Darwin Ate My DCF

Imagine a very clever calculator. You feed it three or four opinions about the future - how fast a company will grow, how long the good times last, how much cash it will throw off - and it hums for a second and prints out a single, exact answer: "This business is worth ₹742 per share." Not "about ₹700." Not "somewhere between ₹500 and ₹900." ₹742. Down to the rupee.

That calculator is real. Grown-ups in the stock market use it every day. It's called a DCF - short for discounted cash flow - and it is one of the most respected tools in all of investing. It looks like science. It has formulas, it has decimal points, it fits neatly in a spreadsheet, and when someone shows you their DCF they usually sound very sure.

Here is the whole chapter in one sentence: that confident number is built on guesses, and if you nudge any one of the guesses even a little, the answer swings wildly - so the precision is fake. The decimal points are not knowledge. They are a costume that guesses put on to look like facts.

That doesn't mean the DCF is useless or that you should throw it away. It means you should treat it the way you'd treat a friend who is often right about the big shape of things but hopeless with exact times - worth listening to for the direction, never to be trusted to the minute. This chapter is about learning to hear the difference between a number that knows something and a number that only sounds like it does.

The book this guide follows takes its lessons from Charles Darwin, so let's start with him - gently, in kid words, because the science is the whole trick.

Before Darwin, people looked at a living thing - a giraffe with its impossibly long neck, an eye that can see, a bird that knows exactly when to fly south - and reached for a precise, tidy explanation: someone must have designed each part on purpose, measuring every bone to the millimetre. It felt scientific because it was neat and complete. Every detail had an exact reason.

Darwin's idea was messier and, it turned out, far truer. He said nobody measured anything. Living things just vary a little - some giraffes are born with slightly longer necks than their brothers and sisters, by pure chance. In a place where the tastiest leaves are high up, the longer-necked ones eat a bit better, live a bit longer, and have a few more babies, who tend to inherit the longer neck. Do that for thousands and thousands of years and you get a long-necked animal - not because anyone planned the exact length, but because a rough advantage, repeated over an enormous stretch of time, slowly piles up.

Notice the two very different ways of being right here. The old story was precise and wrong: it gave an exact designer and an exact plan, and the plan didn't exist. Darwin's story was rough and right: it couldn't tell you the exact neck length of any single giraffe, or predict which baby would be born longest, but it captured the true engine - small advantages, compounding over deep time. Vague on the details, dead-on about how the world actually works.

There's a second thing Darwin's idea teaches us, and it matters just as much for investing. He was working across deep time - spans so long the human mind can't really picture them. Over a few years, nothing much seems to happen to a species; over a few million, a fish grows legs. The important changes hide out at the far end of time, where our imagination gives out. A spreadsheet has exactly this same shape, and exactly this same weakness: as we'll see, most of a business's value sits far out in the future, in years we can no more picture than Darwin could picture a million springs. We're comfortable being vague about deep time in biology. We are strangely willing to be precise about it in a spreadsheet - and that's the mistake.

That contrast - a neat wrong answer versus a fuzzy right one, and our habit of pretending to see clearly into a fog we can't - is exactly the trap waiting for us in the spreadsheet. Hold on to it. We're about to meet the giraffe's evil twin: a number so precise it fools you into thinking it's true.

Before we can see why the DCF fools people, we have to be fair to it and understand what it's honestly trying to do - because the goal itself is a genuinely beautiful idea.

A business is really just a machine for producing cash, year after year, hopefully for a very long time. When you buy a slice of a business (a share), what you're really buying is your slice of all the cash that machine will ever hand out - this year, next year, ten years from now, thirty years from now. So a sensible question to ask is: what is all of that future cash worth, in total, today? A business is worth the cash it will hand its owners over its whole life, pulled back to what that future cash is worth in your hand right now.

