Practice the Thing You Want to Do

There is a satisfying kind of improvement that can keep you occupied for years without bringing you much closer to what you wanted.

You become excellent at the exercise. You know how to earn the points, complete the routine, anticipate the example or answer the familiar question. The score rises. The larger task remains awkward.

This can happen without fraud, laziness or stupidity. The exercise may teach something real. You may be working hard. The missing question is where the improvement goes.

If you want to interview people, a beautifully organized question list is useful only insofar as it helps an actual conversation. If you want to use a language, recognizing a word on a screen is one task; understanding it in a hurried conversation is another. If you want to analyze a problem at work, repeating a demonstrated calculation is a beginning. Choosing the right calculation in an unfamiliar case asks more.

Respect the beginning. Then find out what comes next.

Practice deserves a destination.

A remarkable result with a boundary

In 1980, researchers reported a striking case of learned memory skill. One undergraduate, identified as S.F., increased his digit span from seven to seventy-nine after more than 230 hours of laboratory practice over twenty months. He developed ways of coding numbers and organizing retrieval. The performance was extraordinary. It was also specific: when tested with letters after three months of training, he recalled only about six consonants.¹

The timing matters. The letter test did not occur at the end of the whole programme. Nor did the study show that one person’s general memory capacity had expanded elevenfold. It demonstrated how far performance on a particular task could move through learned strategies in one case.

There is plenty of ambition in the accurate version. A performance that once looked like a narrow ceiling changed dramatically. We do not need to erase the method or the boundary to find that encouraging.

Ask what your own ceiling might be made of. Are you attempting the task as a beginner while imagining that experienced people use the same method more forcefully? Perhaps they organize the material differently. Perhaps they recognize patterns you have not learned. Perhaps they use external tools because the real task permits them.

Watching a skilled person can help you identify questions, but ask them to explain the work. A smooth performance can conceal the decisions that make it possible. Where did they look first? Why did they reject one option? How did they notice the error? What would they do if the familiar cue were absent?

You may discover that the thing you called talent includes learnable procedures. You may also discover demands that require substantial time or support. Both findings improve the plan.

Find the part that breaks

Leah tries the course exercise. She understands what the final report should show. She follows the demonstrated sequence successfully. Then the columns appear in a different order, and she becomes stuck.

Her first explanation is that she forgot everything. A more useful description is that she learned a sequence without yet learning how to identify the relevant fields in a changed file.

That is a teachable distinction. The next practice can vary the layout while keeping the underlying task similar. She can explain why each field is needed before choosing it. A teacher can check whether she understands the relationship or is merely following the appearance of the example.

The same approach works as a practical question in many skills. Where does the attempt first become unreliable? At understanding the instruction, selecting a method, executing a step, checking the result or adapting when something changes?

Do not assume the most visible error is the first one. A poor ending may begin with a misunderstood question. A rushed performance may begin with overambitious preparation. A wrong answer may come from an unnoticed input rather than a lack of advanced knowledge.

Find the point you can work on. Make it small enough to inspect but keep it connected to the larger task. Practicing isolated parts forever can create another kind of mismatch. At some point the parts must be brought together under conditions resembling the use you care about.

This is where a capable teacher earns their place. They can help distinguish a foundation that needs work from a detail that can wait. They can also stop you from practicing an error with increasing fluency.

If you cannot obtain that help, use the best reliable examples and feedback available, and remain cautious about what your progress proves. A lack of instruction is a real limitation. It is not improved by pretending every learner can invent the curriculum alone.

Hours have contents

The appeal of an hours-to-mastery formula is obvious. It makes an uncertain undertaking resemble a payment plan. Put in the required amount and collect the result.

Research does not justify that guarantee. In a preregistered observational replication involving thirty-nine violinists, the two more accomplished groups did not differ significantly in their reported practice accumulated by age eighteen, while the less accomplished group reported less. Practice and skill were associated, but the study could not assign a causal value to each hour or explain every difference between players.²

This is a narrow study, not a final accounting of expertise. It is enough to make us suspicious of a universal receipt for greatness.

An hour is a container. Ask what is inside it. Repetition of something you already do well? Attention to a specific weakness? Confused struggle without correction? A realistic performance? Rest needed to make further work sensible? Discussion with someone who can see what you have missed?

These are different uses of time. Counting them can help you understand your commitment, but the count does not tell the whole story.

You may need substantial repetition. You may need a different method. You may need access to equipment that makes the right practice possible. You may need to recognize that the standard you want would require a commitment you do not wish to make. None of these answers can be obtained from a clock alone.

Be especially cautious when somebody sells certainty by converting a group finding into your personal destiny. A statistical association does not say that you will reach a particular rank after a particular number of sessions. Nor does unexplained variation authorize a confident story about your genes, character or hidden potential.

Your practical concern is the quality and relevance of the next period of work. Keep an eye on the result, not only the sacrifice.

Ask where the gain traveled

A large online study tested six weeks of cognitive training. Participants improved on the tasks they practiced, but the study found no convincing generalized benefit on the untrained cognitive benchmarks it measured. More than eleven thousand participants met the study’s testing and participation criteria, out of a much larger group who initially registered.³

The finding concerns those programmes, outcomes and participants. It does not establish that learning never transfers. It makes a valuable demand: test the thing you hope will improve.

If the product promises better performance at work, a rising game score is an incomplete answer. If a course promises confidence in real conversations, completion of familiar exercises is an incomplete answer. If your own plan promises clearer writing, the number of pages drafted is an incomplete answer until you examine whether a reader understands them.

A proxy can be convenient and still be only a proxy. The danger begins when convenience makes it the unquestioned goal.

