How to think about luck

You get in a fight, someone swings at you and you duck under, throw a punch and knock the person out.

If you are a trained martial artist, it wasn’t luck.
If you’ve never trained anything, it was luck.

The same outcome, but two different explanations. Why?

If you rewinded history a 1000 times and got the same configuration of atoms, you would get the same result 1000 times (physics is deterministic).

But if you placed this same person in a 1000 fights, in different cities, or just on different days, or different opponents, or different type of terrain, or different anything – you would expect this person not to easily slip a punch and knock the opponent out. In fact, you would give a 1/1000 chance (or similar small number) for this happening. However, you would give, say, 800/1000 to a trained martial artist.

We call a positive, successful event with a small probability luck.

By the very definition of luck, everybody who has ever done something extraordinarily unexpected has had luck.

Let’s take Bill Gates for an example. It’s 1955, and you place a 1000 Bill Gates around the world (at age 0). Would you expect a 1000 multi-billionaires at age 60? By definition, no, because in 1955, there was surely much more than a 1000 individuals with sufficient Bill Gates-ness, but we only see one emerge as a multi-billionaire. Hence, our probability estimate is more correct than not: we would expect 0 or a very small number of such individuals – but definitely not 1000 or more.

And this is just 1000 persons, and it’s still 1955. Imagine if you removed all restrictions, temporal, spatial, whatever. If you put a Bill Gates in 200 B.C., you have no guarantee that he would rise to a prominent position equivalent to a billionaire today. He might have greater chances simply because he is smart, but that’s all you have, a slightly better chance which you can’t quantify.

Another way to think about it: right at this moment, there are probably millions of people around the world who are: high IQ, high conscientiousness, high creativity, fast learners with integrity and drive. But you don’t see millions and millions of people with Bill Gates-like success. You see only a handful – hundreds and thousands, despite the fact that there aren’t any significant gaps in the attributes of Bill Gates and these millions of super-talented people.

That is what is meant by luck – your internal attributes play a large role, but your success is not independent of your environment.

However, this goes against the example in the beginning – if you put a trained martial artist anywhere and anytime, you would give a high chance of seeing a scenario where the martial artist wins easily. Sometimes no, sometimes circumstances interfere, but you would mostly expect the martial artist to emerge victorious. What is the difference here?

As far as I can tell, training to fight and winning a fight are very closely related – if you do everything right, you will probably (but not definitely) win. We clearly see the link between the move repeated in the training room and the same move performed during a confrontation. However, certain other areas of human life, including building a multi-billion company, do not have clearly visible correlations between preparation and outcome. It’s obvious that you need the preparation as it gives you a boost in your probability to ‘win’. However, it’s muddled, and we conclude that the role of uncertainty and luck is greater here.

I am not yet sure if this conclusion is correct or not. But in practical terms, looking at luck from this perspective can be detrimental since it acts as a self-fulfilling prophecy. “Mathematically, I have very small chances of building wealth, so I will not try.” But this conclusion gives you exactly zero chances to reach wealth (maybe more, maybe you get struck by blind luck, but the chances are nearer 0 than 1).

In practice, building an accurate model of the world and then executing the best reasonable strategy of building yourself is the way to go. You want to max the attributes which allow you to better navigate the world, and then try to use them. You still might not reach your goal, whatever it may be. Bad luck still may happen. But you will definitely get a better chance, and that’s all you can hope for anyway.


Practice everything

” You practice gymnastics, I practice everything!”

Mindfulness is the most important meta-skill.

If you are mindful all the time, you perform every action with care, and find ways to improve something – either the action itself, or something else that is not related.

When you go up stairs, you are mindful of every step – what a great opportunity to practice walking silently!

When you go to throw out the trash, what a great opportunity for graceful movement, wielding the bag of trash like a kettlebell.

When you walk around the neighborhood and observe a bar or a ledge, what a great opportunity to perform several pull-ups and increase your strength!

