in time.

459
00:51:15,910 --> 00:51:23,500
So if I want to know what this limit is, I can replace it with five.

460
00:51:23,500 --> 00:51:30,520
And that's going to be equal to one plus. I'm here because I'm going off to the limits of time equals infinity.

461
00:51:30,520 --> 00:51:33,220
This is just one over five.

462
00:51:33,220 --> 00:51:49,610
Once we come up in further and further in enough time steps and that equation, you should recognise from there because it's the same equation again.

463
00:51:49,610 --> 00:52:05,130
And it's going to have the same solution again. And so the occurrence of these types of relationships.

464
00:52:05,130 --> 00:52:10,200
Isn't so much, it isn't so much a deep secret.

465
00:52:10,200 --> 00:52:15,640
It's more exactly as Henry Henry Pryor, said Henry Ford.

466
00:52:15,640 --> 00:52:24,480
It's it isn't so much. It isn't so much a deep secret, but it's more exactly as Poincaré originally proposed.

467
00:52:24,480 --> 00:52:31,680
What I've really shown here is that a quasi we get the same solution for these two different mathematical problems.

468
00:52:31,680 --> 00:52:38,010
Ultimately, for problems like this, for pure mathematics problems which aren't coupled to data,

469
00:52:38,010 --> 00:52:46,000
then they do ultimately just prove that a quasi and rhetorically was a rhetorical question that Poincaré posed.

470
00:52:46,000 --> 00:52:53,220
The answer is yes. And when you look at philosophy of mathematics, they come to the same answer.

471
00:52:53,220 --> 00:52:59,340
What you're doing when you decouple a problem from reality is you're just proving that a equals eight.

472
00:52:59,340 --> 00:53:09,940
And in that way, you can never expect a secret number of any type to inform what your form something about the structure of your life.

473
00:53:09,940 --> 00:53:16,610
But. Then the question could mathematics hold the answers we're all looking for?

474
00:53:16,610 --> 00:53:27,620
I believe it can. But only when we let data into our lives as well is when we combine.

475
00:53:27,620 --> 00:53:32,800
Data in the form of all the things that happened to us.

476
00:53:32,800 --> 00:53:37,330
With a model, a mathematical description of the world.

477
00:53:37,330 --> 00:54:08,297
That we can actually start to understand what's happening and better control our lives and maybe become more successful and happier people.