3 ways to make better decisions by thinking like a computer Tom Griffiths
If there’s one city in the world
where it’s hard to find
a place to buy or rent,
it’s Sydney.
And if you’ve tried
to find a home here recently,
you’re familiar with the problem.
Every time you walk into an open house,
you get some information
about what’s out there
and what’s on the market,
but every time you walk out,
you’re running the risk
of the very best place passing you by.
So how do you know when
to switch from looking
to being ready to make an offer?
This is such a cruel and familiar problem
that it might come as a surprise
that it has a simple solution.
37 percent.
(Laughter)
If you want to maximize the probability
that you find the very best place,
you should look at 37 percent
of what’s on the market,
and then make an offer
on the next place you see,
which is better than anything
that you’ve seen so far.
Or if you’re looking for a month,
take 37 percent of that time –
11 days, to set a standard –
and then you’re ready to act.
We know this because
trying to find a place to live
is an example of an optimal
stopping problem.
A class of problems that has been
studied extensively
by mathematicians and computer scientists.
I’m a computational cognitive scientist.
I spend my time trying to understand
how it is that human minds work,
from our amazing successes
to our dismal failures.
To do that, I think about
the computational structure
of the problems
that arise in everyday life,
and compare the ideal
solutions to those problems
to the way that we actually behave.
As a side effect,
I get to see how applying
a little bit of computer science
can make human decision-making easier.
I have a personal motivation for this.
Growing up in Perth
as an overly cerebral kid …
(Laughter)
I would always try and act in the way
that I thought was rational,
reasoning through every decision,
trying to figure out
the very best action to take.
But this is an approach
that doesn’t scale up
when you start to run into
the sorts of problems
that arise in adult life.
At one point, I even tried
to break up with my girlfriend
because trying to take into account
her preferences as well as my own
and then find perfect solutions –
(Laughter)
was just leaving me exhausted.
(Laughter)
She pointed out that I was taking
the wrong approach
to solving this problem –
and she later became my wife.
(Laughter)
(Applause)
Whether it’s as basic as trying to decide
what restaurant to go to
or as important as trying to decide
who to spend the rest of your life with,
human lives are filled
with computational problems
that are just too hard to solve
by applying sheer effort.
For those problems,
it’s worth consulting the experts:
computer scientists.
(Laughter)
When you’re looking for life advice,
computer scientists probably aren’t
the first people you think to talk to.
Living life like a computer –
stereotypically deterministic,
exhaustive and exact –
doesn’t sound like a lot of fun.
But thinking about the computer science
of human decisions
reveals that in fact,
we’ve got this backwards.
When applied to the sorts
of difficult problems
that arise in human lives,
the way that computers
actually solve those problems
looks a lot more like the way
that people really act.
Take the example of trying to decide
what restaurant to go to.
This is a problem that has
a particular computational structure.
You’ve got a set of options,
you’re going to choose
one of those options,
and you’re going to face
exactly the same decision tomorrow.
In that situation,
you run up against
what computer scientists call
the “explore-exploit trade-off.”
You have to make a decision
about whether you’re going
to try something new –
exploring, gathering some information
that you might be able
to use in the future –
or whether you’re going to go to a place
that you already know is pretty good –
exploiting the information
that you’ve already gathered so far.
The explore/exploit trade-off
shows up any time you have to choose
between trying something new
and going with something
that you already know is pretty good,
whether it’s listening to music
or trying to decide
who you’re going to spend time with.
It’s also the problem
that technology companies face
when they’re trying to do something
like decide what ad to show on a web page.
Should they show a new ad
and learn something about it,
or should they show you an ad
that they already know there’s a good
chance you’re going to click on?
Over the last 60 years,
computer scientists have made
a lot of progress understanding
the explore/exploit trade-off,
and their results offer
some surprising insights.
When you’re trying to decide
what restaurant to go to,
the first question you should ask yourself
is how much longer
you’re going to be in town.
If you’re just going to be there
for a short time,
then you should exploit.
There’s no point gathering information.
Just go to a place
you already know is good.
But if you’re going to be there
for a longer time, explore.
Try something new,
because the information you get
is something that can improve
your choices in the future.
The value of information increases
the more opportunities
you’re going to have to use it.
This principle can give us insight
into the structure
of a human life as well.
Babies don’t have a reputation
for being particularly rational.
They’re always trying new things,
and you know, trying to stick them
in their mouths.
But in fact, this is exactly
what they should be doing.
They’re in the explore
phase of their lives,
and some of those things
could turn out to be delicious.
At the other end of the spectrum,
the old guy who always goes
to the same restaurant
and always eats the same thing
isn’t boring –
he’s optimal.
(Laughter)
He’s exploiting the knowledge
that he’s earned
through a lifetime’s experience.