When you drive to work, what a great opportunity to glide through traffic with minimal required effort! (Or perhaps to listen to an audiobook and increase your knowledge about the world)

Being mindful about everything will lead to practice, and then to improvement. Being mindful is not all that is required for progress – several pull-ups here and there will not lead to great increases in physical strength, as a good program would, but being mindful will present opportunities you might otherwise miss. Being mindful will also lead to different opportunities for different people. Where one may find an opportunity to practice an athletic movement, another may find something else entirely. The principle, however, is the same: keeping your eyes and ears open will open doors and present possibilities. It will be up to you to use or not use these possibilities.

I don’t want to be the best in the world

This is a text about my preferences. I don’t offer any facts, so if you don’t like reading pure opinions, skip this one.

I am not interested in becoming the very best in anything right now. The reason is simple: becoming so good that you’re in the top 20% takes quite a lot of effort for most things, but it’s pretty doable still. But getting to top 10%, top 5%, top 1% – that just takes a disproportionate amount of time and effort – or optimization power.

Let’s illustrate with an example: if you’re reasonably athletic and learn reasonably well (i.e. you’re not a prodigy), it would take you between 5 and 10 years of consistent practice (3-5 times a week) to become a black belt in Brazilian jiu-jitsu. This seems like a real good trade-off: do your 10 year stint, train hard and consistent, become a black belt.

But becoming one of the very best black belts in the world, getting to the level of, say, Gordon Ryan, that takes a crazy amount of optimization invested. Here’s a graph to show what I mean:

So, the relationship is as we would think – the more optimization power you invest into a skill, the more the skill grows. The more hours you spend training BJJ, the better you are at it. But it’s not that simple – up to a point, the more time you spend learning, the more you will get out of it. But when you get fairly good at something, every following improvement will require more and more time. At one point, you get to something like diminishing returns – a lot of time invested, for very small improvements in skill.

And I, personally, am not interested in these small improvements. I want to get decent, or maybe even very good for some things, but after that, I’m not interested in chasing the 0.01% improvements. This is the part of the curve where I stop being interested:

Why? Well, I want to do more than one thing in life, but I want to become good at things. I don’t want to be someone who knows a bit of everything, but can’t really do anything. “But you should pick a field and get good at it!” But picking one thing, one field, one activity, one sport – that’s like picking a pre-made class in an RPG game.

And yeah, if you’re really passionate about one thing, you go for it. Me? I always pick a custom class, and pick and choose the things that I’ll get good at. The result is good, both in games and, hopefully, in real life. Here it is, in one tweet:

So, I guess I do want to become the best in the world. Just not in one of the default classes. I want to become the best in the world in being me.


The year is coming to an end and it’s time to reflect on mistakes made, successes achieved and lessons learned.

Grabbing an opportunity beats planning. The world is far too chaotic for complex plots, and we have limited resources to invest. This doesn’t mean that you shouldn’t plan at all – only that the universe will sometimes throw an opportunity at you. This opportunity will require that you drop most or all of your stuff for some period of time – but if you manage to do it, it will pay off. Lesson for 2019: be more mindful of hidden opportunities, and grab them decisively if they present themselves.

Even being rash is better than being indecisive. Lady Luck favours the bold. An action bias is bad in politics and in organizations, but I have a strong feeling that it’s quite beneficial for individuals. Lesson for 2019: when you see a possibility, take it! Bravery is the prerequisite for doing good in the world.

When you say yes to one thing, you say no to another. Spreading yourself too thin is a great way to not excel at things. Novelty is great, but just adding without subtracting means that you literally don’t have enough hours in the day to do productive stuff. Choose carefully where you spend your time. Lesson for 2019: relax – do the one thing first, and then do the second thing, and so on. You cannot do it all at the same time, so don’t try. Get one thing done, then move on to the next.

