English-Video.net comment policy

The comment field is common to all languages

Let's write in your language and use "Google Translate" together

Please refer to informative community guidelines on TED.com

TEDGlobal 2011

Tim Harford: Trial, error and the God complex

Filmed
Views 1,750,507

Economics writer Tim Harford studies complex systems -- and finds a surprising link among the successful ones: they were built through trial and error. In this sparkling talk from TEDGlobal 2011, he asks us to embrace our randomness and start making better mistakes.

- Economist, journalist and broadcaster
Tim Harford's writings reveal the economic ideas behind everyday experiences. Full bio

It's the Second World War.
00:15
A German prison camp.
00:17
And this man,
00:20
Archie Cochrane,
00:23
is a prisoner of war and a doctor,
00:26
and he has a problem.
00:29
The problem is that the men under his care
00:32
are suffering
00:35
from an excruciating and debilitating condition
00:37
that Archie doesn't really understand.
00:40
The symptoms
00:43
are this horrible swelling up of fluids under the skin.
00:45
But he doesn't know whether it's an infection, whether it's to do with malnutrition.
00:48
He doesn't know how to cure it.
00:51
And he's operating in a hostile environment.
00:53
And people do terrible things in wars.
00:56
The German camp guards, they've got bored.
00:58
They've taken to just firing into the prison camp at random
01:01
for fun.
01:03
On one particular occasion,
01:05
one of the guards threw a grenade into the prisoners' lavatory
01:07
while it was full of prisoners.
01:10
He said he heard suspicious laughter.
01:13
And Archie Cochrane, as the camp doctor,
01:15
was one of the first men in
01:18
to clear up the mess.
01:20
And one more thing:
01:22
Archie was suffering from this illness himself.
01:24
So the situation seemed pretty desperate.
01:27
But Archie Cochrane
01:30
was a resourceful person.
01:32
He'd already smuggled vitamin C into the camp,
01:35
and now he managed
01:38
to get hold of supplies of marmite
01:40
on the black market.
01:42
Now some of you will be wondering what marmite is.
01:44
Marmite is a breakfast spread beloved of the British.
01:47
It looks like crude oil.
01:50
It tastes ...
01:52
zesty.
01:54
And importantly,
01:56
it's a rich source
01:58
of vitamin B12.
02:00
So Archie splits the men under his care as best he can
02:02
into two equal groups.
02:05
He gives half of them vitamin C.
02:07
He gives half of them vitamin B12.
02:09
He very carefully and meticulously notes his results
02:12
in an exercise book.
02:15
And after just a few days,
02:17
it becomes clear
02:19
that whatever is causing this illness,
02:21
marmite is the cure.
02:24
So Cochrane then goes to the Germans who are running the prison camp.
02:27
Now you've got to imagine at the moment --
02:30
forget this photo, imagine this guy
02:32
with this long ginger beard and this shock of red hair.
02:34
He hasn't been able to shave -- a sort of Billy Connolly figure.
02:37
Cochrane, he starts ranting at these Germans
02:40
in this Scottish accent --
02:42
in fluent German, by the way, but in a Scottish accent --
02:44
and explains to them how German culture was the culture
02:47
that gave Schiller and Goethe to the world.
02:50
And he can't understand
02:52
how this barbarism can be tolerated,
02:54
and he vents his frustrations.
02:56
And then he goes back to his quarters,
02:59
breaks down and weeps
03:02
because he's convinced that the situation is hopeless.
03:05
But a young German doctor
03:10
picks up Archie Cochrane's exercise book
03:13
and says to his colleagues,
03:16
"This evidence is incontrovertible.
03:20
If we don't supply vitamins to the prisoners,
03:25
it's a war crime."
03:28
And the next morning,
03:30
supplies of vitamin B12 are delivered to the camp,
03:32
and the prisoners begin to recover.
03:35
Now I'm not telling you this story
03:39
because I think Archie Cochrane is a dude,
03:41
although Archie Cochrane is a dude.
03:43
I'm not even telling you the story
03:47
because I think we should be running
03:49
more carefully controlled randomized trials
03:51
in all aspects of public policy,
03:53
although I think that would also be completely awesome.
03:55
I'm telling you this story
03:59
because Archie Cochrane, all his life,
04:01
fought against a terrible affliction,
04:04
and he realized it was debilitating to individuals
04:08
and it was corrosive to societies.
04:12
And he had a name for it.
04:14
He called it the God complex.
04:16
Now I can describe the symptoms of the God complex very, very easily.
04:19
So the symptoms of the complex
04:23
are, no matter how complicated the problem,
04:26
you have an absolutely overwhelming belief
04:29
that you are infallibly right in your solution.
04:32
Now Archie was a doctor,
04:36
so he hung around with doctors a lot.
04:38
And doctors suffer from the God complex a lot.
04:40
Now I'm an economist, I'm not a doctor,
04:43
but I see the God complex around me all the time
