How To Explain Things Real Good

Some people just want to watch the world learn.

You're here, because you want to help others learn. I'm here, because Hari thinks I'm good at helping others learn. We'll see how quickly he regrets inviting me.

So here's the deal: I ramble for 15 minutes about how to explain complex ideas, do a quick Q&A, then, because this is a small class, I might be able to do some fast 1-on-1 feedback on your project ideas. Also I'll give Hari the slides to give to you later. Sound good? Let's go!

Today I'll to try to explain how I explain. But first:

Some Theory about why I'm not giving you Some Theory

Explaining stuff is a skill. When training skills, people usually think about Theory versus Practice. Theory is easy to share, but hard to apply to day-to-day work. Practice is easy to apply, but hard to share –– for example, I can't talk you into being able to ride a bike.

But I think there's something in the middle: Tricks! (or Heuristics, Rules-of-Thumb, whatever you call it.) Tricks are easy to share, but also easy to apply!

But, of course, there's a catch: the vast majority of tricks are false. Like, "Beer before liquor, never been sicker. Liquor before beer, you're in the clear." That's... false? Not true. It does not work.

Y'know, I only recently realized you can actually test advice? Like all products, advice can be methodically tested, modified, and iterated upon. The idea that we should test our ideas does not come naturally to human beings.

So, before I share my favorite Tricks for explaining stuff, I want to stress that actually tested & iterated upon these tricks. Of the hundreds of tricks I've heard, I've thrown away the ones that don't work, and the ones that technically work but are too easy to mess up. After 8 years of making explorable explanations, there's only a few tricks that have survived the test of time.

So without further blabber, here's...

Nicky's Top 5 Tricks On How To Explain Things Real Good

The first step to helping someone learn, is to help them want to learn. So:

Trick #1: Show what made you care.

Most textbooks are fact-dumps, and seem to not care about helping you care. But sometimes, it's even worse when they do try to help you care. Here's some other Tricks I've been told for making one's audience care, that can work, but are too easy to mess up:

"Make it interesting" -> becomes, "How do you do, fellow kids? Math is COOL!"

"Make it relevant" -> becomes, "Math is important, so you can calculate tips at restaurants!"

"Create a curiosity gap" -> becomes clickbait. I should clarify my "Top 5 Tricks" title was ironic.

The problem is that these Tricks make you second-guess your audience. Solution: don't second guess what would make your audience care about some topic. Just show them what made you care about that topic.

An interesting puzzle? A clever creation? A personal story? Show me what made you care.

Important: I said, "show WHAT made you care", not "tell WHY you care." If you try to report WHY you care, it can be way too abstract.

For example, if you asked me, "why do you care about group theory?" I might say, "Group theory unifies several branches of mathematics, and has applications in quantum field theory." Which is true, but to a total novice, that's abstract gobbledygook.

It's better to ask me, what made you care about group theory?" Only years after I learnt about group theory, did I finally care about it, thanks to this puzzle:

Why do mirrors flip left-and-right, but not up-and-down?

...which doesn't even sound like a math question! Yet it is, and its solution is a great introduction to group theory. (Don't ask me to explain mirrors right now, I only have 15 minutes.)

The point is, to avoid second-guessing & being too abstract: Don't tell me why someone might care, show me what made you care.

But, speaking of showing & telling... once you've helped your audience want to learn, how do you deliver? Well:

Trick #2: Show, then tell.

"Show don't tell" is good for fiction, not so much for non-fiction. But fact-dumping educational material does the opposite: "tell, don't show." Look, I love Wikipedia, I give Jimmy my monies each month, but... well, here's their introduction to a "symmetry group":

In group theory, the symmetry group of a geometric object is the group of all transformations under which the object is invariant.

// gif

Obviously, it's better to Show and Tell. But it's still easy to accidentally mess up, if you show & tell in the wrong order.

For example, I recently read this classic masterpiece of math exposition, titled Gödel's Proof. It's fantastic! But when they explained what a "formal mathematical system" was, they started by telling me a page of technical definitions about theorems & axioms, and I was totally lost.

Only on the next page did they show me an analogy: chess. Chess pieces are like math symbols. An arrangement of pieces is like an arrangement of symbols, called Theorems. The rules on how to move from one arrangement to another, are like the rules for how to prove one theorem from another. The starting arrangement in chess, are like the starting Theorems in math, called Axioms.

And I was like... "that is the most beautiful explanation I've ever heard of pure mathematical logic."

"...Why didn't you start with that?!"

So: don't just show and tell, show then tell.

Concrete, then abstract.

Familiar, then unfamiliar.

How do we do this? Easy! Just remember the acronym:

PEA! With an A.

Pictures, Examples, Analogy.

Show me a picture like this, then tell me, "a fractal is a shape that contains a smaller copy of itself."

Show me the example of how a square (a "geometric object") can be flipped and rotated ("transformation") yet its square shape doesn't vary ("invariant")... then tell me, "the symmetry group of a geometric object is the group of all transformations under which the object is invariant".

Show me a chess analogy for pure mathematics, then tell me the technical definitions.

Show, then tell.

So that's how you explain one idea. But – hear me out, I know this sounds crazy – what if you want to explain more than one idea?

Well, do not do a listicle. Again, my "Top 5 Tricks" was ironic. Don't give me a disconnected listicle, give me a connected story.

How? With this:

Trick #3: Therefore & But, not And Then

Imagine you have a story that goes:

Have you noticed a lot of recent Hollywood movies feel like that? A series of grand set-pieces that should be emotionally moving on paper, and yet... it feels hollow. I reckon that's because each event is connected with an "AND THEN", not a "therefore" or "but". If events aren't consequences of each other, they won't feel consequential.