Chronic use of social media is a cause of bad feelings. At least for me – if I spend too much time on Facebook, I feel bad. Not because I feel unworthy, or because I feel like other people are leading more interesting lives than I do. I literally feel bad at a physical level, in my body. I cannot keep my focus on things and I’m always sorta anxious and itching to check out my news feed. I don’t know enough about biology, but I think it has something to do with disrupted dopamine in my body. Lesson for 2019: be very sparing in your use of social media. It’s a thin line to walk – if unsure, better to block it all than to keep it.

Knowing your OCEAN stats is like knowing your main attributes in an RPG. Openness to Experience, Conscientiousness, Extraversion, Agreeableness, Neuroticism. If you think about these things in more ‘mechanical’ terms, you can design your life around them. Very extroverted? Why aren’t you going out with friends? Low agreeableness? Why don’t you compete more? Lesson for 2019: Go out more. Do more competitive stuff. Work hard on your goals. Enjoy books and music. Do risky things.

Shut up. Rambling is dumb and serves no one. Being concise in speech is a virtue. You can’t fuck something up if you just adopt a habit of silence, but you can easily fuck things up if you can’t keep your mouth shut. Nobody asked you for advice? Why are you giving it? Even worse, you have a history of giving someone unsolicited advice, and the person never takes it. Doesn’t matter if it’s good advice – shut up! .

Have a great 2019 everyone.

Confidently wrong


I feel like I have something to say, so I start to write. But the ideas in my head remain unformed and I have trouble putting them out. What is it that I want to say?


How do you have a good life? This is a question that many people try to answer and they all contribute to some extent, but individuals vary and there is no advice that is always valid for everyone. Or is there? For example, ‘don’t do to others what you wouldn’t want done to yourself’ is pretty good general advice. I have trouble thinking of exceptions there.


The state of reality is definitive. What I mean by that is that the atoms that compose the universe are as they are. They move according to certain physical rules and that’s about it – when you zoom out, you see cities being built, cars being driven, dogs chasing cats, and planets revolving around suns. But under the hood, it’s just a bunch of small balls that bump around.

The state of human knowledge about reality is probabilistic. Our brains cannot keep track of every single atom in the universe, and then build up a mental image of what will happen next. Instead, we have ‘categories’ of things, like ‘cars’ or ‘dogs’. These are all atoms, but these different arrangements of atoms do very different things, and we look at them from a zoomed-out perspective. When we say we know things, we mean that we assign some probability to some atoms being arranged in a particular constellation.

But since the tools we use to experience the world are imperfect, we only ever get partial data from the universe, and must model in our heads the best (the most probable) explanation of how things actually are. This is only one part of our imperfections – the very architecture of our brains, how we think about things – that also has its shortcomings, but more on that later.


Complex topics are complex because what you thought is simple and general turns out to be nuanced and particular. In a way, you take one ‘chunk’ of knowledge (‘This is X’) and then say: ‘This is X, but actually Y in this case, but keep in mind that in this other case it’s actually Z. Actually, this way of thinking doesn’t work – it would probably be better to think about X, Y and Z as instances of a general phenomenon A, which is contrasted with B.’

In some cases I find myself automatically reacting to such sentences and saying ‘you’re overcomplicating!’ And while this can be true in some cases, I think it’s my general reluctance to accept that reality is more complicated than I naively thought.

Scott Alexander is the best person that I know for taking complex topics and presenting their nuances in an accessible way. He has a way of talking about different opinions without the reader feeling like they just lost their time listening to just some opinions. And he has a way of talking about facts which are very different and detailed, but without the reader feeling lost in too much new data. And he has a way of talking about theories and presenting how these theories could very well be wrong – but without lack of confidence while presenting them.


Confidence while presenting only probabilistic estimates of how the world works is very important. We are apes; we like individuals who are confident. And this is an obvious failure mode: get a confident leader who is actually wrong and you screw up the entire tribe. Confidence shouldn’t play any part – only facts. But it’s difficult to just bypass our inborn architectures, and we must take them into account. Therefore, learning how to say ‘probably X, but I’m not totally sure’ without sounding like a weakling is important – maybe not in the rationalist community – but probably yes in the general community. And as far as I know stuff about human biology, this is important for one’s own mental health – not feeling like an underdog, feeling like you’re on the top, like you’re wanted reproductive material. Your social standing is directly related to your level of serotonin, and this directly correlates how good you feel about yourself (and life expectancy and other stuff).


