# Chapter 28: The Book: the long record

*The Analyst's Handbook, Sifter's textbook for student investors · By Matt Lang and Bree Lang · The record · https://sifter.finance/textbook/ch28-book*

A mandate in one line, a portfolio run for months at real prices, and a reason on every trade. The exercise a reader weighs most.

Where this is on Sifter: Training Floor · Exercises · Book

## Why a book that runs for months

The three exercises so far each take one claim and settle it. The Book takes no single claim at all. It is a portfolio, $100,000 of paper money at real prices, run for as long as you like, and what it records is not whether you were right but how you decided, one trade at a time, with the reason you gave on the day you gave it.

A Sprint record shows whether your confidence can be trusted. A Book shows whether you can hold a position through a bad month, whether you sell for the reason you said you would, and whether the reasons you give in March still make sense when someone reads them in November. None of that can be shown in two weeks or written the night before.

The page opens on one line, "Run a portfolio for months, and say why on every trade." This chapter goes through that line.

## The page

Before a book exists, the page asks three things in order. First, "What is this book trying to do?", with a box for one line and an example under it: "beat the S&P 500 by 3 points over 12 months with a tracking error under 6%, holding 8 to 12 stocks." That line is the mandate, and it is pinned to the book for as long as the book runs.

Second, the stocks. You search for one, and it appears as a row with a weight, a box that says "Why [ticker]? One line.", and a question, "How sure?", answered from 1 to 5. Every row needs all three before the book can start, and the weights cannot add to more than 100 percent. A line under the rows keeps count: how many stocks, what share is invested, and what share is still cash.

Third, the two doors. "See the evidence" opens once two or more stocks have weights, and "Start the book" opens once the mandate is written and every row is complete. Under the lock the page says what pressing it does: buys at today's prices, every memo goes on your card, and rankings start at day 30.

## The simplest mandate

The mandate is the Book's version of the sentence with a blank in it, and the simplest one is the example the page offers. Beat the S&P 500 by some number of points over twelve months, with a tracking error under some number, holding some number of stocks. Three numbers and a benchmark, which is the same shape the Quarter's briefs had, except that here you write it.

If you write your own line, the book is scored as an information ratio, the excess return over the S&P per unit of tracking error from Chapter 7. If you would rather start from a guide, the page offers five: information ratio, total return, Sharpe ratio, and total return or Sharpe ratio capped at five stocks. Picking one writes a line for you, which you can edit, and sets what the book is scored on. The status line under the book's value says which, in the words "scored as," for the life of the run.

A book on its own is hard to judge, because nobody knows what it was for. A book with a mandate can be read against the mandate, and a book that said "under 6% tracking error" and ran at 16 has told the reader something even if it made money. Everything else on the page is measured against that one sentence.

### Check yourself

A student picks the guide "Total return, five stocks," runs five stocks for a year, and beats the S&P by 9 points with a tracking error of 16 percent. The Overview tab prints an information ratio. What is the book scored on?

- Total return, because the guide set it, and the information ratio is printed as context. **(correct)** Right. The guide writes the mandate and sets the goal the book is scored on, and the status line says "scored as Total return" for the whole run. The Overview prints the return, the excess, the information ratio, the tracking error and the worst fall for every book, whatever it is scored on, because a reader wants to see them, but only the goal enters the score.
- The information ratio, because that is what the Overview shows. The Overview shows the same seven numbers for every book. What a book is scored on is set by the mandate, either your own line, which means information ratio, or the guide you picked, and the status line under the value says which.
- Tracking error, because 16 percent breaks the mandate. Tracking error is never the score. A total-return guide does not set a tracking-error limit at all; a custom line that promised one would be read against it by a reader, but the score would still be the information ratio the custom line is scored on.

## Look before you type

Once two or more stocks have weights, "See the evidence" opens the same three lines you met on the Quarter, read on your draft. The first says what a book at these weights ran over the years these stocks share, its beta and its tracking error. The second looks inside: which stock holds the most of the risk, and which sector holds the most of the book. The third is the calendar, how many of your stocks report in the next three months.

Read them the way the Quarter chapter taught, with one difference. On the Quarter you were about to predict two numbers and be scored on them. Here you are about to write a mandate, and the first line is the fastest way to find out whether the mandate you wrote and the book you built are the same thing. A line that promises a tracking error under 6 percent above a draft whose history ran at 14 is a mismatch to fix before the lock, not after.

Once the book is live the same three lines are behind the Evidence tab, read on whatever you hold today, so you can come back to them after every trade.

## Start the book

Pressing "Start the book" buys every row at today's prices, in the weights you set, and from that moment the page changes. The book's value leads, with its return next to the S&P's over the same days, then one status line, the book's number, when it started, how many days it has run, and what it is scored as. Then the mandate. Then a "Trade" button on the right, and under all of that six tabs.

Overview is the return, the excess, the information ratio, the tracking error, the worst fall and the days, a chart of the book against the S&P from the same start, and the last three decisions. Holdings is each stock with its weight, its return since the buy, and the memo you wrote on it. Decisions is the whole log, which the next section is about. Evidence is the three lines on today's holdings. Analytics and Ranking are the compete surface, for those who want it, and the ranking does not count a book until it is 30 days old.