That "in your hand right now" part is the clever bit, and it has a name: discounting. Here's the idea in kid terms. Would you rather have ₹100 today or ₹100 handed to you in five years? Today, obviously - because ₹100 today you can use, or grow, or keep safe, while ₹100 in five years is stuck in the future and, thanks to prices slowly rising, it'll buy less by the time it arrives. So a rupee promised far away is worth less than a rupee in your pocket now. The further away the promise, the more we shrink it. Discounting is just the rule for shrinking future rupees down to their honest value today.

So the DCF does three steps, and they're all sensible:

1. Guess the cash. Estimate how much spare cash the business will produce each year, for many years into the future.
2. Shrink each year. Discount every future year's cash back to today - the further out, the more it shrinks.
3. Add it all up. Total those shrunk-down amounts into one number: the business's value today.

There is genuinely nothing wrong with that recipe. If we actually knew all the future cash, the DCF would hand us the perfectly correct price and investing would be solved. The recipe isn't the problem. The problem is hiding inside the word that appears in step one and never really leaves: guess.

And it's worth pausing on why the discounting step, sensible as it is, quietly makes the guessing worse. Discounting shrinks far-off rupees, yes - but it doesn't shrink them to nothing, and it shrinks them at a rate you also guessed. So the two hardest things to know - how much cash arrives in the distant future, and how hard to shrink it - are multiplied together, guess times guess. When you multiply one uncertain number by another uncertain number, the fog doesn't add up; it compounds. A model that starts with three honest "I'm not sure" inputs can end with an output that is far less sure than any single input - and yet it prints that output as one clean, sure-looking number. The arithmetic is exact. What it's exact about is a pile of maybes.

You might think, "So the number's a bit shaky - who cares, I'll just take it with a pinch of salt." Here's why it's more dangerous than that.

A number that looks exact doesn't just sit there quietly. It acts on your brain. When the spreadsheet prints ₹742, that decimal-point confidence quietly hushes the sensible voice in your head that was about to ask, "Wait - is this a good business? Will it even be around in fifteen years? Do I actually understand how it makes money?" The precise number feels like it has already done all that thinking for you. It hasn't. It has only done arithmetic on your guesses. But it feels like knowledge, and that feeling is the trap.

There's a name for this in the book's spirit: false precision. It's when the form of an answer - lots of digits, a tidy formula - promises more certainty than the content can possibly deliver. A weather app that says "73.4% chance of rain at 3:47 PM" isn't more accurate than one that says "probably rain this afternoon." It's just more confident-looking, and that extra confidence is unearned. In the market, unearned confidence is exactly what makes people put in more money than they should, and hold on longer than they should, because a number told them they were sure.

And here's the sharpest part. Being precisely wrong is more expensive than being roughly right, because the roughly-right person knows they're unsure and leaves themselves a cushion, while the precisely-wrong person, trusting their ₹742, leaves no room for the mistake they can't see coming. It is safer to be roughly right about a business - good, durable, understandable - than to be precisely wrong about it in a spreadsheet, because a fuzzy answer keeps you humble and an exact one makes you reckless.

There's one more reason this matters more than it seems, and it's about who usually wins the argument. In a room full of investors, the person with the tidy spreadsheet and the confident ₹742 tends to sound like the grown-up, and the person who says "I don't know exactly, but this looks like a strong business at a fair-ish price" tends to sound woolly and unserious. So the false-precise voice wins the meeting - not because it's more right, but because it's more sure-sounding. Certainty is persuasive even when it's empty. Learning to distrust the shiny number isn't just protecting your own wallet; it's learning not to be talked out of good, humble judgement by someone whose only real advantage is that their guesses came dressed in a spreadsheet.

Let's stop talking about this in the abstract. The fastest way to feel how flimsy the ₹742 is, is to build one and then poke it.

A DCF has a few dials on it - the guesses you feed in. The two most powerful dials are:

• The growth dial: how fast you think the cash grows each year.
• The discount dial: how hard you shrink future rupees (this is really a measure of how risky and uncertain the future feels).

