Don't Predict. Calculate.

Nineteen pages. That is what she puts on the table, and it is genuinely good work. Five-year revenue projections. Market analysis. Competitive landscape research. Pricing strategy. Go-to-market plan. Customer acquisition cost. Lifetime value. Cohort retention curves. Everything a well-funded MBA program taught her to believe matters. Weeks of it. She is proud of the document. She should be. It is coherent and structured and every number traces back to a defensible assumption. I read it for 30 seconds. Then I ask one question. What is the probability that any of these assumptions is still true in 2 years? She looks at me like I asked her to name the colour of Tuesday. Nobody has ever put that question to her. In school, they asked whether the assumptions were reasonable. In her last job, they asked her to implement the plan. Not one person asked whether the plan survives contact with reality. I don't know, she says. I'm not sure. Right, I say. So start there. Because that is the only question in the document that actually matters. This is not pessimism. It is the difference between someone who survives a decade like the one coming and someone who gets flattened by it. Most people hold a plan. A plan is a prediction wearing a suit. And you cannot predict. Here is what you can do instead.

Start with the macro, because the macro does not care about your effort. Economic cycles. Currency moves. Interest rate regimes. Inflation, then deflation. Debt loads. Recessions. Wars. Supply chain breaks. Technology acceleration. Pandemics. Government policy swings. In 2003 I was inside Lehman Brothers on the day the power went out and the trading floor went dark. In 2008, 164 years of institutional permanence evaporated over a single weekend. That is weather. You do not change weather. You do not outwork weather. You get positioned before it arrives. The macro right now reads: higher prices, higher debt, and softer demand for knowledge work as AI commoditizes it. You can be the single best software engineer on the planet and still get dismantled by acceleration that makes your specific stack obsolete. That is not a verdict on you. It is a reading of the conditions. Then the micro. Where you live and what it costs you. What skills you are compounding. How you spend 5pm to 1am, the hours that actually change your trajectory. These are your navigation settings. If the macro says demand for knowledge work is falling and your micro says you live somewhere your only viable income is a job that funds the rent, you are not just fighting the macro. You are fighting it with extra weight bolted to your legs. Then the pace of the technology itself. A major economic cycle turns every 5 to 7 years. AI does not wait for the cycle. It can reprice an entire labour market in 18 months. That is the variable most people cannot calibrate, because nothing in the last 30 years taught them to. Three layers. Weather, settings, speed. You control one of them.

I have been through 6 of these transitions. Not because I am the smartest person in the room. Because I price probabilities instead of making predictions, and I move on that arithmetic earlier than almost everyone. In 2006 the luxury property market was visibly detached from the income that was supposed to pay for it. People were borrowing 10 times what they earned. So I ran the question. What is the probability that prices rise forever? Zero. What is the probability of a correction? High. I started moving out in 2006. Then 2007. Before it was visible, before the newspapers got there. When 2008 landed I was already out of the way. The nightclub worked the same way. I opened it in the late 1990s, when the category was still expanding. By 2001 I could see the probability curve bending. So I was already building elsewhere. Banking. Property. When the club finally had to close it cost me a business, not my balance sheet. Lehman was the same arithmetic. I loved that job. The rigour, the complexity, the sheer intellectual pressure of it. But I never assigned a 100% probability to Lehman existing forever. I was building relationships with people, not a career inside a bank. I was learning how capital actually moves, not cementing myself inside one capital structure. When it fell I lost a job I had assumed was permanent. I did not lose my identity, because I had never lodged my identity there in the first place. And in 2013 the Cyprus bail-in took savings from people who had been told their money was safe. Mine went with them. Even that was priced. It was painful, it was not fatal, and the difference between those two words is the entire discipline. Six collapses. Never leveraged through any of them.

