The Dominoes Fallacy
The Difference Between Complexity and Difficulty
I’ve set a goal this year of running a half marathon.
Really, running a half marathon is simple. In fact, running an ultra-marathon is simple.
Two steps:
- Run.
- When you hit your goal, stop.
But running an ultra-marathon, or any significant distance for that matter, is not easy.
I believe we fundamentally misunderstand the difference between complexity and difficulty. In many facets of the world, we perceive complexity and difficulty as the same thing, but really they are not.
Look at making a beautiful steak dinner. Let’s say, for argument’s sake, you want to make a 2-Michelin-star steak dinner. You’re thinking garlic mashed potatoes, carrots, asparagus, steak, red wine reduction, and creme brulee for dessert. You need some basic ingredients - steak, good red wine, potatoes, butter, garlic, carrots, asparagus, lemon, salt, eggs, sugar, vanilla, and so on.
The point is that any one of the steps to making a 2-Michelin-Star dinner is actually quite simple:
Boil Potatoes. Nailed it.
Mash potatoes. Nailed it again. On a roll here.
Microwave garlic and butter together, add to potatoes. I’m a superstar. Call me Chef Ramsay from now on.
Each step is like this - none of them, on their own, are all that difficult. The reason we associate difficulty with completing this task is because we get overwhelmed with the complexity of the issue.
I found this everyday in my engineering degree. Solve for the total combined stress in the beam. Well, crap. That’s really hard, isn’t it?
When actually, it’s really not.
Draw a Force Body Diagram. Label the stuff you have, and the stuff you don’t. Figure out what equations apply to this scenario. Figure out if you can use one of those to find the stuff you don’t have. Plug in the stuff you have into the equations to find the stuff you don’t. Repeat as needed.
Overall, the algorithm is quite simple, but it is certainly complex.
My final exam in university was Robotic Manipulators. We had to calculate the forces required, output speeds, and fully design a control system to move a robotic arm from one position, to performing an action in another position - by hand. That’s not a simple task. But once you know the process, each step is rather trivial: multiply these things together, add these ones up, and solve for this variable.
Not a single operation in the entire class was harder than grade 11 math.
That’s beautiful, isn’t it?
But why do we think that robots are so difficult?
Because of the complexity.
I tend to think of complexity as “number of steps”.
Making a 2-Michelin-Star dinner is more complex than making PB&J or Tomato Soup, but fundamentally, none of the steps are any more challenging. It isn’t harder, just more complicated.
If you start to dissociate complexity with difficulty, you will start to realize that the vast majority of things we view as really challenging in the world are in fact quite straight-forward - there are just a lot of steps.
Here’s another way to think about it.
When I learned about Fatigue Life in Solid Mechanics (how long it takes something to fail over time and repetition) we used this thing called a “stress factor”. It was symbolized by the greek letter Psi. It was calculated by the affectionately named “k-factors”. They were labelled as follows:
Psi = ka*kb*kc*kd*ke*kf
Each k-factor represented its only area of significance - ka is temperature, kb is material, kc is the kind of notch in the system around where your stresses build up, and so on.
It’s simple to calculate each k-factor, and then you just multiply them all together. If something isn’t relevant to the calculation, it is gets a 1, so as not to change anything (ex. at room temperature).
Most people think of complexity like this:
Complexity = Multiplication of the difficulty of all of the steps
Say that 1 is the difficulty of something very simple, like putting on a pair of pants. No one bats an eyelash about how hard it is to put pants on - unless you’re a university student, but we’ll ignore that for now.
So if you’re making your bed:
Complexity = 1.00001*1.00001*1.00001*1.00001
There may only be 4 steps, and each one is easy.
But if you’re running a marathon:
Complexity = 1000
There’s only really 1 step — run — but it is far from easy.
But I fundamentally think this is the wrong way to approach this problem.
I refer to this as the “Dominoes” Fallacy - Dominoes, because there are many moving parts, but each individual step is quite simple. When you arrange an enormous array of Dominoes, setting up any individual Domino is not a hard task, but it certainly is a complex endeavor. These are fundamentally different things, however, and we cannot think of them as being the same.
I think about it like this:
Difficulty = How hard is each individual task on its own
Complexity = How many tasks are there
Under this framework, there are 4 possible scenarios, which I’ve given examples for below:
Simple and Easy: Making your bed
Simple and Hard: Running a marathon
Complex and Easy: Making a perfect steak dinner
Complex and Hard: Designing a space shuttle
Most people believe that you tackle difficult and complex problems the same way, but you really can’t treat them as the same entity at all - you must deal with them differently.
Running a marathon is a hard task, but it isn’t complex. There really isn’t any strategy or planning required - it is all about execution in the moment.
To tackle hard problems, we must dip into the part of our brain that deals with mental grit, creativity, and innovation.
Making a 2-Michelin-star dinner isn’t hard, but it is complex. There requires a great deal of strategy and planning, and execution in the moment is less important. If you can plan out your steps methodically, and continually perform them correctly, accomplishing a complex task is no sweat at all. It is really just accomplishing a long series of simple tasks.
Peter Thiel in Zero to One talks about the importance of both humans and computers, and how we are fundamentally different systems. Humans are amazing at solving hard problems - we can think abstractly, be innovative, and make connections where there were previously none. Computers are amazing at solving complex problems - they can do routine, simple steps, in remarkably high volumes, with amazing accuracy and precision.
When you face a problem in life, business, relationships or otherwise - ask yourself - is this hard, or complex?
Complex tends to lend itself to automation, where hard lends itself to ingenuity. Complex problems can be solved if they can be ordered (Computer Science problem) but hard problems can be solved if they can be untangled (Mathematics problem).
It is at the intersection of these two concepts - the hard complex problems - that the true magic happens. We call this innovation. Peter Thiel founded Palantir which specializes in large-scale data analytics based on this principle - using computers for complex challenges, and humans for hard ones. They just raised $800 million.
The next time you run into a problem, ask yourself these questions:
Is this hard, or complex?
If it is hard, what about it makes it hard? If it is complex, what about it makes it complex?
Can I find a solution to the difficulty? Can I automate the complexity?
Do I need both?
This is a good framework and baseline to start with. I’ve used this exact technique to automate a huge majority of things in my life, such as setting up systems to take vitamins, dealing with fitness and dietary goals, dealing with exam questions at school, and designing complex interacting systems in my career.
I even aced my Robotics course.
As for the ultra-marathon? Well I’m working on it.
Hat tip: Peter Thiel (Zero to One), Derek Sivers (Tim Ferriss Podcast)
