Un-Wrong Yourself

Reading “The Undoing Project” by Michael Lewis. It was on my Christmas Gift List.

It’s a book about how two Nobel Prize winning Israeli psychologists, Tversky and Kahneman, colaborated in thought provoking, scientific based research on how the human mind works. Their story and work is full of gems of knowledge.

If Michael Lewis was a wine cellar, this book would be a really fine wine.

Schermafbeelding 2017-01-04 om 12.35.16.png

Accidently, the Freakonomics Podcast had an episode with Michael Lewis about “The Undoing Project”.



one of the pet areas of Daniel Kahneman is “decision making”: how can we make better decisions?

I’m reading the part about how Tversky and Kahneman produced their first paper “Belief in the Law of Small Numbers”.


And this got me thinking.

What is the best way to a approach a creative idea?


FIRST STEP is to assume one thing and one thing only:



My colleague (A) came up with this story idea. He thinks it’s a great idea. But obviously he’s wrong about that.

Another colleague (B) looked at the story idea by my first colleague (A) and found it to be total crap. He hated it. But, the thing is: he’s wrong also. (B) is almost always wrong.

Meaning: the probability that they are wrong is equally big. They are almost always wrong. Everybody is. Even the experts. Nobody knows nothing. Certainly in TV and Film.

It’s sort of like Schrödinger’s Cat. This particular story idea is equally great and equally crap. No matter what anyone thinks about it.

When you create something or you are confronted with a new idea, you only do yourself a service by realizing that you are most likely dead wrong about it.

Don’t trust yourself, because you are untrustworthy. You are really crap in knowing either way.

So, the question becomes:

How can you un-wrong yourself?

How can you shift the odds in knowing if the idea is crap or great?


The answer to that becomes the SECOND STEP :


Split your idea in tiny parts. In ATTRIBUTES and in COMPONENTS. What is it about this idea? What does it contain? What does it hold? What is necessary for it to work? What is assumed? How will it work? etc.

By sequencing you still don’t know nothing. It’s just laying down the pipes.

You might form an opinion about each tiny part. But that will not help you. You won’t come very far because you are almost always wrong, remember?


STEP 3 would then be:


I mean, really know it. Become an expert in why things suck.

This has an echo of  the Charlie Munger advice to always invert:

“All I Want To Know Is Where I’m Going To Die So I’ll Never Go There.”

You really need to be ably to distinguish when ideas are certified dead meat.

But to become an expert in such an area, you will first need a collection of so called AGENTS. Kahneman talks about it in his Google Talk. How people are really bad at abstract ideas, but are great in Agents. It’s around minute 00:23:25.

You need to acquire a wonderful collection: a very practical, unique set of rules (or algorithms) where you are a 100% guaranteed that when you use one of those rules, the result will definitely suck. When applied it will be a 100% sure failure.

I suggest you turn those rules or algorithms in tactile AGENTS. You represent one rule by one object. And you have those objects placed in front of you.

So now you can pick up such Agent and ask yourself: does this rule apply to this idea?

Does it or doesn’t it? And you go on to the next one. You checklist every agent-of-suck.

When one of the Agents that makes thing suck is relevant to the idea or one component or one attribute: tell the creator of the idea to go back to the drawing board. He or she needs to stop wasting your time. Instead s/he needs to:


Yet, when none of the Agents that make things suck are relevant to the idea presented… then the odds are shifted already, somehow. You might be on the right path.

Now, it can be that you are not so good at judging yourself as an “expert of suck”. So you might not know really well why things suck. This actual idea might suck really bad, only you don’t know it.

Because, you know: You. Are. Almost. Always. WRONG.

Nah, you better can go on to the next step.


STEP FOUR would be :


Take every little attribute or component and come up with at least 5 better solutions for the function of that attribute or component.

Know what it needs to do. Name that function. Make it tactile. Turn that function into an Agent. And then… come up with something better.

For example, the TITLE. What is it the title needs to do? What are the functions of a title? Now, find a better solution for the one you have.

But how would you know if it is any better?

Turns out we are pretty great in competition between two small things.

For example, it might be hard to know which actress is any better: Mila Kunis or Scarlett Johanson? But we can judge fairly well who of those two has the most beautiful eyes. Or who has the best range of accents. Or who has the best comical timing. Etc.

When it’s small and we compete two thing against each other, then we might get it more right than wrong.


Now, Tversky observed that this is easy when you have a choice between 2. But not when you have a choice between 3 or more.

In the book is described how people might like “coffee” over “tea” and “tea” over “chocolate”, but then again “chocolate” over “coffee”. There is no real hierarchy, no transitivity in choices like that, because the choice is too complex. People think different about “chocolate” when placed against “tea” versus when it’s placed against “coffee”. They imagine different traits.

Kahneman is absolutely against trusting your gut in complex matters.



So, you must keep the area of choice as small and clear as possible, for intuition to have merit. Even better is that you frame the choice as the choice by somebody else, but yourself. You pick an outsiders view.

And you must delay judgement until the end.

For example, you have a few working titles. First, you go over all of them.

You ask yourself: which one is the best summary of the story for somebody who doesn’t know the story? Everybody gives a score of 1 to 5. You keep the score hidden.

Then, you stack them against each other. You have on title in competition with the other one. Which one you like best, in relation to this one focused question “which one is the best summary of the story, for somebody who doesn’t know the story”?

You put the winner on the side, the looser on another side.  Now, you pick the next working titles. And ask the same question. Put the winner on the side, the looser at the other side. And so forth.

The looser is not done, though. You keep the “loosing” titles against all winning titles. Would you switch one?

Then you rank them. Grade the winners. Until you have an overall winner.

Then it’s on to the next question. You gather all working titles from the start. You ask: which one is the most engaging?

And so forth. This way you put every working title in competition for every function.


It’s a bit of a hassle, I know. It’s not instant. It’s not fast. There’s work and expertise involved.

There’s a process. 

I ain’t saying this is the ultimate process. It’s just what came to me when reading about Kahneman and Tversky. As a possible process. An improvement to what is usually out there.

But then again: I’m almost always wrong. So…





Geef een reactie

Vul je gegevens in of klik op een icoon om in te loggen.

WordPress.com logo

Je reageert onder je WordPress.com account. Log uit /  Bijwerken )

Google photo

Je reageert onder je Google account. Log uit /  Bijwerken )


Je reageert onder je Twitter account. Log uit /  Bijwerken )

Facebook foto

Je reageert onder je Facebook account. Log uit /  Bijwerken )

Verbinden met %s