More generally,
knowing about
the explore/exploit trade-off
can make it a little easier for you
to sort of relax and go easier on yourself
when you’re trying to make a decision.
You don’t have to go
to the best restaurant every night.
Take a chance, try something new, explore.
You might learn something.
And the information that you gain
is going to be worth more
than one pretty good dinner.
Computer science can also help
to make it easier on us
in other places at home and in the office.
If you’ve ever had
to tidy up your wardrobe,
you’ve run into a particularly
agonizing decision:
you have to decide what things
you’re going to keep
and what things you’re going to give away.
Martha Stewart turns out
to have thought very hard about this –
(Laughter)
and she has some good advice.
She says, “Ask yourself four questions:
How long have I had it?
Does it still function?
Is it a duplicate
of something that I already own?
And when was the last time
I wore it or used it?”
But there’s another group of experts
who perhaps thought
even harder about this problem,
and they would say one of these questions
is more important than the others.
Those experts?
The people who design
the memory systems of computers.
Most computers have
two kinds of memory systems:
a fast memory system,
like a set of memory chips
that has limited capacity,
because those chips are expensive,
and a slow memory system,
which is much larger.
In order for the computer to operate
as efficiently as possible,
you want to make sure
that the pieces of information
you want to access
are in the fast memory system,
so that you can get to them quickly.
Each time you access
a piece of information,
it’s loaded into the fast memory
and the computer has to decide which item
it has to remove from that memory,
because it has limited capacity.
Over the years,
computer scientists have tried
a few different strategies
for deciding what to remove
from the fast memory.
They’ve tried things like choosing
something at random
or applying what’s called
the “first-in, first-out principle,”
which means removing the item
which has been in the memory
for the longest.
But the strategy that’s most effective
focuses on the items
which have been least recently used.
This says if you’re going to decide
to remove something from memory,
you should take out the thing which was
last accessed the furthest in the past.
And there’s a certain
kind of logic to this.
If it’s been a long time since you last
accessed that piece of information,
it’s probably going to be a long time
before you’re going to need
to access it again.
Your wardrobe is just like
the computer’s memory.
You have limited capacity,
and you need to try and get in there
the things that you’re most likely to need
so that you can get to them
as quickly as possible.
Recognizing that,
maybe it’s worth applying
the least recently used principle
to organizing your wardrobe as well.
So if we go back
to Martha’s four questions,
the computer scientists
would say that of these,
the last one is the most important.
This idea of organizing things
so that the things you are most
likely to need are most accessible
can also be applied in your office.
The Japanese economist Yukio Noguchi
actually invented a filing system
that has exactly this property.
He started with a cardboard box,
and he put his documents into the box
from the left-hand side.
Each time he’d add a document,
he’d move what was in there along
and he’d add that document
to the left-hand side of the box.
And each time he accessed
a document, he’d take it out,
consult it and put it back in
on the left-hand side.
As a result, the documents would be
ordered from left to right
by how recently they had been used.
And he found he could quickly find
what he was looking for
by starting at the left-hand
side of the box
and working his way to the right.
Before you dash home
and implement this filing system –
(Laughter)
it’s worth recognizing
that you probably already have.
(Laughter)
That pile of papers on your desk …
typically maligned
as messy and disorganized,
a pile of papers is, in fact,
perfectly organized –
(Laughter)
as long as you, when you take a paper out,
put it back on the top of the pile,
then those papers are going
to be ordered from top to bottom
by how recently they were used,
and you can probably quickly find
what you’re looking for
by starting at the top of the pile.
Organizing your wardrobe or your desk
are probably not the most pressing
problems in your life.
Sometimes the problems we have to solve
are simply very, very hard.
But even in those cases,
computer science can offer some strategies
and perhaps some solace.
The best algorithms are about doing
what makes the most sense
in the least amount of time.
When computers face hard problems,
they deal with them by making them
into simpler problems –
by making use of randomness,
by removing constraints
or by allowing approximations.
Solving those simpler problems
can give you insight
into the harder problems,
and sometimes produces
pretty good solutions in their own right.
Knowing all of this has helped me
to relax when I have to make decisions.
You could take the 37 percent rule
for finding a home as an example.
There’s no way that you can
consider all of the options,
so you have to take a chance.
And even if you follow
the optimal strategy,
you’re not guaranteed a perfect outcome.
If you follow the 37 percent rule,
the probability that you find
the very best place is –
funnily enough …
(Laughter)
37 percent.
You fail most of the time.
But that’s the best that you can do.
Ultimately, computer science
can help to make us more forgiving
of our own limitations.
You can’t control outcomes,
just processes.
And as long as you’ve used
the best process,
you’ve done the best that you can.
Sometimes those best processes
involve taking a chance –
not considering all of your options,
or being willing to settle
for a pretty good solution.
These aren’t the concessions
that we make when we can’t be rational –
they’re what being rational means.
Thank you.
(Applause)