04:45
in my fellow economists.
04:47
I see it in our business leaders.
04:49
I see it in the politicians we vote for --
04:51
people who, in the face of an incredibly complicated world,
04:53
are nevertheless absolutely convinced
04:57
that they understand the way that the world works.
05:00
And you know, with the future billions that we've been hearing about,
05:03
the world is simply far too complex
05:06
to understand in that way.
05:08
Well let me give you an example.
05:10
Imagine for a moment
05:12
that, instead of Tim Harford in front of you,
05:14
there was Hans Rosling presenting his graphs.
05:16
You know Hans:
05:19
the Mick Jagger of TED.
05:21
(Laughter)
05:23
And he'd be showing you these amazing statistics,
05:25
these amazing animations.
05:27
And they are brilliant; it's wonderful work.
05:29
But a typical Hans Rosling graph:
05:31
think for a moment, not what it shows,
05:33
but think instead about what it leaves out.
05:36
So it'll show you GDP per capita,
05:39
population, longevity,
05:42
that's about it.
05:44
So three pieces of data for each country --
05:46
three pieces of data.
05:48
Three pieces of data is nothing.
05:50
I mean, have a look at this graph.
05:52
This is produced by the physicist Cesar Hidalgo.
05:54
He's at MIT.
05:56
Now you won't be able to understand a word of it,
05:58
but this is what it looks like.
06:00
Cesar has trolled the database
06:02
of over 5,000 different products,
06:04
and he's used techniques of network analysis
06:07
to interrogate this database
06:12
and to graph relationships between the different products.
06:14
And it's wonderful, wonderful work.
06:16
You show all these interconnections, all these interrelations.
06:18
And I think it'll be profoundly useful
06:21
in understanding how it is that economies grow.
06:23
Brilliant work.
06:26
Cesar and I tried to write a piece for The New York Times Magazine
06:28
explaining how this works.
06:30
And what we learned
06:32
is Cesar's work is far too good to explain
06:34
in The New York Times Magazine.
06:36
Five thousand products --
06:40
that's still nothing.
06:43
Five thousand products --
06:45
imagine counting every product category
06:47
in Cesar Hidalgo's data.
06:49
Imagine you had one second
06:51
per product category.
06:53
In about the length of this session,
06:55
you would have counted all 5,000.
06:58
Now imagine doing the same thing
07:00
for every different type of product on sale in Walmart.
07:02
There are 100,000 there. It would take you all day.
07:05
Now imagine trying to count
07:08
every different specific product and service
07:10
on sale in a major economy
07:13
such as Tokyo, London or New York.
07:15
It's even more difficult in Edinburgh
07:17
because you have to count all the whisky and the tartan.
07:19
If you wanted to count every product and service
07:22
on offer in New York --
07:24
there are 10 billion of them --
07:26
it would take you 317 years.
07:28
This is how complex the economy we've created is.
07:31
And I'm just counting toasters here.
07:34
I'm not trying to solve the Middle East problem.
07:36
The complexity here is unbelievable.
07:39
And just a piece of context --
07:42
the societies in which our brains evolved
07:44
had about 300 products and services.
07:46
You could count them in five minutes.
07:48
So this is the complexity of the world that surrounds us.
07:51
This perhaps is why
07:54
we find the God complex so tempting.
07:56
We tend to retreat and say, "We can draw a picture,
07:59
we can post some graphs,
08:02
we get it, we understand how this works."
08:04
And we don't.
08:07
We never do.
08:09
Now I'm not trying to deliver a nihilistic message here.
08:11
I'm not trying to say we can't solve
08:13
complicated problems in a complicated world.
08:15
We clearly can.
08:17
But the way we solve them
08:19
is with humility --
08:21
to abandon the God complex
08:23
and to actually use a problem-solving technique that works.
08:25
And we have a problem-solving technique that works.
08:28
Now you show me
08:31
a successful complex system,
08:33
and I will show you a system
08:35
that has evolved through trial and error.
08:38
Here's an example.
08:40
This baby was produced through trial and error.
08:42
I realize that's an ambiguous statement.
08:46
Maybe I should clarify it.
08:49
This baby is a human body: it evolved.
08:51
What is evolution?
08:54
Over millions of years, variation and selection,
08:56
variation and selection --
08:59
trial and error,
09:02
trial and error.
09:04
And it's not just biological systems
09:07
that produce miracles through trial and error.
09:09
You could use it in an industrial context.
09:11
So let's say you wanted to make detergent.
09:13
Let's say you're Unilever
09:15
and you want to make detergent in a factory near Liverpool.
09:17
How do you do it?
09:20
Well you have this great big tank full of liquid detergent.
09:22
You pump it at a high pressure through a nozzle.
09:25
You create a spray of detergent.