This reality is complicated: in this reality, it’s not enough to be confident. If things were simpler, you could just take, say, a Conor McGregor approach to life, and be extremely confident in absolutely everything. Things would probably go your way more often than not, but you could also be exposing yourself to grave mistakes and miss opportunities to change your mind.

And if things were simpler, you could also take the opposite approach, being uncertain about things, because you know you could be wrong. And you’re right, you could be wrong! You probably are very wrong in just about everything. But:

separates the body from the mind

This reality is complicated, because it requires that you be confident while saying that you’re only partially sure about something. If you go the pure confidence route, you’ll screw up something because you will be wrong. If you go the pure uncertainty route, you’ll screw up yourself because you’ll lower your social standing and your serotonin levels.

How I learn

I’m not an expert in learning and I don’t try to brand myself as one. But I think I’m decent at learning new things, and I want to list the things that help me. Also, this very text is an example of one of the listed techniques.

+ Every part of this post will have linked resources in the end

Table of contents:

  • Habits
  • Deep work
  • Spaced repetition
  • Testing and recall
  • Don’t DiSSS me
  • A Feynman summary
  • The tribe
  • The tricks

0. Habits

👏Habits 👏are 👏the most 👏important 👏thing.

If you don’t build habits, you have to spend willpower. And willpower is a finite resource. So building a habit should be step 0 – the framework you use to start learning. Some specifics:

  • Trying to change more than 1-3 habits at a time is very difficult; better to change only one behavior and then move to the next one
  • Using D͏͇͓̹̠̟̯ͅ ̙̖̬̟̓ͧ̓̈̓̌̈͠Ȋ̧̦͆̂ ͕̪͉̭̫͆ͬ̈́̀S̘ͮ̈̇͠ ̻͔͙̣̝̥͛̏ͪ̒͌͘C̣͕̗ ̬̼̬͉̥̖̤͊̊ͤ̀̓̅ͤI̝̽̽̊ͬ̑ ̦͓̪̩P̯̯̱ͯ͒̄͆͛͂͗ ̹̬̩̥̝̖͍̅L̤͇͖̪̬̋͌ ̞̲̮̳̱͖̃ͬͯ̏̓̽͘ͅI̴͔͕͛ ̣̦͙̫Ṋ̺̩̞͓ͮͪͅ ̦͈̪̈́E̶̞̾̅̓ͅ   is a good thing, but don’t force yourself to do things you don’t feel like doing for more than a couple of days – if you consistently find yourself hating your work and wanting to do something else, you’re not lazy, you’re just interested in that something else. Find a way to do the something else!
  • Procrastination falls into 2 categories:
    • you procrastinate on everything: you should get some of that discipline and get to work
    • you procrastinate on some things, but are really engaged with and interested in (in essence, motivated to do) other things: provided that these other things are valuable and not just time-wasters, do these other things and quit the things that bore you


1. Deep Work

When you want to learn, multitasking is a big no. You probably want to master new or complex material – you can’t have your mind wandering off, checking social media and doing other low effort things. You have to be really laser focused.

And this means that you can’t steal a couple of minutes here and there and expect improvements in a new, complex skill. Most people need time to get into a focused mode. I recently saw an Instagram ad that featured a guy sitting on the toilet with a phone in his hand. It went something like this: ‘Got a minute? Learn coding on your phone.’ Like, maybe it works, I dunno. But, in general, this is my reaction to such things:


Okay, so 0 and 1 were both prerequisites to even begin learning. If you have the two of them down, you can move on to actual steps to learn things.

I used to think that speed reading was one of the first skills one needs to learn because it made everything else faster and more efficient. I think there are some merits to changing how we read, and I’ll just briefly describe what I mean here, but the overall benefits and promises of speed reading are far too blown out of proportion.