The dangerous thing - the thing the spreadsheet never shows you - is how touchy the final number is. Nudge a dial by a hair, and the answer doesn't nudge; it lurches. Let me show you the shape of it before we do rupees.

Look at what happened. Nobody changed the company. Nobody discovered new facts. Someone just felt a little more optimistic about growth - 12% instead of 8%, a difference you couldn't tell apart in real life - and the "value" nearly doubled, from ₹560 to ₹1,010. If the true engine of your answer is that jumpy, then the last two digits of ₹742 aren't telling you anything. The honest output of this machine was never a number. It was always a range, and a wide one.

Now let's turn the dials ourselves, with real rupees.

Let's build a tiny, honest DCF together. illustrative

Picture an ordinary Indian business - call it a plain, mid-sized maker of packaged goods. This year it produced ₹100 crore of spare cash for its owners. We want to know what the whole business is worth today, based on the cash it'll make from here on.

To answer, we must feed the machine three guesses:

• Growth: we guess the cash grows 10% a year for the next ten years.
• Discount: we guess future rupees should be shrunk at 12% a year (this stands for how risky and uncertain we feel the future is).
• The long tail: after ten years we assume it keeps making steady cash forever, growing slowly at 4% a year.

Turn the crank on those three guesses and the machine prints a value of roughly ₹1,850 crore. There it is - a confident, exact-looking answer. If someone showed you only this, you'd nod: the business is worth about ₹1,850 crore.

But we built it, so we know a secret the number hides: every one of those three inputs was a guess dressed as a fact. We didn't know growth would be 10% - we picked it. We didn't know 12% was the right shrink-rate - we felt it. Nobody can know what a business earns nine years from now; the world is far too surprising. The ₹1,850 crore looks like a measurement, the way you'd measure a table with a ruler. It is nothing of the sort. It is arithmetic performed neatly on three hopeful opinions. A figure like ₹1,850 crore means nothing on its own; it's only as good as the story of assumptions behind it, and here the story is three guesses about a future no one can see.

Let's prove the guesses are load-bearing by changing exactly one of them.

We'll leave the business completely alone. Same company, same ₹100 crore of cash this year, same everything - except we'll turn the discount dial from 12% to 14%. illustrative

Why might someone reasonably use 14% instead of 12%? Because the future felt a bit riskier to them - maybe interest rates in the country are higher, maybe this industry feels more uncertain. It's not a crazy change. It's the sort of thing two sensible, honest analysts would simply disagree about over a cup of chai. Neither is being silly.

With the discount dial at 14% instead of 12%, the same machine now prints roughly ₹1,450 crore instead of ₹1,850 crore. Four hundred crore of "value" vanished - a fifth of the whole business - and nothing about the business changed. Not its factories, not its products, not its customers, not a single rupee it will actually earn. All that changed was one person's private feeling about how risky the future is, expressed as a two-percentage-point nudge on a dial.

Now stack the dials. Suppose an optimistic analyst uses 12% growth and a 10% discount, while a cautious one uses 8% growth and a 14% discount. Both are being reasonable. The optimist's machine might print around ₹2,600 crore; the pessimist's around ₹1,150 crore. Same business. Same day. Same spreadsheet formula. And a gap of well over ₹1,400 crore between two honest people - more than the value of the entire company, sitting inside the disagreement about a few guesses.

This is the punchline you must feel in your bones: the DCF didn't tell us what the business is worth. It told us what we already believed, translated into rupees. If you feed it hope, it prints a big number and calls it value. If you feed it fear, it prints a small one. The machine has no opinion of its own. It's a mirror wearing a lab coat.

And notice a small horror hiding in the numbers above. Two-percentage-point moves - 12% to 14%, 10% to 12% - are small. They're the kind of difference you can't feel and can't defend; ask an honest analyst why 12% and not 13% and the truthful answer is usually "it felt about right." Yet those undefendable little differences are the very levers that moved the answer by hundreds of crores. That's the signature of false precision: the inputs you can't pin down are exactly the ones the output leans on hardest. If the thing driving your answer is a number you picked by feel, then the answer is a feeling too - no matter how many decimal places it wears when it comes out the other end.

