Keep a forecast journal

Everything in this book comes down to one simple habit, and now it's time to make it real. The superforecasters weren't smarter than the loud experts on TV. They just did one thing the experts didn't: they kept score. On paper. Dated. Honest. Every call written down when they made it, and checked when the result came in.

Here's why that matters so much. Think about playing a video game without the high-score screen, or a football match without the scoreboard. Ask yourself afterwards how you did, and your memory will happily tell you: "amazing, obviously." It replays the screamer you scored and skips the three sitters you missed. Memory is a sneaky editor. It keeps the wins in bright colour, quietly deletes the losses, and rewrites the near-misses as if you'd nailed them all along.

That's why almost everyone thinks they're a good predictor - the "record" they're checking has been secretly edited to say so. The score-sheet is the cure. Paper doesn't forget the misses, and a note with a date on it can't be rewritten once you know the answer. Commit the call to paper with a number and a date, before the outcome - and before hindsight - can rewrite it.

Let's slow down on the sneakiest editor of all, because it has a name and it robs almost everyone: hindsight. Once you know how something turned out, your memory quietly reaches back and rewrites what you thought beforehand to match. A stock crashes, and you feel - honestly, sincerely - "I always had a bad feeling about that one," even though a week earlier you were thrilled to own it. A stock soars, and you remember being "sure" all along, though at the time you were nervous. It doesn't feel like lying, because you're not; your memory has genuinely edited the tape. This is why "learning from experience" so often teaches nothing: the experience you're learning from has already been doctored to make you look wise. The only defence is a note written before the answer arrives, in ink your future self can't secretly change.

Here's the everyday version. Imagine Haridya predicts the score of every cricket match her favourite team plays, but only out loud, to herself, and never writes anything down. At the end of the season, ask her how good her predictions were and she'll say "pretty good, actually." Of course she will - she's remembering the two matches she called correctly and quietly forgetting the eight she didn't. Now imagine she'd written each prediction in a notebook before the match. The notebook would tell a different, truer story: two out of ten. That gap - between the two-out-of-ten on paper and the "pretty good" in her memory - is the exact gap that quietly empties investors' accounts. The notebook doesn't make Haridya predict better. It just stops her from believing she's better than she is, which is the first step to actually improving.

This is the last chapter because it's where everything else gets tested. Starting from the base rate, changing your mind in small steps, hunting for what proves you wrong - the score-sheet is the tool that finally tells you whether all of it is actually working, or whether you just feel like it is.

The setup is deliberately plain. Every entry needs four things - and a call missing any one of them isn't a forecast, it's a mood, and moods don't go on the score-sheet.

The date. When you made the call, written before the result. This is what stops your memory sneaking in and rewriting it later.

The number. A percentage, not a word. "65%", not "I like it." The number is what lets you check your aim later - do the things you call "65%" actually happen about 65% of the time?

The target. The exact, checkable thing. Not "does well" but "sales grow more than 12%," or "margin stays above 18%," or "the stock beats the market average." Something a future fact can tick right or wrong.

The deadline. By when. "By March 2027." Without a deadline the call floats forever - always "still playing out," never actually scored.

Then two habits around the entries. First, look at it every few months - sit with the notebook, tick off what's finished, and leave the unfinished calls alone. Second, once a year, score your aim: gather all your calls where you said "80%" and ask what fraction actually came true. If your 80%s happen 80% of the time, your aim is good. If they only happen 55% of the time, you're overconfident - and now you know by exactly how much.

That's the whole machine. Four things, a check every few months, one honest reckoning a year. It's boring on purpose - the power is in actually doing it, not in it being clever.

The once-a-year reckoning deserves a closer look, because it's where the notebook turns from a diary into a mirror. The method is simple enough for a class 5 student: sort all your finished calls into buckets by how sure you'd said you were, then, in each bucket, count what fraction actually came true. All your "60%" calls in one pile - did about 6 in 10 happen? All your "80%" calls in another - did about 8 in 10 happen? That's it. You're checking whether your sureness matches reality, bucket by bucket.

Reading the mirror is where it gets useful. If every bucket's "happened" bar sits below its "said" bar - your 60s land 55%, your 70s land 58%, your 80s land 55% - you're overconfident, and now you know by how much: your sureness needs pulling down. The rarer opposite, "happened" bars above "said," means you're too timid and could back your strong calls harder. And a bucket that lands right on target tells you where your judgement is genuinely sound. You couldn't feel any of this from the inside; it only appears once the calls are sorted and counted. The gap between the two bars is the single most valuable number the whole notebook produces, because it's the one that tells you how much to trust yourself next time.

Let's run it for twelve months with real money on the line. illustrative

Start a plain notebook - or a spreadsheet, doesn't matter. Over the year you write down twenty calls, each with the four things. "3 Feb: 75% this company beats the market average by March 2027." "18 May: 60% the margin stays above 18% for the next two quarters." "9 Sep: 80% this fund beats its benchmark over the year." You feel good - most of your entries are 70%, 75%, 80%.

