Nona Blog

Superforecasting — why you should read it!

I used to be all about them business books. I was a massive believer that the more you read the better you become — reading a book or two a month. Unfortunately I was retaining next to nothing.

I changed my strategy to rather understand and retain as opposed to focusing on quantity. This means I need to be very specific as to what I read. There is less time to waste on books that could be blog posts and most business books I’ve come across fall into this category.

This focused approach lead me to Superforecasting

I was first introduced to this book in The Knowledge Project’s interview with Phillip Tetlock, the co-author of the book. It ignited my curiosity and I put it on my reading list, bought the book and left it lying around for months.

Finally, when I I started reading, I couldn’t tear myself away. I read a chapter every other day, finishing it quite quickly and thoroughly enjoying it.

Superforecasting book cover


I highly recommend you give it a look if you want some revision on human psychology and how to become better at forecasting, and thinking in general.

1. Doubt is the key

Sadghuru said something that completely changed my life: 

“Conclusion is death”. 

Why was this so powerful ? Because I was always taught to be confident and know what I want. Leaving no room for doubt. 

However, this idea of being 100% sure of something closes the door to any other possibility. This is when you get sucker punched.

Since you’re human, you will be wrong at some point. We all are — deal with it. When you’re wrong, and don’t see it coming, it will hit you like a wrecking ball. Not the Miley Cyrus sing-a-long-song kind. The sledge-hammer-to-the-face kind.

Doubt on the other hand allows for error. It says: “I think something works like this but I’m not certain”. This little bit of humility is the difference between being destroyed by the wrecking ball or seeing something ominous coming and dodging it at the last second.

This means you constantly adjust based on newer information. For example if I forecast that my wife won’t be pregnant six months from now  but three months into it she has morning sickness, should I stick to my initial forecast or adjust?

In a world that despises doubt — it may be worth rethinking our feelings toward it.


2. Small increments make for solid progress

I’m sure you’ve all heard that it’s the little things that make the difference. I know, I’ve heard it many times but I always forget to apply it. 

When things seem overwhelming, take a step back and do the smallest thing you can to improve the situation. 

Make a forecast based on what you know right now and iterate around that, updating frequently.

3. Ask the right questions 

Often a problem can be solved by simply rephrasing the question or breaking it into smaller solvable problems.

  • What information would be helpful in answering this question?
  • What information is actually irrelevant and can be ignored?
  • Can this question be inverted to make it easier to solve?

At varsity we were taught Fermi’s method of deduction where you use simple mind experiments to answer seemingly impossible questions.

Here’s a great short video explaining the process.

Another great way to solve problems is to invert them. A method Charlie Munger is famous for promoting and is well explained here but here’s a small example:

I want to lose weight? 


What am I doing right now that makes me overweight? 

This is critical when forecasting. You have little to go on but need to make an accurate estimate. Take the question. Invert it and/or break it down into small solvable chunks, you might surprise yourself with the accuracy you start to achieve.

4. Follow a process

What is the best way to lose weight?

A diet you say.

What kind of diet?

The one you can stick to.

Forecasting and most things in general become easier when you follow a process. It reduces the cognitive load needed when approaching a similar problem. 

E.g. I don’t want to eat bad food. Therefore I don’t buy bad food. This means I need to exert a lot of effort to either leave the house or make a decision of ordering a delivery to get the bad food. 

This is an example of a process. 

When it comes to forecasting a process that could work is the following:

1. Read up on the topic you are forecasting
2. Imagine why your forecast is wrong
3. Make as precise a forecast as you can e.g. 53% likely 
4. Document why you made that forecast (so you can improve later when revising)
5. Regularly update your forecast based on new information

5. Keep score

My mom loves to tell me whenever I come to her with some bad news that she had a feeling something bad was going to happen. 

This is utterly ridiculous since she always has a bad feeling. She then uses the 1 in 1000 times it actually manifests to rationalise to herself that she can feel bad events coming. Clearly not scientific.  

If the odds of something happening are 1/1000 this is a pretty unlikely event and can de discarded as negligible. 

The only reason my mom doesn’t see it that way is because she doesn’t keep score. She allows availability bias to cloud the actual numbers. 

If my mom were to look at the results — if she kept score — she would realize that her feelings are horribly inaccurate. In all likelihood these feelings should be avoided when deciding if something bad is going to happen.

And that’s a wrap

These were the five things I took away from Superforecasting:

  1. Doubt is the key
  2. Small increments make for a sturdy whole
  3. Ask the right questions
  4. Follow a process
  5. Keep score 

I highly recommend giving it a read!

Nona designs and builds intuitive software for FinTech businesses. If you’d like to accelerate your FinTech project, book a consultation with us!

Dominic Bauer

Full Stack Developer - Nona