The Skill-Luck Continuum: Why You Must Know Which One Drove the Outcome

You’ve had the best eighteen months of your career. Every position you opened ran your way, or every deal you chased closed, depending on what you trade in. The numbers are not subtle. You’re top of the table, the bonus is real, and people have started asking you how you do it. You’ve started answering, because by now you have a story: you read the setup faster than the room, you sized the bets others flinched at, you trusted the call when it was unpopular. The story holds together, and the results back it.

Here’s the part the story never mentions. A run this good would also show up if you were merely decent and the variance happened to break your way for a year and a half. Those two worlds look identical from the inside. Same wins, same table position, same admiring questions, same confident answer. The only thing that separates them is something you can’t see from where you’re standing: how much of this was you, and how much was the deck you happened to be dealt. Get that proportion wrong and you’ll learn exactly the wrong lesson from the best stretch you’ve ever had.

The evidence

Start with a finding that should unsettle anyone on a streak. Take a set of firms with identical capabilities, no skill differences at all, and let their results run with normal random variation. Some of them will pull ahead and stay ahead for long stretches, purely by chance. Watched from outside, those firms look like persistent outperformers, the kind that get written up for their strategic brilliance. The pattern of sustained advantage looks the same whether a real edge produced it or pure randomness did, and observers reliably credit the edge. A streak is what skill and luck both produce, which is exactly why it tells you so little on its own about which one you have.

It gets sharper. When you model talent and luck running together across many careers, the people who end up most successful are almost never the most talented. They tend to be moderately talented people who caught the most lucky breaks. Talent matters, but as a threshold: below some bar you’re out, and above it luck takes over the variance. The spread in success comes out far wider than the spread in talent, and that extra spread is luck doing the work. Read that against your own table position. Being near the top is consistent with being the best in the room, and just as consistent with being good enough and well-placed when the variance arrived.

So the question is not whether you have skill. You clearly cleared the bar or you wouldn’t be in the seat. The question is how much of this particular run to read as a verdict on your judgement, and that depends on where your work sits on a line. At one end are activities where the same input gives the same output almost every time: chess, a hundred-metre sprint, playing a written piece of music. At the other end are activities ruled by chance: a roulette spin, a lottery ticket. Most real work, trading, closing deals, hiring, betting on a product, sits in the middle, and the middle is where streaks lie to you the most.

How it works

There’s a fast way to find where on that line your work sits, and it’s almost too simple. Ask whether you could lose on purpose. A chess player can throw a game on demand any time they like, because they control the outcome end to end. Try the same in your seat. Could you reliably produce a losing quarter, on command, without anyone able to tell you were doing it deliberately? If you can’t manufacture a bad result at will, that’s because the result was never fully yours to manufacture. The same wall that stops you losing on purpose is the luck in your outcomes, and the harder it is to lose on cue, the more of your good run belongs to the deck rather than to you.

This is why your mind reaches for the skill story so fast. You see an outcome and you build a cause, instantly and without effort, and a coherent cause feels true. But coherence is a feature of the story, not proof it’s right. The world has more noise in it than your sense-making will admit, and a long lucky run is exactly the shape of noise that’s easiest to mistake for a signal about yourself.

The lesson your best run hands you is only worth as much as your honest estimate of how much of it you actually caused.

How to use it

The basic move is to stop reading a single result as a verdict and start estimating a proportion. After an outcome that matters, ask: if I made this exact call a hundred times under similar conditions, how many times would I get this result? You don’t need a real number. The difference between roughly ninety and roughly forty changes everything about what you take away, and you can feel which neighbourhood you’re in.

The part worth getting right is how to set that number honestly instead of letting the streak set it for you. Four features of the work push the answer toward luck, and you can check them directly. How wide is the spread of possible outcomes from the same decision? How long before you find out whether a call was good? How many independent shots do you actually get in a career? How tangled is the chain between what you did and what happened? Wide spread, slow feedback, few shots, tangled causes: that’s the luck end, and your hundred-rep number should drop accordingly. A trader gets a fast, clean signal on a position and many independent trades a year, so skill surfaces sooner. Someone making a handful of large bets a year, each taking years to resolve, could run hot for a decade on variance alone and never get enough shots to know. Be most suspicious of your number precisely where the bet is biggest and slowest, because that’s where the lucky run and the skilled one look identical for the longest.

Then judge the thing you can actually control, which is the process rather than the score. In high-luck work a single result is a poor read on whether you decided well, so grade the reasoning instead. Were the right factors weighed? Was the bet sized to the real uncertainty rather than to your confidence? Was the conviction earned or borrowed from the last three wins? Over many decisions a sound process beats a sloppy one. Over a handful, the variance drowns the signal, and your streak is a handful.

The hardest move comes mid-run, when everything’s working and nobody around you wants the audit. The danger isn’t the good mood, it’s that you’ll size the next bet to the streak instead of to the odds and push harder exactly as your read of your own edge runs furthest ahead of the evidence for it. Run the hundred-rep check on your current hot hand while it’s still hot, not in the post-mortem after it breaks. If the honest number is closer to fifty than ninety, don’t quit. Keep the same process and refuse to let the run inflate your sizing, so that when the variance turns, as in a high-luck game it will, it costs you a bad quarter and not the whole account.

Why it matters

There’s a real cost on each side of getting this wrong, and they’re mirror images. Read luck as skill and you turn reckless: you repeat the call that happened to land, raise the stakes, and treat a coming reversal as bad luck rather than the average finally showing up. Read skill as luck and you turn passive: you shrug off a genuine edge as a fluke, stop pressing the advantage you actually have, and underinvest in the one thing that does move your odds. The streak pushes you toward the first error. A long enough slump pushes you toward the second. Both are expensive, and both come from the same failure to ask which force was driving.

What you’re really building is a sense of how much to trust any lesson a result hands you. A win that came from a sound process is worth repeating, while a win that came from variance will teach you the wrong thing and do it most convincingly right when you’re winning. Knowing the difference is what separates the people who genuinely sharpen over a career from the ones who just stack up outcomes and mistake the pile for proof. Your best eighteen months can make you better or it can set you up to give it all back. Which one depends on whether you were honest about how much of it was the deck.