Choose at least one check close to actual use. For Leah, that might be a new practice file with different errors and a request to explain her choices. For Ben, it might be an interview with an unfamiliar but willing guest on a topic he has prepared, with permission and clear expectations. In both cases, the check should remain appropriate to the learner’s stage.

You can create a ladder of tasks without pretending every rung is automatically crossed. First perform the practiced exercise. Then vary one important feature. Then attempt a suitable version of the real activity. At each point, observe what holds and what breaks.

Three separate results

The figure below keeps three questions apart. It is easy to collapse them when you are excited about progress or discouraged by a setback.

Scroll sideways to see all columns. Keyboard: focus the table and use the left and right arrow keys.

Text equivalent: What actually transferred

Part of the figure	Meaning
Book	Stop Worshipping Your Limits
Title	What actually transferred
Starting point	Check each claim separately.
Practised task	Did performance improve on what you rehearsed?
Nearby task	Does the skill carry to a different but related task?
Real-world outcome	Does it help with the outcome you actually value?
Reminder	Improvement on one task does not establish transfer to another.
Status	Original thinking aid. Not a diagnostic tool or validated scale.
How to read the diagram	The three fields are separate claims requiring separate evidence. Their order is a reading order, not a causal chain, required sequence or promise that transfer occurs. Equal panels do not encode measured quantities or effect sizes.

Figure 3. What actually transferred. This original comparison separates performance on the practised task, a nearby task and the real-world outcome. Record each result independently; improvement in one field does not establish improvement in the others.

Ben can use it after several practice recordings. In the first field, he notes whether he asks a follow-up during a familiar rehearsal. In the second, he notes what happens with a different guest or an unexpected answer. In the third, he asks whether the recording produces a clear, useful contribution that the programme could responsibly use.

The third field may reveal problems beyond interviewing. The sound is poor. The topic is too broad. The guest’s consent does not cover the intended use. The editing removes necessary context. These findings do not erase improvement in listening. They show that the valued outcome depends on additional work.

That is how a larger capability is assembled. Several tasks have to cooperate.

Keep the results separate enough to guide practice. If the rehearsal improves but the unfamiliar conversation does not, Ben can investigate adaptation. If both improve but the recording remains unusable, he can identify the production problem. A single confidence score would conceal these differences.

You can use the same questions in your own field without borrowing this example’s measures. What did you practice? What related task did you check? What happened where the skill actually matters?

An honest answer of not yet checked is better than an invented transfer.

Make difficulty useful

Difficulty has acquired a glamorous reputation. People tell us to leave comfort behind as though discomfort were a reliable guide to value. It is not.

Some difficult tasks are useful because they require the capability you want. Some are difficult because the instructions are poor, the equipment is broken or the setting is hostile. Some are beyond a safe level for your present preparation. The sensation of struggle does not identify which kind you are experiencing.

Ask what the difficulty is asking you to do. Is it bringing attention to a relevant decision? Is it exposing a gap you can address? Is correction available? Can you attempt it safely? Or are you simply repeating confusion with no way to tell what should change?

For a beginner, an easier version can sometimes reveal the structure of the task. For a more advanced learner, the same version may provide little new information. The useful challenge depends on what you can currently do and what you are trying to learn. There is no moral prize for selecting the most punishing exercise.

Leah does not need to process the most complicated file in the company on her first day. She needs a progression that makes the relevant operations understandable, then checks them under more realistic conditions. Ben does not need his first interview to be a hostile public encounter. He needs a chance to hear himself, notice a missed opening and try again.

Serious practice can be demanding without becoming theatrical. The work should have a reason beyond making you feel that you have paid in suffering.

Leave room for judgment

Not every part of a skill can be reduced to a repeated move. You may need to decide what matters, how to adapt, when to seek help and when a familiar procedure is inappropriate. Practice should eventually include those decisions.

A checklist can help you remember necessary steps. It cannot always choose the right goal. A worked example can show a method. It may not tell you when another method is better. A script can support a difficult conversation. It may become an obstacle if you stop hearing the person in front of you.

As you improve, ask yourself to explain a choice. Why this approach? What information would change it? What could go wrong? Which part am I uncertain about? You are trying to make the knowledge available for use, including use in circumstances that do not match the demonstration exactly.

Be willing to return to a simpler task when a gap becomes visible. That is a practical adjustment. Experienced people still rehearse foundations. The important thing is to know why you are returning and how you will check whether it helps.

Also make room for enjoyment. A person can become so devoted to correcting weaknesses that the activity loses every source of pleasure. Play a piece you like. Make something with the skill you already have. Have a conversation without grading it. A life devoted solely to identifying deficiencies would be a strange outcome for a book about possibility.

You are developing a capability so that you can use it. Use what is already available while you continue learning.

A better relationship with the score

By now, Leah’s practice results have changed. She can do the operation in the familiar file. She is slower in a new layout but can explain the problem. The instructor identifies one error she had not noticed. Her picture of the skill becomes more detailed.

She could tell herself that she is still too slow and therefore the effort means nothing. She could tell herself that the first success proves she has found her destiny. Neither account is as useful as the actual record.

The record says what she can do, where she needs help and what to practice next. It also gives her a better basis for deciding whether the work interests her.

That is an achievement in judgment as well as performance.

Keep your ambition large enough to matter and your measures close enough to reality to teach you. Let a score report on the task it measures. Refuse to make it the judge of every capacity you possess.

Then practice the thing you want to do. Find its parts. Learn a method. Receive correction. Bring the parts together. Check what travels into the world beyond the exercise.

A limit may move a little or a great deal. The movement will be more useful when you know what moved.