09:27
Then the spray dries. It turns into powder.
09:30
It falls to the floor.
09:32
You scoop it up. You put it in cardboard boxes.
09:34
You sell it at a supermarket.
09:36
You make lots of money.
09:38
How do you design that nozzle?
09:40
It turns out to be very important.
09:43
Now if you ascribe to the God complex,
09:46
what you do is you find yourself a little God.
09:48
You find yourself a mathematician; you find yourself a physicist --
09:51
somebody who understands the dynamics of this fluid.
09:54
And he will, or she will,
09:57
calculate the optimal design of the nozzle.
10:00
Now Unilever did this and it didn't work --
10:03
too complicated.
10:05
Even this problem, too complicated.
10:07
But the geneticist Professor Steve Jones
10:10
describes how Unilever actually did solve this problem --
10:13
trial and error,
10:16
variation and selection.
10:18
You take a nozzle
10:20
and you create 10 random variations on the nozzle.
10:22
You try out all 10; you keep the one that works best.
10:26
You create 10 variations on that one.
10:29
You try out all 10. You keep the one that works best.
10:31
You try out 10 variations on that one.
10:34
You see how this works, right?
10:36
And after 45 generations,
10:38
you have this incredible nozzle.
10:40
It looks a bit like a chess piece --
10:42
functions absolutely brilliantly.
10:44
We have no idea
10:47
why it works,
10:49
no idea at all.
10:51
And the moment you step back from the God complex --
10:53
let's just try to have a bunch of stuff;
10:55
let's have a systematic way of determining what's working and what's not --
10:57
you can solve your problem.
11:00
Now this process of trial and error
11:02
is actually far more common in successful institutions
11:04
than we care to recognize.
11:07
And we've heard a lot about how economies function.
11:09
The U.S. economy is still the world's greatest economy.
11:12
How did it become the world's greatest economy?
11:16
I could give you all kinds of facts and figures
11:19
about the U.S. economy,
11:21
but I think the most salient one is this:
11:23
ten percent of American businesses
11:26
disappear every year.
11:29
That is a huge failure rate.
11:32
It's far higher than the failure rate of, say, Americans.
11:35
Ten percent of Americans don't disappear every year.
11:37
Which leads us to conclude
11:40
American businesses fail faster than Americans,
11:42
and therefore American businesses are evolving faster than Americans.
11:45
And eventually, they'll have evolved to such a high peak of perfection
11:48
that they will make us all their pets --
11:51
(Laughter)
11:54
if, of course, they haven't already done so.
11:56
I sometimes wonder.
11:59
But it's this process of trial and error
12:02
that explains this great divergence,
12:04
this incredible performance of Western economies.
12:08
It didn't come because you put some incredibly smart person in charge.
12:11
It's come through trial and error.
12:14
Now I've been sort of banging on about this
12:16
for the last couple of months,
12:18
and people sometimes say to me,
12:20
"Well Tim, it's kind of obvious.
12:22
Obviously trial and error is very important.
12:24
Obviously experimentation is very important.
12:26
Now why are you just wandering around saying this obvious thing?"
12:28
So I say, okay, fine.
12:31
You think it's obvious?
12:33
I will admit it's obvious
12:35
when schools
12:37
start teaching children
12:39
that there are some problems that don't have a correct answer.
12:42
Stop giving them lists of questions
12:45
every single one of which has an answer.
12:48
And there's an authority figure in the corner
12:50
behind the teacher's desk who knows all the answers.
12:52
And if you can't find the answers,
12:54
you must be lazy or stupid.
12:56
When schools stop doing that all the time,
12:58
I will admit that, yes,
13:00
it's obvious that trial and error is a good thing.
13:02
When a politician stands up
13:04
campaigning for elected office
13:07
and says, "I want to fix our health system.
13:09
I want to fix our education system.
13:11
I have no idea how to do it.
13:13
I have half a dozen ideas.
13:16
We're going to test them out. They'll probably all fail.
13:18
Then we'll test some other ideas out.
13:21
We'll find some that work. We'll build on those.
13:23
We'll get rid of the ones that don't." --
13:25
when a politician campaigns on that platform,
13:27
and more importantly, when voters like you and me
13:30
are willing to vote for that kind of politician,
13:33
then I will admit
13:35
that it is obvious that trial and error works, and that -- thank you.
13:37
(Applause)
13:40
Until then, until then
13:44
I'm going to keep banging on about trial and error
13:47
and why we should abandon the God complex.
13:49
Because it's so hard
13:52
to admit our own fallibility.
13:55
It's so uncomfortable.
13:57
And Archie Cochrane understood this as well as anybody.
13:59
There's this one trial he ran
14:02
many years after World War II.
14:04