  • Skimming: Getting a general overview of the material before delving into the details is probably good. You first look at things in general and only then do you seek out specifics.
  • Asking yourself what you want to find out. Speed reading is most effective when you read relatively easy texts and when you don’t aim for full assimilation of the entire text, but for getting out some specific info. Want to find out one specific date, number, name, key concept? Speed reading is your friend. Want to absorb a novel, complex, academic paper, or a mathematical proof? You probably can’t breeze through that.

2. Spaced repetition

Now, on to the real stuff. When you learn new material, I don’t know of a more reliable method for keeping it in memory than spaced repetition. Spaced repetition has been studied for a long time and it seems to be the most useful thing for internalizing new material. It also works great with mnemonics, which I’ll describe later on.


  • (linked above; Gwern probably has the best overview of the literature)
  • ANKI – it’s free, it’s good, there’s a ton of already available decks

3. Testing and recall

In the Learning How To Learn course, a distinction is clearly made between things that people believe work, and things that actually work (in terms of effective learning).

Things that don’t work as well: drawing concept maps, re-reading the material, cramming and just trying to ‘plow through’.

Things that work better: trying to recall the material before re-reading it. Testing goes hand in hand because testing is just applied recall. And this is just something that sounds right: the best domains are those that actually test things, not just live in abstract, theoretical realms.

In practical terms, this can mean doing quizzes and sample tests – and this could be a good thing for academic endeavors – but I personally prefer what I call Feynman summaries (explained in item 5).

4. Don’t DiSSS me

Deconstruction, Selection, Sequencing, Stakes: DSSS or DiSSS for pronunciation. Tim Ferriss uses this technique to learn things, and it’s a very good system. First, you’ll ‘deconstruct’ whatever you want to learn and observe the ‘moving parts’. This is also what they call ‘chunking’ in the Learning How To Learn course.

Then, the process of selection where you apply the famous Pareto principle: which 20% of the subject matter will lead to 80% of the results? A common example in the accelerated learning community is languages: in general, around a 100 words of any language comprise 80% of its spoken volume. Think about how much you say ‘I’, ‘yes’, ‘here’, ‘could’, ‘for’ in comparison to ‘comparison’, ‘zesty’ or ‘blanket’. The first 100 words are the spoken language – you only need to fill in the remaining vocabulary when you know them, and this is far easier than doing the reverse.

I have a personal story on this: I used to teach beginner Swedish in a language school. I redesigned the curriculum, and instead of doing the course by the book, I focused exclusively on the first 100 words of spoken Swedish. The result: my students spoke much better after the same period of time. Drilling sentences like ‘I am here’, ‘He will not’, ‘Who are you?’, ‘Where is that?’ etc. brought real, actual results. It would be good to test this in a real scientific experiment, but from my personal, anecdotal experience, doing the 80/20 analysis works. (Check out this link for an orthogonal/contrarian view.)


5. A Feynman summary

This is my own combination of two methods that work really well. Method 1 is to explain something in very simple terms – this is also called the Feynman technique. After explaining, you identify the gaps in your knowledge, and then you go research some more to fill in these gaps.

Method 2 is to write a summary of what you know. This means that you test yourself, that you practice recall, but you’re not doing quizzes, you’re writing an essay. Combine the Feynman explanation with this written format of testing and you get the Feynman summary – a tool that helps you understand, teach, recall, and test your own knowledge.


6. The tribe

Being a lone wolf is limited – to learn something, it’s much better to connect yourself to people that are learning the same thing (or that already know the same thing).

If you’re the smartest person in the room, you’re in the wrong room.

Study groups are recommended in the Learning How To Learn course – and I also remember reading about the effectiveness of learning in pairs, but I can’t find the link to that. I have no personal experience to offer here, but having a tribe to keep you accountable generally seems like a good idea. The only prerequisite is that the team needs to be focused on the work, not on having fun.


