Split View: 잘한다는 것의 구조 — 자기평가, 피드백, 좁은 범위

The Structure of Being Good — Self-Assessment, Feedback, and a Narrow Scope

Introduction — This Post Is a Map

The wish to be good at something runs strong in everyone, yet what being good is actually made of is a question people ask surprisingly rarely. This post is a short answer to that question and, honestly stated, a map. The deep survey of each region is already being done by other posts, so the job here is to draw the whole structure on one sheet and point at where the deeper material lives.

If the first post in the series dealt with the force to begin and the second with the allocation of that force, this one deals with the circuit that turns force into skill. To write the conclusion first, being good is not a single lump called talent but a structure assembled from three parts. Calibrated self-assessment, a fast feedback loop, and a narrow scope.

Before drawing the map, let me clear away one misunderstanding. Being good and being recognized as good overlap, but they are not the same. Recognition is skill multiplied by visibility and by relationships, and the later terms of that multiplication were handled in the effort installment. What this post deals with is the first term, the circuit in which skill itself gets made.

1. Calibrated Self-Assessment — Starting From Knowing Where You Are

The first part of improvement is, paradoxically, not skill itself but the accuracy of your estimate of your own skill. However good the map is, if your current position is wrong then the entire route is wrong.

And this estimate goes systematically astray. In Kruger and Dunning's 1999 study, performers in the bottom 25 percent estimated themselves to be somewhere around the top 40 percent. The problem of self-assessment — including the fact that there is considerable statistical objection to the way this result is consumed in popular culture — is handled properly in the series' metacognition installment.

There are reasons the estimate goes astray. First, familiarity feels like skill. A codebase you have looked at for a long time, a tool you use often, blurs the line between knowing something well and handling it well. Second, there are no records. Few people have written down last quarter's prediction alongside the actual outcome, so most self-assessment leans on memory, and memory edits in your favour. Third, the feedback that reaches you has already been filtered. People speak gently to your face, and out of that gentle speech we remember the pleasant parts longest.

So the practical rule is a single one. Do self-assessment by prediction and scoring rather than by feeling. Writing down your score before you sit the exam and comparing it with the actual score; predicting where you will be pulled up before the code review arrives. The prediction has to be recorded for the error to be visible, and the error has to be visible for calibration to happen.

Let me give a constructed example. Two people work through the same problem book. A reads the solution, marks it as known once it makes sense, and moves on. B, the moment the problem appears, predicts one of three things first. I can solve this now, I can solve it with a hint, I cannot solve it. Then, after actually attempting it, B counts the mismatches between prediction and result. Three months later the two problem books hold different data. A's data is progress through the book; B's data is progress plus the error rate of B's own estimates. The one who ends up holding a map of their skill is B.

2. The Feedback Loop — The Circuit That Turns Repetition Into Practice

The second part is the circuit in which the result of an attempt comes back as information that corrects the next attempt. Repetition without this circuit is not practice but rote continuation. It is why ten years of work sometimes amounts to one year of experience repeated ten times. The whole picture of this problem is in the deliberate practice installment.

The cheapest supply of feedback is testing yourself. As the experiment by Roediger and Karpicke showed, testing yourself is one of the rare acts that measures skill and increases it at the same time. For code, explaining it to someone else; for writing, letting it sit a day and reading it as though it were another person's; for sport, filming yourself. All of them are devices for turning your own performance into an observable object.

Feedback has two quality axes: speed and specificity. The further apart the attempt and the result, the blurrier the link between what caused what; the blunter the result, the less the point to be corrected is pinned down. That was good is a warm thing to hear, but it puts no information at all into the loop. So there are two questions when designing a loop. How quickly can I learn the result of this attempt? And does that result tell me what to change in the next one?

It is also worth knowing that the baseline supply of feedback differs by field. Sport and games return results immediately, so the loop turns for free. In fields where results are late and blurry — writing, design, teaching, management — the loop has to be built deliberately. If skill is stubbornly refusing to grow in some field, suspecting the supply chain of feedback before suspecting your talent is the right order.

Better feedback comes from people better than you. As I wrote in the table tennis post, skill mostly grows in the seat where you lose. An enjoyable rally with a comfortable opponent maintains you but does not grow you. Discomfort is closer to a signal that the feedback loop is running.

3. A Narrow Scope — A Plan to Raise Everything Raises Nothing

The third part is setting the scope. The wish to be good naturally spreads into the wish to be good at everything, and when the front line is wide, two things collapse. Feedback blurs, and progress becomes invisible. If you cannot specify what improved, competence is not supplied either.

