Here’s a question that decides how you spend the next decade.
Is success mostly luck?
Because if it is, the sensible strategy is exposure. Take more shots, be in more rooms, buy more tickets, and accept that the distribution will do what it does.
And if it isn’t, the sensible strategy is something else entirely.
Most people answer this badly, and they answer it badly in a specific way: they look at outcomes, notice enormous variance, and conclude the process must be random.
That’s an observation error, and it’s one you can catch in yourself right now.
You only ever see the last link.
You see the exit, the promotion, the book on the shelf, the business that worked. You don’t see the seven years of sequence that produced it, because sequences aren’t visible from outside and nobody narrates them accurately afterwards.
So you’re looking at a set of endpoints and inferring randomness from their spread. You’d reach the same conclusion about trains if you only ever saw arrivals.
What a Causal Chain Is
A causal chain is the ordered sequence of conditions that must each occur for a given outcome to be produced. Outcomes are not generated directly but as the terminal state of such a sequence, which means the probability of an outcome is the compounded product of the probability of each link. This has a mathematical consequence that is unintuitive and decisive: a ten-step sequence in which each step has a seventy per cent chance of completion has an overall probability under three per cent, while raising each step to ninety-five per cent raises the overall probability to roughly sixty per cent. The leverage in any project therefore lies almost entirely in per-step certainty rather than in effort, motivation, or the desirability of the outcome. Because outcomes cannot be acted on directly — there is no operation that makes a result more likely without acting on its causes — attempts to influence outcomes directly produce no change, which is the structural reason that goal-focused effort so often fails.
Three properties, then what to do with them.
Property One: The Arithmetic Is Brutal
Take a project with ten steps.
You’re reasonably confident about each one. Not certain — you know roughly what to do, you’d probably get there. Call it seventy per cent per step.
Seventy per cent, ten times over, is 2.8 per cent.
Not seventy. Not “pretty likely.” Under three chances in a hundred.
Now change nothing about the project, the effort, or the person. Just resolve each step until it’s obvious rather than approximate — ninety-five per cent instead of seventy.
Ninety-five per cent, ten times, is roughly sixty per cent.
Same project. Same you. From a three per cent chance of completion to a sixty per cent chance, purely by moving each step from about right to certain.
That’s the whole argument for granularity and it’s why “break it into smaller steps” is directionally right and practically useless. Size isn’t the variable. Certainty per step is the variable, and size is only a rough proxy for it.
It also explains something that looks like a personality difference and isn’t. The person who reliably finishes things isn’t more determined than you. They’re operating at higher per-step certainty, and the difference compounds across the sequence until it looks like a difference in character.
From outside it reads as grit. It’s arithmetic.
Property Two: Order Is Load-Bearing
A chain is not a set.
This sounds obvious and it’s routinely ignored. People produce a list of everything that needs to happen and then work on whichever item is most appealing, which is usually the one they’re best at.
But links unlock other links. Some steps are only available after earlier ones have completed, and doing them out of order means doing them at low certainty — you’re attempting something before its preconditions exist.
The practical error looks like this: someone wanting to change career studies for a qualification before they’ve spoken to anyone doing the job. The qualification is a real link. It’s link six. Link one is a conversation, and doing six before one means you’ve committed a year to a direction you haven’t verified.
That’s not laziness or avoidance. It’s a sequencing failure, and it’s extremely common in conscientious people, because the out-of-order link is usually the one that feels most like work.
The cheap early links are almost always conversations, and almost nobody does them first.
Property Three: Some Links Aren’t Tasks
Third property, and it’s the one that connects this to everything else on this site.
Some links in a chain are not actions. They’re states.
You can’t complete “be someone people bring opportunities to” by doing it on a Tuesday. You can’t tick off “have the judgement to know which of these is the real problem.” Those are conditions of the person rather than items on a list, and they’re frequently the highest-leverage links in the entire sequence.
Which means a chain that requires you to be different has a link that can’t be executed, only developed.
This is why purely tactical approaches stall. You can build a flawless sequence and find that link four requires a capability you don’t have, and no amount of task management produces it.
It’s also the whole reason the four-hundred-year-old material at the centre of this site matters. A man whose life was predicted accurately for twenty years didn’t break the forecast by executing a better plan. He worked on the generator — the character producing his decisions — because that’s where the binding link was. That’s the method, and it’s the same claim in older language.
Then What Is Luck?
Real, and misplaced in most people’s model.
Luck determines which chains are available to you. Where you were born, who your parents knew, what happened to be hiring the year you graduated, whether the person you sat next to turned out to be useful. All of that is genuine and none of it is earned, and pretending otherwise is the ugliest thing this genre does.
What luck doesn’t determine is whether a chain converts once you’re on it.
Two people are handed the same opportunity. One compounds it for a decade. The other has it evaporate in eighteen months and genuinely cannot tell you where it went.
The difference isn’t luck — they had identical luck. It’s what happened to the sequence afterwards.
So the accurate model is: luck is an input to the chain, not an alternative to it. More exposure gives you more chains to work with, which is a real strategy and worth pursuing. It does nothing about your conversion rate, and conversion rate is where the compounding lives.
Someone with mediocre luck and high per-step certainty will out-produce someone with excellent luck and low certainty, over a long enough period, because the first person converts what they get and the second doesn’t.
That’s not a comforting claim and I’m not offering it as one. It’s just where the leverage is.
Building One Backwards
Here’s the practical method, and the direction matters.
Most people build chains forwards: first I’ll do this, then that, then hopefully… Forward construction produces plausible-sounding sequences that aren’t causal, because you’re generating steps from what you can imagine doing rather than from what the outcome actually requires.