He wanted to test out
14:06
the question of, where is it
14:09
that patients should recover
14:11
from heart attacks?
14:13
Should they recover in a specialized cardiac unit in hospital,
14:15
or should they recover at home?
14:18
All the cardiac doctors tried to shut him down.
14:21
They had the God complex in spades.
14:24
They knew that their hospitals were the right place for patients,
14:27
and they knew it was very unethical
14:30
to run any kind of trial or experiment.
14:32
Nevertheless, Archie managed to get permission to do this.
14:35
He ran his trial.
14:37
And after the trial had been running for a little while,
14:39
he gathered together all his colleagues
14:41
around his table,
14:43
and he said, "Well, gentlemen,
14:45
we have some preliminary results.
14:47
They're not statistically significant.
14:49
But we have something.
14:51
And it turns out that you're right and I'm wrong.
14:54
It is dangerous for patients
14:57
to recover from heart attacks at home.
14:59
They should be in hospital."
15:01
And there's this uproar, and all the doctors start pounding the table
15:04
and saying, "We always said you were unethical, Archie.
15:06
You're killing people with your clinical trials. You need to shut it down now.
15:09
Shut it down at once."
15:12
And there's this huge hubbub.
15:14
Archie lets it die down.
15:16
And then he says, "Well that's very interesting, gentlemen,
15:18
because when I gave you the table of results,
15:20
I swapped the two columns around.
15:23
It turns out your hospitals are killing people,
15:27
and they should be at home.
15:29
Would you like to close down the trial now,
15:31
or should we wait until we have robust results?"
15:34
Tumbleweed
15:38
rolls through the meeting room.
15:40
But Cochrane would do that kind of thing.
15:43
And the reason he would do that kind of thing
15:46
is because he understood
15:48
it feels so much better
15:50
to stand there and say,
15:52
"Here in my own little world,
15:54
I am a god, I understand everything.
15:56
I do not want to have my opinions challenged.
15:58
I do not want to have my conclusions tested."
16:00
It feels so much more comfortable
16:03
simply to lay down the law.
16:05
Cochrane understood
16:08
that uncertainty, that fallibility,
16:10
that being challenged, they hurt.
16:12
And you sometimes need to be shocked out of that.
16:14
Now I'm not going to pretend that this is easy.
16:18
It isn't easy.
16:21
It's incredibly painful.
16:23
And since I started talking about this subject
16:25
and researching this subject,
16:27
I've been really haunted by something
16:29
a Japanese mathematician said on the subject.
16:31
So shortly after the war,
16:33
this young man, Yutaka Taniyama,
16:35
developed this amazing conjecture
16:38
called the Taniyama-Shimura Conjecture.
16:40
It turned out to be absolutely instrumental
16:42
many decades later
16:45
in proving Fermat's Last Theorem.
16:47
In fact, it turns out it's equivalent
16:49
to proving Fermat's Last Theorem.
16:51
You prove one, you prove the other.
16:53
But it was always a conjecture.
16:57
Taniyama tried and tried and tried
17:00
and he could never prove that it was true.
17:03
And shortly before his 30th birthday in 1958,
17:06
Yutaka Taniyama killed himself.
17:09
His friend, Goro Shimura --
17:13
who worked on the mathematics with him --
17:15
many decades later, reflected on Taniyama's life.
17:17
He said,
17:22
"He was not a very careful person
17:25
as a mathematician.
17:27
He made a lot of mistakes.
17:29
But he made mistakes in a good direction.
17:32
I tried to emulate him,
17:36
but I realized
17:39
it is very difficult
17:41
to make good mistakes."
17:43
Thank you.
17:46
(Applause)
17:48

▲Back to top

About the speaker:

Tim Harford - Economist, journalist and broadcaster
Tim Harford's writings reveal the economic ideas behind everyday experiences.

Why you should listen

In the Undercover Economist column he writes for the Financial Times, Tim Harford looks at familiar situations in unfamiliar ways and explains the fundamental principles of the modern economy. He illuminates them with clear writing and a variety of examples borrowed from daily life.

His book, Adapt: Why Success Always Starts With Failure, argues that the world has become far too unpredictable and complex for today's challenges to be tackled with ready-made solutions and expert opinions. Instead, Harford suggests, we need to learn to embrace failure and to constantly adapt, to improvise rather than plan, to work from the bottom up rather than the top down. His next book, Messy: Thriving in a Tidy-Minded World will be published in September 2016. 

Harford also presents the BBC radio series More or Less, a rare broadcast program devoted, as he says, to "the powerful, sometimes beautiful, often abused but ever ubiquitous world of numbers."

He says: "I’d like to see many more complex problems approached with a willingness to experiment."

More profile about the speaker
Tim Harford | Speaker | TED.com