Now the part that offends people who believe in merit. Like a Virgin was not the best pop song ever written. It was not the most innovative either. It landed in a window: limited competition at that exact point in pop history, enormous cultural attention arriving at the same moment, the right artist with the right image, global distribution exactly when distribution mattered. Release it 2 years earlier and it fails. Too early. Release it 2 years later and 10 other artists have already occupied the space. The song is good in all three timelines. The difference is timing, not quality. 20% talent. 80% probability. 80% cultural moment. 80% the wind blowing in your direction while your sail is up. Your plan has the same ratio. 20% strategy, 80% conditions. You own the 20%. You are a passenger in the 80%. So the question is never is this a good plan. The question is what are the probabilities this plan can work given the conditions, because some conditions make the probability zero no matter how good the plan is. I have watched brilliant founders with immaculate decks walk into a zero. Nothing about their execution was wrong. The window was closed.

You cannot predict. You can only position. Those are not the same skill.

Here is how I actually build. Map A to Z. Then find every point where the plan can die. A: you have the idea. B: you understand the market. C: you can raise capital. D: you can hire. E: you can build the product. F: you get traction. G: you reach profitability. H: you survive the competition. I: you survive a recession. J: you keep the team motivated. K: you adapt when the technology moves under you. And on until Z: you exit for the number you actually needed. Most people are excellent at A through E. Some reach F. Almost nobody has a plan for G through Z. Then they hit F, the economy turns, and the conversion rates they modelled assumed a growing market rather than a contracting one. So for every letter, ask two things. What is the probability this step succeeds given the conditions? And what is the backdoor if it fails? A to B the short way. If that closes, A to C to F to G. If that closes, A to K to M to Z. 10 routes to the same target, not 1 perfect route. The person with 10 ways to reach the goal beats the person with 1 perfect way. Every time. Not because they are smarter. Because they cannot be stopped by a single closed door. The 5-year projection collapses the moment the market contracts, because a projection is a straight line and the world does not produce straight lines.

Two students, same month, same question. One has an offer from Google. One has an offer from a seed-stage startup. Which should I take? That is the wrong question, and it is the question they always ask. The real chain is: what are the conditions? Growth or contraction? What is the probability of a recession in the next 24 months? Of layoffs in that sector? Of the startup's market being repriced? Then the micro: where do you live, what does it cost, what do you learn in each version of the next 2 years, who do you meet. Then the weights: if you want income, the safe seat wins. If you want to learn how companies are actually built, the startup wins. If you want optionality, the startup wins harder, because you will learn the machine instead of one component of it. Given all that, the answer is not obvious. The framework just forces you to think instead of feeling. Most people feel the fear and take the safe seat. Some feel the excitement and take the risk. Almost nobody prices the actual structure, which is why almost nobody can explain afterwards why it worked or why it didn't. And the clock is running on your skill. In 2020 roughly 1,000 people in the world could do what you do at your level. In 2025 it was 10,000. In 2026 it is 50,000. By 2030 it will be 500,000. Not because you got worse. Because education got cheaper, faster and global, and because the tools got better than the credential. So the skill is not the advantage. The skill is being commoditized in real time. The advantage is the part that cannot be commoditized: your judgment, your relationships, the trust people place in you, and your ability to see where something is broken and move toward it. That is not trainable in 12 weeks. It only accumulates by showing up for years. The framework scales from a career decision to your personal finances to geopolitics. It changes nothing about your ambition and everything about your exposure. Do not predict. Calculate. Replace the five-year projection with a distribution. Revenue at 80% of plan: 5% likely. At 40%: 20%. At 10%: 40%. Total failure: 35%. Now nothing about the straight line matters. What matters is that in 4 of 5 scenarios you are still in the game, and in the fifth you fail fast and redeploy. Given most of what you believe will happen will not, what is your actual path to safety? Not technical skill alone. Too dependent on one thing. Not money. It runs out. Not one income stream. Not one country. Not one identity. Positioning: low fixed costs, multiple income streams, multiple skills you can move between, and the ability to relocate when a market stops working. I am 49 and still in the fight. If I had exited with billions I would not be writing this book, I would be sailing. I am here because I never priced any of it at zero. Not Lehman, not the property, not the club, not my own industry. That is the whole method. Sit with it long enough and it stops being frightening. It becomes arithmetic.