Build backwards.
Start at the outcome and ask one question, repeatedly:
What has to be true immediately before this?
Not “what would help.” What must be the case, without which this cannot happen.
Then ask it again of that condition. And again. Keep going until you hit something that’s true today.
The chain you get is usually shorter than you feared and different from what you’d have written forwards. And it has a specific property that forward-built lists don’t: every link is load-bearing, because you generated it by asking what was necessary rather than what seemed useful.
Two things you’ll notice doing this.
You’ll find links you’d never have listed. Usually a person you need to know, or a piece of information you’ve been assuming, or a capability you don’t have. Those are the real chain and they were invisible from the front.
You’ll find the first link is embarrassingly small. Backward construction reliably terminates in something like “send one email.” That’s not an anticlimax — it’s the correct answer, and it’s the same terminus as the decomposition method in the Fog, arriving from the other direction.
Where the Effort Actually Goes
The sentence I’d want you to keep from this piece:
Stop trying to make the outcome likely. Make each link certain, and the outcome takes care of itself.
You cannot act on an outcome. There is no operation that raises the probability of a result without acting on its causes. Wanting it more doesn’t. Visualising it doesn’t — as covered in why manifestation isn’t working, the emotionally inhabited version actively reduces effort.
What you can act on is one link at a time, and the question for each is not “am I motivated” but “what would make this step near-certain?”
Usually the answer is unglamorous. More information. A conversation with someone who’s done it. Removing a dependency. Making the step smaller. Finding out what it actually costs.
Ten minutes of that per link, across ten links, moves your project from three per cent to sixty.
That’s the highest-return activity available to you and it feels like nothing while you’re doing it.
Three Hard Lines
This is not “trust the process.” That phrase describes faith in an unexamined sequence, which is the opposite of what’s being proposed. A process you haven’t decomposed and can’t assign certainties to isn’t a process — it’s a hope with a schedule. The whole value here is in measuring the links, and a slogan that discourages measurement is worse than nothing.
This is not a claim that outcomes are controllable. They aren’t. Sixty per cent is sixty per cent, and a well-built chain still fails four times in ten. What’s controllable is the conversion rate, and improving a conversion rate is a completely different activity from guaranteeing a result. Anyone promising the second thing is selling something.
This is not a denial of luck or structural advantage. Both are real and both are large. Two people running identical chains get different outcomes, and some of that is unearnable. The argument is that within whatever set of chains you have access to, conversion is the variable you own — not that everyone has the same set.
The Cost
Every piece here names its price. This one has two and they’re both real.
First: building the chain tells you how far away you are.
Right now the outcome sits at an unmeasured distance, which is uncomfortable and also permits optimism. Build the chain backwards and you’ll get eleven links, several of which you haven’t started, one of which requires a capability you don’t currently have.
That’s a number where there used to be a feeling, and the number is usually worse than the feeling.
Some people stop right there, and I’d rather warn you than have you meet it unprepared.
Second: the work is boring, permanently.
Raising per-step certainty is research, conversations, small clarifications, and removing ambiguity. None of it feels like progress toward anything. There’s no moment of momentum, no sensation of the thing coming together.
It just quietly moves a project from three per cent to sixty, and then a year later the outcome arrives and everyone tells you that you got lucky.
Common Questions
Is success random?
Outcomes vary enormously, which makes the process look random from outside — but you only ever observe endpoints, not the sequences that produced them. Outcomes are the terminal state of a causal chain, and the probability of the outcome is the compounded product of the probability of each link.
How much of success is luck?
Enough to matter and not in the place most people put it. Luck largely determines which chains are available to you, which is genuinely unearned and often decisive. It doesn’t determine whether a chain converts once you’re on it, and conversion is where compounding happens.
Can you actually engineer an outcome?
Not directly — there’s no operation that makes a result more likely without acting on its causes. What you can engineer is per-step certainty, and the arithmetic means that moving each step from approximate to obvious changes the overall probability by an order of magnitude.
Why do some people get results consistently?
Usually because they operate at higher per-step certainty, and the effect compounds across the sequence until it looks like a difference in character or determination. From outside it reads as grit; structurally it’s arithmetic.
How do you find the chain rather than inventing one?
Build backwards. Start at the outcome and repeatedly ask what must be true immediately before, until you reach something true today. Forward-built lists tend to contain steps you can imagine doing rather than steps the outcome requires.
What if a link isn’t a task?
Some links are states rather than actions — a capability, a piece of judgement, a way of operating. Those can’t be executed, only developed, and they’re frequently the highest-leverage links in the sequence. A chain that requires you to be different has a link that no amount of task management will complete.
Where to Start
Thirty minutes. One page. One outcome.
Take the thing you most want and write it at the bottom of the page. Bottom, not top — you’re building upwards.
Then ask, once: what has to be true immediately before this?
Write that above it. Then ask the same question of that condition. Keep going until you write something that is already true today.
Don’t optimise the chain. Don’t worry whether it’s the best route. Just get one honest sequence on the page.
Then do one thing: go along the chain and mark each link certain or approximate.
That’s the whole exercise, and the marking is where the value is. Every “approximate” is a place where you’re losing probability, and every one of them can be converted with an hour of unglamorous work — a conversation, some information, a smaller version of the step.
You’ll almost certainly find that the first approximate link is near the beginning, and that you’ve been working on link seven for months.
I’ve spent twenty-five years working on how ordinary people develop capabilities that look impossible from the outside, and the thing that separates people who reliably get where they’re going isn’t ambition, talent, or luck.
It’s that they know what the next link is, and they’ve made it certain before they touch it.