7. The tricks

And finally, there are many tricks that are useful in learning new material. Here’s a list:

Pomodoro technique is simple: set an alarm for 25 minutes and turn off all distractions. No mobile phones, email, social media, or any other sort of distraction. Then focus on your subject. Read, write, watch a video, whatever – but be focused for 25 minutes. Take 5 minutes of reward. This means chocolate, social media, browsing, gaming. You need to feel good about your Pomodoros. Do a couple of cycles and then call it a day. Note that you might want to avoid Pomodoros for things that require 2-3 hours of deep focus. For me, Pomodoros are a good way of doing things that I don’t feel like doing.

Mnemonics are memory techniques. The most popular ones are number systems (like PAO) and memory palaces (loci). There are many others, but I like these two the most. The great thing about mnemonics is that they can work well with other things listed in this article, as well as with each other.

Planning. Make a plan for studying the night before – this is good because it gives you peace of mind: you already know what you have to do that day. Also plan your quitting time – it doesn’t make sense to constantly be in open loops. Set a finishing time and then feel at ease for having accomplished your load for the day. This way, you allow yourself to relax and you let your diffuse mode of thinking make new connections.

Sleep is absolutely, critically important. No sleep = no learning. Optimization freaks talk about the Uberman and the Everyman sleep cycles – I don’t know about these, maybe they work well – but sleep in general is critical.


This list will probably expand in the future. Updated on 30/11/2018.

Failing is training

The argument to moderation is a fallacy in which a person claims that the truth lies between two extremes. For example, you have the pro-vaccination arguments, and the anti-vaccination arguments, and since they are opposing each other, you concede that both arguments have some merit and that the truth is probably somewhere in the middle of the spectrum: maybe some vaccines, but not all.

And while going for balance may be a good thing, there is no guarantee that it will be a good thing.

If I haven’t convinced you yet, consider a society which tries to find a balance between slavery and non-slavery. It’s pretty fucked up to compromise with slavery, and it’s obviously wrong.

But if not used to make an argument, the Yin and the Yang can be useful for thinking about yourself. I’ve been looking for balance for a long time, and I will probably still be looking for balance for quite some time after this text.

One place where I’ve been looking for balance is between self-criticism and self-acceptance. If you’re too accepting of yourself, you’re soft and you don’t hold yourself to a high enough standard. That means that you cannot grow and do great things. On the other hand, if you smother yourself with criticism, you’re unhappy and always unsatisfied. None of these extremes look appealing: balance to the rescue.

Here’s one balancing thought that came to my mind yesterday:

I’m not failing; I’m training.

This isn’t an excuse – for myself, I am certain of this. I think that there are people who can twist anything to suit their needs, but I am fairly certain that at this point in time, I’m not using this as an excuse.

Here’s what I mean:

Let’s say that you have a track record of actually doing things. You’re an achiever – maybe not the achiever, maybe not a cocaine-modafinil IQ 180 monster, but you actually do things.

You have projects, and you try new things, and you branch out to new areas that interest you. This branching out could be, for example, trying out a new hobby, starting with a new sport, or applying to jobs in a new industry.

Of course you’re not successful or well-known. The world is not one big pond: you have many disconnected ponds, and being good or respected at one thing does not mean that anyone from the other pond knows about you. A successful boxer will fail for a long time before getting good at Jiu-Jitsu – but for him (or her), the pressure is much higher because (s)he is already a well-known boxer.

A small voice whispers to your ear then. It says: “Just quit, why are you embarrassing yourself here?”

This voice leads you to abandon new hobbies, new sports, new jobs, new relationships. Why? To maintain your high status – which you originally earned by starting at the very bottom.

Or a voice will whisper to your ear: “Maybe you’re not as good as you thought you were.” Maybe you continue with the new thing, but now you’re uncertain – because you’re used to win, and now you’re not winning, and now you’re permanently dissatisfied.

Not really necessary to say, but all of this is obviously stupid. Failing is training, and if you pursue new directions in your life – if you’re growing – then you will fail a lot, until you stop failing.

I’m not failing; I’m training.