Customers' Yachts Ch. 2: The Forecasting Business — Selling Certainty Nobody Has

Investment Background

Chapter 1 established the misalignment. This chapter covers its principal product.

Schwed's treatment of forecasting is the book's most famous section. And his argument is far subtler than "forecasts are inaccurate."

His argument is that the forecasting business exists independently of forecast accuracy.

The Wall Street Translation

The Asymmetry in How Forecasts Are Scored

Schwed observed something true in 1940 and still true today:

People who make bold forecasts are remembered when right and forgotten when wrong.

That asymmetry creates a rational behavioral strategy:

• Make a bold, specific forecast, and if it happens to be right, your career is made.
• If it is wrong, almost nobody follows up, because everyone forecasts and everyone is wrong.

So for a practitioner, the expected value of making bold forecasts is positive — even with zero forecasting ability.

Note the form of that conclusion. It requires nobody to be dishonest. It requires only an asymmetric scoring system.

Which is why financial media is permanently full of specific predictions: not because anyone can forecast, but because the odds facing forecasters are favorable.

A Concrete Mechanism: Scattered Predictions

Schwed describes a more blatant version, and it exists today in various forms.

Imagine a thousand people each randomly predicting whether the market rises or falls next year.

About five hundred will be right.

In year two, those five hundred predict again. About two hundred fifty are right twice running.

After five years, roughly thirty people have five consecutive correct calls.

Those thirty now hold an extraordinary-looking track record produced entirely at random.

They will write books. They will be interviewed. The other nine hundred seventy will not.

This is the same mechanism as the survivorship bias in Chapter 5 of A Random Walk — where it argues that active management track records are not credible.

What this book adds is the supply-side explanation: the mechanism does not merely passively produce an illusion. It creates an active business model.

Why Customers Demand Forecasts

This is Schwed's most insightful passage and the most easily overlooked: the problem cannot be blamed entirely on the supply side.

Customers want forecasts.

Facing uncertainty, the psychological need is for an answer — any answer. An advisor who says "I do not know what markets will do next year; nobody does" leaves the client feeling they provided no value.

An advisor who says "we project 8% next year, overweight technology and healthcare" feels professional.

Both may have exactly the same forecasting ability: zero.

This is how Chapter 1's selection mechanism actually operates: customer preferences select for confident practitioners.

So this is not a story of an industry deceiving customers. It is a story of an industry supplying what customers demand. And what they demand does not exist.

Division of Labor With the Rest of the Library

Boundaries are needed, since "forecasting does not work" appears repeatedly here.

Book	What it says about forecasting
A Random Walk	The evidence — the historical record of professional forecasts is poor (efficient market argument)
Thinking, Fast and Slow	The psychology — the illusion of validity, why experts are overconfident in their own forecasts
Against the Gods ch05	The epistemology — some futures carry no computable probability at all (Knight)
This book	The business model — why forecasts are mass-produced and sold despite being known not to work

Four books, four different levels.

Kahneman explains why forecasters believe themselves. Knight explains why some things are unforecastable. Malkiel supplies the evidence on how poor the record is.

And this book answers a question none of the three addresses: given all that, why does the business thrive?

Because its business model does not depend on accuracy.

Today's Forms

An honest update to 2026 is required, because the forms changed.

The 1940 form: broker telephone recommendations, investment newsletters, market commentators.

Today's forms:

• Annual outlook reports — every December, every institution publishes next year's target level. Almost nobody reviews last year's accuracy.
• Continuous financial media commentary — every market move requires an explanation, and that explanation is manufactured after the close.
• Algorithmically driven content — push notifications engineered to create urgency.
• Confidence on social media — a particularly pure form, because it has no accountability mechanism at all.

The channels changed. The asymmetry did not: right is remembered, wrong is forgotten.

A concrete test you can run yourself: find any large institution's annual target published last December, then compare it to the actual result. The exercise is usually startling, and more notable still: almost nobody performs it, including the institutions issuing the forecasts.

Executable Trading Rules

1. For any forecast, first ask "where are this person's wrong forecasts recorded?" The chapter's most practical line. If no such record exists, their correct forecasts carry no information. A forecaster with no record of failures is a forecaster with no track record.

2. Treat "I don't know" as a mark of professionalism rather than incompetence. The concrete form of Chapter 1's third rule. In genuinely unknowable domains, admitting unknowability is the only honest position.

3. Do an annual forecast review. Concretely: write down the forecasts you read this year that influenced a decision, and check them next December. The exercise permanently changes your attitude toward forecast content, and it takes one year.

4. Distinguish "forecast" from "conditional plan." "Markets will fall next year" is a forecast and useless. "If markets fall 30%, I will take the following actions" is a plan and extremely useful. The latter requires knowing nothing about what will happen.

5. Do not extend "forecasting does not work" into "planning is unnecessary." A common overreach. Chapter 5 of Against the Gods gives the correct position: calculate where frequency data exists, buffer where it does not. Neither is forecasting.

Relevance to a Retirement Portfolio

This chapter has a very concrete application for retirement investors, and it concerns a real sum of money.

Your retirement plan requires no forecasts to function.

Worth developing, because it is counterintuitive. A retirement plan appears full of assumptions about the future — returns, inflation, longevity.

But note the nature of those assumptions: they are not forecasts. They are parameters of a distribution.

Forecast (useless)	Planning parameter (useful)
"The S&P will reach 6,000 next year"	"The historical range of long-run real returns is 5–7%"
"Inflation will fall to 2% next year"	"Here is the historical distribution of inflation; I test against the worst case"
"Technology will lead"	"I hold the whole market, so I need not know which sector leads"

Nothing in the right-hand column requires anyone to forecast anything.

And that is the most underrated virtue of a low-cost broad index fund: it is the one product that explicitly requires no forecasting.

It is therefore also the product this industry is least eager to sell — because it needs no ongoing advice, generates no trades, and requires no annual outlook report.

Chapter 1's question has its clearest answer here: if you buy a total-market index fund and leave it alone for thirty years, the industry earns approximately nothing from you.

That is neither a coincidence nor a conspiracy. It is a direct consequence of Chapter 1's structural misalignment.

Chapter 3 covers how that misalignment shows up in fees: the costs you can see, and the ones you cannot.