The Map Is Not the Territory - Why the numbers we trust in finance are sometimes only simplified versions of reality

There is a moment every serious investor has experienced at least once.

You open a brokerage research report. Twelve pages. Dense with tables, charts, scenario analysis, sector comparisons. At the bottom of page one, in confident bold text, sits a price target: ₹2,342 per share.

Not ₹2,300. Not "around ₹2,300 to ₹2,400." Exactly ₹2,342.

Something about that precision feels reassuring. It feels like someone has done the work. Run the numbers. Figured it out. The decimal place, specifically, signals rigour. This is not a guess. This is analysis.

Except — and this is the thing the report does not say — it is built almost entirely on assumptions. Assumptions about revenue growth five years from now. Assumptions about profit margins in year three. Assumptions about what the business will be worth at the end of the forecast period, stretched out into perpetuity. Change any one of those assumptions by a small amount, and the ₹2,342 becomes ₹1,900. Or ₹2,900. The number is precise. The confidence underneath it is borrowed.

This is not a criticism of research analysts. They are doing something genuinely useful. But it is an illustration of a problem that runs through all of financial modelling — a problem that a Polish-American philosopher identified clearly in 1931, long before anyone had built a DCF spreadsheet.

The Map and the Territory

Alfred Korzybski was not a finance person. He was a scientist and philosopher who spent his career studying how human beings use language and symbols to represent reality — and how catastrophically wrong things go when we forget that the symbol is not the thing itself.

His most famous line was simple: "The map is not the territory."

A map of Bengaluru is useful. It helps you navigate. It shows you roads, landmarks, distances. But it is not Bengaluru. The map has no traffic on Hosur Road at 6pm. It has no waterlogging during the monsoon on Sarjapur. It does not know that the shortcut through the layout has been blocked for six months by construction. The moment you mistake the map for the actual city, you stop paying attention to what is in front of you — and that is when you get lost.

Korzybski's insight, when applied to finance, is both obvious and deeply underappreciated.

Every financial model — every DCF, every P/E multiple, every credit rating, every risk framework — is a map. It is a structured, simplified representation of a reality that is far more complex, more dynamic, and more surprising than any spreadsheet can contain. The model is useful precisely because it simplifies. And it is dangerous precisely because it simplifies.

The map is not the territory. But in finance, we forget this constantly.

The DCF — The Most Seductive Map

The Discounted Cash Flow model is the intellectual centrepiece of fundamental equity analysis. It is taught in every MBA programme, used by investment banks, private equity firms, and individual investors. It has a rigorous mathematical structure. It rests on a genuinely sound idea: the value of any asset is the present value of all the cash it will generate in the future.

The problem is not the idea. The problem is what happens when you try to implement it.

An analyst building a DCF must make a series of forecasts: revenue growth for the next five years, operating margins, capital expenditure, working capital changes, and a discount rate that reflects the riskiness of the business. Each of these is an estimate. Each estimate carries uncertainty. But the model takes all of them as inputs, runs the calculation, and produces a single output — a "fair value" per share — with no visible uncertainty attached to it.

Here is where it gets genuinely important to understand. In most DCF models, the terminal value — the estimated value of the business beyond the explicit forecast period — accounts for roughly 60% to 80% of the total calculated enterprise value. This is the part that assumes the business will grow at some constant rate forever, or can be sold at some assumed multiple at the end of year five.

And this terminal value is ferociously sensitive to the assumptions you plug in. Research on DCF modelling has shown that terminal value depends on the gap between the discount rate and the perpetual growth rate, Even a half-percentage-point change in the discount rate can materially change the implied valuation, particularly when a large proportion of value comes from distant cash flows.

On a large company, this is not a rounding error. This is thousands of crores of "value" appearing or disappearing based on a number the analyst typed into one cell.

The model is not lying to you. It is doing exactly what it was designed to do. The discipline matters. The number at the end — the ₹2,342 — does not. What the DCF actually produces, when used honestly, is a way to understand what assumptions the current market price implies, and whether those assumptions seem reasonable. That is its genuine value.

Mistaking the output for a fact about the territory is where the trouble begins.

The P/E — The Map Everyone Carries

If the DCF is the map used by professionals, the Price-to-Earnings ratio is the map that everyone carries in their pocket. Simple, quick, intuitive: divide the price of a share by its earnings per share, and you have a sense of how "expensive" or "cheap" it is.

But even this apparently simple map requires asking: which price? Which earnings?