So the plans of people whose skill grows are generally narrow. Not I want to be good at presenting, but this month I fix only the structure of the first three minutes. My backhand is weak, so this week I count only backhand receives. A narrowed scope looks shabby, but it is the only size at which feedback is sharp and progress is observed. If you want to sketch a roadmap of what to narrow first, breaking a skill into sub-units with something like skill path is one way.

There is an order to narrowing, too. Start with the bottleneck. Find the one sub-skill that eats most into overall performance and repeat that alone. Pull apart the fear of someone who says presenting frightens them and it is usually not the whole presentation but two points, the first thirty seconds and the unexpected question, and those two points can be practised separately. Practice that improves somewhere other than the bottleneck is pleasant but barely changes overall performance. People want to practise what already works, and the real thing aims precisely at the weakness.

As a sense of size, narrow enough that a change is observable within two weeks is generally about right. If two weeks of practice show no change at all, the scope is still too wide or there is no feedback.

4. What the Three Parts Look Like Turning Together

The three parts do not operate separately; they turn as one circuit. Drawing a single revolution looks like this. First you predict. In this presentation the opening will go well and I will get stuck in the questions. Then you execute narrowly. This week's practice is question-and-answer simulation only. Then you score. Did I get stuck where I expected, or somewhere I did not expect? If it was somewhere unexpected, that point becomes the scope of the next revolution. A narrow scope sharpens feedback, sharp feedback calibrates self-assessment, and calibrated self-assessment tells you what to narrow next. That is the cycle.

There is nothing dramatic about this circuit. It is less romantic than the narrative of talent, and in exchange you can start it today. Set the size of one revolution at a week and a year becomes fifty-two chances to calibrate. The difference between someone who has turned fifty-two revolutions and someone who has merely passed through a year is, seen from outside, commonly called talent.

There is one side effect worth knowing in advance. Once the circuit starts turning, it feels for a while as though your skill has dropped. It has not actually dropped; errors that were invisible have started to become visible. The self-assessment of someone who has begun predicting and scoring usually comes down once and then rises again along with the actual measurements. Read that descending stretch as failure and switch the circuit off, and your mood recovers while your map goes blank again.

Where This Advice Does Not Apply

This map is not a map of every situation. Let me mark out three cases.

• When you do not yet know what you want to be good at — in the exploration stage, a feedback loop and a narrow scope are premature optimization. What someone who does not know which well to dig needs is not the technique of digging deep but permission to dig shallowly in several places, and that belongs to the territory of direction and small bets, covered in the next post.
• When being good is not the point — not every activity has to be an improvement project. As I wrote in practising being an amateur, one thing you do not have to be good at is a counterweight for a life all by itself. Start scoring your hobbies too and there is nowhere left to escape to.
• When you are already depleted — a feedback loop is a circuit that consumes energy. Turning the circuit harder on an empty tank is not an answer but a way of hurrying the breakdown along. The post about stopping comes first.

Closing — The Depth Is in the Other Posts

To summarize: being good is a structure made of an accurate current position, a working feedback circuit, and a front line of observable size, and when any one of the three is missing, the efficiency of the other two falls with it. A narrow scope without feedback becomes an elaborate repetition of the same mistake, feedback without self-assessment keeps only what you wanted to hear, and self-assessment without scope ends in vague confidence or vague anxiety.

Let me add one thing for balance. This is the circuit of purposeful practice, not the circuit of a whole life. Someone who spends every waking hour inside a calibration loop does not last. Distinguishing the time you turn the circuit from the time you measure nothing and simply enjoy yourself is part of the design too, because it is the time spent only enjoying that lets you endure the time spent turning the circuit.

This post is a map, so it stops here. The conditions of practice are filled in by the deliberate practice installment, the structure of skill and relationships by the table tennis post, and the traps of self-assessment by the metacognition installment. The expertise series that began on the same day is a neighbour of this map. How practice ought to be structured is handled by the practice-structure installment, and how far skill in one domain carries over to another by the transfer installment.

Self-development fundamentals series — Previous: What It Means to Work Hard · Next: When Your Purpose Is Unclear

References

All sources below were accessed on 2026-08-16.

• Kruger & Dunning (1999), Journal of Personality and Social Psychology — the mismatch between self-assessment and actual performance
• Roediger & Karpicke (2006), Psychological Science — the experiment showing that testing increases long-term memory