Contrast that to this story:

That is the kind of modern fable that'll be carved into our culture, like "cry wolf" or "sour grapes". Why? Because it shows the therefore/but loop of every great story, of life itself:

There's a conflict, therefore a character changes something. (Their actions, beliefs, deepest values) But this causes a new conflict. Therefore they change again. But new conflict. Therefore new change. Repeat. In books & in life, that's how you get real character growth.

This "Therefore & But" trick was coined by the creators of South Park, as advice for writing fiction. But a few years ago it was re-popularized by Tony Zhou, creator of the video essay series Every Frame A Painting, because it also works for writing non-fiction.

So here's an example of using Therefore & But for non-fiction: explaining how a piano key works.

Here's a GIF of the inside of a piano key:

// gif

...why. Why is that that complicated.

Alright. Does this help?

// pic

No offense to whoever made those, they clearly put a lot of skilled effort into it. But it's a very And Then explanation. Here's the hammer shank and then here's the capstan and then here's the damper lever, etc.

Contrast that to a recent video from engineer Mark Rober, who used a Therefore/But loop to explain why the piano key is designed the way it is:

Aaaand some other stuff I don't remember. Point is, that's the core therefore/but loop of how any complex idea is designed, whether in engineering, science, philosophy... anything:

You have a problem (or puzzle, or question). Therefore you try a solution (or hypothesis). But this gives rise to a new problem. Therefore you try a new solution. But new problem. Therefore new solution. Repeat. In science & in life, that's how we get real intellectual growth.

So the next time you explain a complex idea, don't give me a disconnected listicle ("And Then"), give me a connected story ("Therefore & But").

As we saw with the humble piano key: you don't need to make up a story.

The story's already there.

So, our first three Tricks – show what made you care, show then tell, therefore & but – these will help make your explanation more beneficial for your audience.



It's not enough to think about benefits, you also need to think about costs. Specifically: the cost of time. Our time being alive is precious, our ultimate non-renewable resource.


We need to respect our audience's time, and be concise. So:

Trick #4: Write a draft, then cut 10%

No matter how good you get at writing, at least 10% of your first draft will always be fluff. Redundant words... non-sequiturs... redundant words... y'know, fluff. Find that fluff, and cut it. So:

  1. Write a first draft, then get a word count.
  2. Multiply that number by 0.9, to get 90% of the word count.
  3. Cut your draft down to meet this new word count.

There is so, so much more I want to tell you about being concise...

...and, I won't.


Trick #5: Do REAL tests, early & often.

In the 4th century BC, Aristotle claimed that heavier objects fall faster. Everyone thought so. It was just common sense.

In the 16th century AD, scientists finally tested the idea. It's false. Two lead balls of different sizes & weights, dropped from the top of a building, will fall with equal acceleration. (As an aside, Galileo himself probably never did this experiment, but we know other scientists did during that time.)

4th century BC. 16th century AD. That's about two thousand years.

Two. Thousand. Years.

That's how long it took humanity to discover something that you could've discovered, with two balls and ten seconds.

Like I mentioned at the start, the idea that we should test our ideas does not come naturally to human beings. It doesn't happen enough in politics. Heck, given science's replication crisis, it doesn't even happen enough in science. But it's especially hard to test our ideas in the creative arts – such as making an explainer article or video. It's hard to put your own creative darlings to a trial by fire.

I'll admit, Trick #5 is still the hardest for me. But I'm trying to get better. For a concrete example of what I mean by, "do real tests, early & often", here's the story of how I tested and iterated upon... this talk!

A few days ago, I met two friends for dinner, and I wanted to test this talk with them. Usually, I only test explainers after I've written & edited my first draft – but this time, I had nothing but an outline. I improvised my entire talk, using this sketchbook to draw diagrams & record feedback.

That's testing early & often.

But I went further, and told my friends: PLEASE HECKLE ME. Because, when I usually test my explainers, I only ask afterwards what parts felt boring, surprising, confusing, insightful. Which is better than nothing, but it means I'm selecting for the feeling of learning, not necessarily actual learning. So I told my friends to heckle me: As soon as something I'm saying is redundant, or makes no sense, interrupt and tell me. I also asked them to summarize what I said & try applying the tricks to their own work, to test for recall & transfer –– actual learning, not just the feeling of it. (But, I reassured them that if they can't remember what I said, that's not their fault as a listener, that's my fault as a communicator.)

That's doing a real test.

So the next time you test an explainer, don't just share it for feedback after you edit the first draft. Get feedback when all you have is an outline and sketchbook, and encourage your friends to heckle you.

In science, politics, life, and art – we should do real tests, early and often.

Because when humanity forgets the idea that we should test our ideas, apparently, we deceive ourselves for 2000 years.

In Conclusion

Those are the 5 Tricks I keep coming back to, that have stood the test of time for me:

  1. Show what made you care
  2. Show, then tell (with PEA – Pictures, Examples, Analogy)
  3. Therefore & But (and the Problem-Solution story loop)
  4. Write a draft, then cut 10%
  5. Do real tests, early & often

So, dear listener, here's our story loop:

Problem: there's a beautiful, important, complex idea you want others to know.


Solution: you write an outline of an explainer – what made you care, concrete examples, therefore & buts – and improvise a talk for friends.


Problem: your friends like it, but they were bored by some parts, and they misinterpreted other parts – the feeling of learning, not actual learning.


Solution: you take in their feedback, write the first draft, and cut it by 10%. Then, you send it out for more feedback.


Your test audience finds more problems.


You iterate on it with more solutions.

REPEAT until your explainer has gone from good to great. You hit "publish", and show it to the world.

THE END. (for now)

That's the story-loop of some people, people like you and me.

Some people, who just want to watch the world learn.

Thank you so much!