Trailing earnings — what the company actually earned last year — may be distorted by one-time events, accounting choices, or a cyclical peak or trough. Forward earnings — what analysts expect the company to earn next year — are forecasts, and forecasts are wrong with impressive regularity. Earnings themselves can be adjusted, normalized, inflated, or obscured by the choices management makes in how they report.

A P/E of 30 might be expensive for a slow-growing commodity business and cheap for a high-quality compounder with durable competitive advantages. The same number on the map means different things in different territories.

In the 2000 technology bubble, many stocks appeared reasonably priced on forward earnings estimates. Those estimates assumed that rapid revenue growth would continue and eventually translate into profits. The estimates never materialized. The map said one thing; the territory said something entirely different — and when investors finally looked up from their maps, the damage was done.

When the Map Says AAA and the Territory Disagrees

In September 2018, IL&FS — one of India's largest infrastructure financing companies — defaulted on a series of debt obligations, triggering a severe liquidity crisis across Indian financial markets.

This was not a surprise that should have surprised anyone who looked carefully at the territory. IL&FS had accumulated roughly ₹91,000 crore in debt. The group had accumulated very high leverage, and several subsidiaries were already showing signs of financial stress before the eventual defaults.

But the credit rating — the map — said something different. Multiple SEBI-registered credit rating agencies had maintained IL&FS at AAA, the highest possible creditworthiness rating, until weeks before the default. According to the parliamentary committee that later examined the crisis, the rating had remained at its top grade even as early warning signs accumulated over years. When the default finally arrived, the rating went, in the words of one report, "straight from AAA to default." There was no middle ground. The map simply did not show the terrain that was actually there.

Credit ratings are maps. Useful, carefully constructed, frequently right — but maps nonetheless. They model risk using available financial data, structured assumptions, and analytical frameworks. What they cannot fully model is what is hidden, what is changing faster than the model can update, or what lies outside the framework entirely.

The IL&FS episode illustrates the danger of treating a credit rating as a complete representation of credit risk. Subsequent regulatory proceedings identified shortcomings in the rating process, including concerns about reliance on management submissions and the adequacy of due diligence.

What George Box Knew

George Box was a British statistician — one of the most influential of the twentieth century. He said something that every person who works with financial models should tattoo somewhere visible:

"All models are wrong, but some are useful."

This is the correct relationship to hold with every financial model you will ever encounter.

Not: models are useless, throw them away. Not: this model is wrong, therefore I will not use it. But: every model is a simplification of reality by definition. The moment it became a model, it left things out. Some of what it left out will matter; some will not. The job of the analyst — and the investor — is to use the model while remaining acutely aware of where the map ends and the unmapped territory begins.

The DCF's real value is not the number it spits out. It is the discipline it forces on your thinking. What revenue growth rate does the current stock price imply? Is that realistic given the company's history and competitive position? Which assumptions drive most of the calculated value? How confident am I actually in those? What would need to change in the world for this investment to go wrong, even if my model says it is cheap?

These questions are what the model is for. The ₹2,347 is a by-product of asking them, not the point of the exercise.

The P/E ratio is useful not because 15 is definitely cheap and 30 is definitely expensive, but because it invites you to ask: what level of growth and profitability is this price implying, and do I believe that story?

The credit rating is useful not as a guarantee of safety but as one input in understanding a structured view of risk — to be supplemented, questioned, and cross-checked against the territory itself.

The Money Vichara for Today

A company is not a spreadsheet. It is thousands of people making decisions every day. It is a culture that may or may not survive the departure of a founder. It is a competitive landscape that shifts in ways no five-year projection can anticipate. It is a regulatory environment, a macroeconomic tide, a technology disruption, a management decision that has not been made yet.

None of these things fit neatly into a cell. None of them will appear in the model unless the person using the model thinks to put them there — and even then, they will be simplified, compressed, and stripped of their actual texture.

The map is useful. The map is necessary. Without maps, we cannot navigate at all, in cities or in markets. The investors who dismiss all models entirely are not wiser than those who use them — they are just navigating blind.

But the investors who mistake their models for reality — who trust the ₹2,342 because the spreadsheet is detailed, who trust the AAA because the rating agency has a long history, who trust the P/E because the number is tidy — these investors are holding a map so tightly that they have stopped looking out the window at the road.

The territory is always more complex, more surprising, and more alive than any map can capture.

So here is today's vichara to carry with you:

The next time you look at a valuation, a rating, or a ratio — ask yourself honestly: am I using this as a map, aware of what it cannot show me? Or have I started to believe that this number is the territory itself?

Because the answer to that question is the difference between a tool and a trap.