![The live Book page: the line "Run a portfolio for months, and say why on every trade.", the value $116,903 with +16.9% and "S&P +7.6% over the same days", the status line "Book 1 · since Apr 2026 · 128 days · scored as Max Return", an empty mandate box reading "This book has no mandate yet. Write one, in your words." with a Pin it to the book button, a Trade button, the tabs Overview, Holdings 6, Decisions 9, Evidence, Analytics and Ranking, the numbers return +16.9%, S&P 500 +7.6%, excess +9.3 pts, info ratio 1.62, tracking error 16.3%, worst fall 9.7%, days 128, a chart of the book against the S&P, and Latest decisions with three rows, each a date, an action such as Updated STRK, its reason and "sure 3".](https://sifter.finance/textbook/figures/ch23-book.png)

*A live Book, four months in. Training Floor › Exercises › The Book, signed in. This book was started before mandates existed, so the box under the status line is still asking for one, and it is scored on total return. The Overview tab shows the numbers, the chart and the last three decisions; the log behind Decisions holds all nine, each with the reason given at the time and how sure the analyst was.*

## Say why on every trade

The second half of the page's opening line, say why on every trade, is the important half. Every buy and every sell goes through the Trade panel, and the panel will not submit without two things beyond the shares: how sure you are, from 1 to 5, and a note of up to 200 characters. A memo on a stock you already hold, written from the Holdings tab, asks for the same two. The log under Decisions keeps all of it, newest first, with filters for buys, sells and updates, and every row shows the date, what you did, the reason you gave, and how sure you were.

That log is the artifact a reader spends the most time on. A return can be luck and a holding can be an accident, but a dated reason cannot be either, because it was written before the outcome and it says what you were thinking when you did not yet know. Read a year of them in order and you can see a person's judgment form: which reasons came true, which ones the analyst abandoned and when, and whether the sureness moved with the evidence or stayed at 3 on every line.

The Book has one rule, and it is on the page. Say why, every time, in words you would be content to have read back to you in a year. "It is a good company" is a reason a machine can write. "Cross-border volume is recovering faster than the street expects, and the next print is in seven weeks" is a reason a person had, and it can be checked.

## Ending a run

A book does not have to run forever. At the foot of the Overview tab is a line that asks whether you are done with this approach, and it explains that a run ends with a post-mortem, a fresh book opens, and you get one reset every six months.

Ending a run is the Book's version of the post-mortem you met on the Sprint, and it asks three things instead of two. Why the run is ending, from five choices: a new approach, it ran its course, the mandate changed, a fresh start after a bad run, or something else. What it taught, in your own words, where the approach worked and where it did not and what you know now that you did not at the start. And what the next book does differently, in a sentence or two. Then you confirm, and the page closes every open position at today's prices, pins the run to your card with its ledger and its post-mortem, opens a fresh book, and uses your reset for the next six months.

The reset is rationed on purpose, so a bad quarter cannot be erased by starting over on a Tuesday. The ended run also stays on the record. It appears under "Earlier books" on the page and on your card, with a line that says every run stays on the record and recruiters see them all. Ending a run for a reason, and saying what the reason was, is part of the record.

## Why the Book is slow

So why does the Floor keep an exercise this slow next to three that settle in weeks? A well-known result in active management holds that a manager's information ratio grows with skill times the square root of the number of independent bets.[1] The Floor splits that result in two. The Sprint supplies the independent bets, many small fast claims, and that breadth is why its calibration score sharpens as quickly as it does. The Book supplies the other term. It asks for a small number of deep positions held through news, through drawdowns and through the long stretches where nothing happens at all, and it shows endurance rather than volume. A card that shows both tells a reader something neither can show alone.

The shrinkage discipline from Chapter 23 applies here too. Six weeks of Book prove nearly nothing, the ranking waits thirty days before it counts you, and the card describes a book by how many months it has been live. That is why every first-year analyst on this platform is told to start the Book early, because time is the one input you cannot add later.

## What reaches your card

The Book arrives on your Sifter Card as a tile that says how many months it has been live, and it opens onto the mandate, the holdings, and the decision log. Earlier runs are listed under the live one, each with its dates and its post-mortem. Nothing on the tile is a score, and the return is printed as context in the same way the Sprint's hit rate is.

A reader who wants to know whether you can run money rather than pick it reads that log, and a year of dated decisions with reasons attached is the hardest thing on the platform to fake, because it cannot be crammed for or written after the fact. The three chapters before this one were about being right at the rate you claim. This one is about running a book for long enough that a reader can see how you decide.

> **Try it on Sifter: Write the mandate first** Open the Book and write the mandate in one line before you add a single stock. Then add your stocks with a weight, a reason and a sureness on each, press "See the evidence," and compare the first line with the mandate you wrote. If they disagree, change one of them before you press "Start the book." Once it is live, never trade without a reason you would be content to read in a year. The card only begins to say something about the Book after months, so start it now.

## Sources

1. Grinold, R. C. (1989). *The fundamental law of active management*. Journal of Portfolio Management, 15(3), 30-37 https://doi.org/10.3905/jpm.1989.409211