Now for the part that catches even careful people, because it's genuinely counter-intuitive. illustrative

You'd think the near years - next year, the year after - matter most in a DCF, since we can guess those best. And it's true we guess them best. But here's the cruel joke: in most DCFs of a growing business, most of the value comes from years far away - the distant future and that "keeps earning forever" tail - precisely the part we understand least.

Let's see it. Take our ₹1,850-crore business again. If you split that value into "cash from the first ten years" versus "cash from year eleven onward, forever," you often find something like this: the first ten years - the part we can actually reason about - might be worth only around ₹700 crore of the total. The remaining ₹1,150 crore - the majority - comes from that faraway "forever" tail, the years so distant that guessing them is barely better than making them up.

Sit with how strange that is. The DCF places most of its bet on the years it can see the least - like judging a whole cricket match by guessing what the score will be in a game a decade from now. That "forever" tail is where a tiny change in the long-run growth guess (4% versus 5%) or the discount dial swings the total by hundreds of crores, because you're adding up an endless stream and small differences, multiplied by infinity, become enormous. The most confident-looking part of the number is built on the foggiest guesses of all. That's not a small flaw at the edge. It's the machine resting most of its weight on its wobbliest leg.

The mistake is almost never "I built a silly model." The people who get hurt build beautiful models. The slip is subtler, and it's about what the beauty does to your judgement.

It starts as pride. You spend two weeks on a spreadsheet - dozens of rows, careful formulas, colour-coded tabs. It's genuinely impressive work. And precisely because you worked so hard on it, you begin to trust its output more than it deserves. All that effort feels like it should have bought certainty. It didn't. It bought a very tidy container for the same three or four guesses you started with. Effort polishes the box; it doesn't upgrade what's inside.

Then comes the quiet cheat almost everyone does without noticing: you already know the answer you want, so you tune the dials until the model agrees. You like the company, you hope it's worth more than its price - so you reach for the slightly-higher growth guess and the slightly-lower discount, both perfectly "reasonable," and lo, the model blesses your hunch with a big confident number. Now you've done something worse than trusting a guess. You've laundered your bias through a spreadsheet so it comes out looking like objective analysis. The decimal points don't just fool other people. They fool you.

Fairness cuts both ways, so let's guard against over-learning this lesson. "DCFs are built on guesses" is true - but a lazy mind can twist it into two wrong conclusions, and both are traps.

The first wrong turn: "So valuation is pointless - just buy good companies at any price." No. The whole reason the DCF exists is that a wonderful business bought at a mad price is still a bad purchase - you can overpay for anything. Distrusting the fake precision of a single number is not the same as distrusting the idea that price versus value matters. It matters enormously. What we're throwing away is the false decimal point, not the discipline of asking whether you're paying too much.

The second wrong turn: "So I'll just use my gut instead of any numbers." Also no - that's how you end up buying hot stories at insane prices with no anchor at all. The numbers still do real work: they force you to say your assumptions out loud. When you build even a rough DCF and discover that the price only makes sense if the company grows at 20% a year for fifteen years straight - a thing almost no business has ever done - the model has done its most valuable job. It didn't tell you the exact value. It told you what you'd have to believe to justify today's price, and let you judge whether that belief is sane.

That is the mature way to hold this tool. Use the DCF backwards and roughly: not to print "₹742" and buy, but to reveal the story baked into the price and ask if you'd bet on that story. The failure isn't using a DCF. The failure is mistaking its confident output for knowledge, and letting a spreadsheet do the judging that only you, looking honestly at the business, can do. A model can carry your judgement of a business; it can never replace it. Feed it a bad judgement, and it will hand that bad judgement back to you polished to two decimal places, looking for all the world like a fact.