A year later, you add it all up. Of your twenty calls, the ones you marked 80% sure came true only 55% of the time. That gap is the whole lesson - and your memory would never have shown it to you. Memory would have handed you the three big wins and buried everything else. The score-sheet is brutally honest, and that's exactly what makes it useful: you're overconfident by about 25 points.

So you act on it. You start saying lower, more honest numbers - and, more importantly, you bet smaller. Where the old, overconfident you might have put ₹80,000 of a ₹1,00,000 idea-budget behind each "sure thing," the new you puts ₹55,000 and keeps the rest back for the misses you now know are coming. Nothing about your stock-picking got smarter. But your bet sizing got honest - and honest bet sizing is what keeps you alive through a cold streak. The notebook didn't turn you into a genius; it made your aim honest, which is worth more. The gap between how sure you felt and how often you were right is the number that changes how much you bet.

And here's why that matters more than being a better stock-picker. Two investors can pick exactly the same stocks with exactly the same hit-rate, and one slowly grows richer while the other slowly goes broke - purely because of bet sizing. The one who bets as if he's 80% reliable when he's really 55% keeps loading too much onto each "sure thing," so a normal run of misses hits him far harder than his wins can repair. The one who has scored her aim, knows she's really a 55%, and sizes every bet to that truth, survives the same cold streak with money to spare. The notebook is the only thing that tells you which of the two you are. It won't hand you better ideas - but it will stop your own overconfidence from turning perfectly decent ideas into a slow bleed, and that alone is worth every dull minute of keeping it.

Here's a small, slightly cheeky project that will teach you more about the Indian market than a year of watching business TV: keep a notebook that scores not just your own calls, but everyone else's too. Every time a confident prediction reaches you - a channel's target for a stock, a Telegram group's "multibagger" tip, a magazine's list of "top picks for the year," your own excited hunch - write one line: the date, exactly what was claimed, and a deadline. Then let time pass and mark each one. You don't need anyone's permission and you don't need to argue with a soul.

A year later, that notebook quietly does something no amount of shouting can: it turns loud, unaccountable voices into plain numbers. The channel that sounded so certain - how many of its calls actually landed? The tipster with the confident forwards - better or worse than a coin flip? Your own hunches - how honest was your sureness? In a market drenched in confident predictions that nobody ever checks, being the one person who does check is a genuine edge, and it costs nothing but a notebook and a little patience. Best of all, it slowly cures you of the most expensive Indian investing habit - trusting a voice because it's loud and confident - by replacing "I feel he's usually right" with "here is his actual record." Score the forecasters around you the same way you score yourself, and you'll stop needing to trust anyone at all.

The first trip-up is never starting - the notebook feels like homework, and the brain that hates being scored is the very same brain that most needs scoring. The second is starting but cheating: leaving out the number so nothing can be marked wrong, or writing calls so vague ("should do well over time") that every outcome counts as a win. A notebook full of un-checkable calls is just memory with extra steps.

The sneakiest trip-up is scoring the result instead of the thinking. A smart, careful call can still lose; a reckless one can still win. If you only ever ask "did it make money?" you'll learn the wrong lessons - punishing good judgement that got unlucky, rewarding bad judgement that got lucky. The notebook is there to grade your aim across many calls, not to cheer or cry over any single one. Score the calibration of your calls across the whole ledger, not the result of the one you remember best.

A forecast journal is the most honest tool in this whole book, but even honesty has fine print, so keep these three cautions.

First, a short notebook can lie with a straight face. Twenty calls, or a single year, simply isn't enough to separate skill from luck - a good run can flatter a poor judge, and a bad run can bury a sound one. So read your early score-sheet gently: it's a rough hint about your aim, not a final grade on your ability. The verdict only firms up over many dozens of calls across good markets and bad. Crowning yourself a superforecaster after one lucky year is just the old highlight-reel trick with a notebook prop.

Second, the notebook grades your aim, not your fortune. Its job is to check whether your "80%" calls really land 80% of the time - your calibration - not to cheer the calls that made money and cry over the ones that lost. A well-reasoned call can still lose to bad luck, and a reckless one can still win; if you let the notebook praise lucky wins and punish unlucky-but-sound calls, it'll teach you exactly the wrong lessons. Grade the honesty of your sureness across the whole ledger, and let single outcomes be the noise they are.

Third, a notebook measures you - it doesn't pick stocks, and it can be quietly gamed. Write vague targets and everything counts as a win; skip the losers and your aim looks perfect; the tool only works if you're honest when nobody's watching, because nobody is. And even a beautifully kept journal that proves your aim is excellent tells you only that you judge probabilities well - never what to buy. It makes you a more honest, better-calibrated reader of companies and forecasters, which is the entire and sufficient goal. This is a book about a method of reading, not a source of tips, and the last page of it points the same way as the first: keep score to stay humble and size your bets honestly, never to convince yourself you've found a sure thing.