Practice is one of the most misunderstood words in high performance. It sounds simple because everybody knows what it means in ordinary conversation: do more of the thing, repeat it, spend time on it, work harder, accumulate attempts.
But high performance practice is not defined by repetition.
It is defined by what the repetition changes.
Practice is useful when the next attempt becomes meaningfully different from the last.
This article continues the How High Performance Works series. It follows How High Performance Foundations Work and How High Performance Diagnosis Works.
The foundations article asks what has to be stable before high performance can be sustained. The diagnosis article asks what is actually limiting the system. Practice begins after that diagnosis, because practice has a job only when the system knows what it is trying to change.
The Direct Answer
High performance practice works when the performer repeatedly attempts a task that is close enough to the current edge of capability to produce useful error, receives enough information to adjust the next attempt, revisits the skill after delay, varies the conditions enough to prevent brittle memorisation, and gradually transfers the capability into realistic performance conditions.
In simple form:
- Diagnose the constraint.
- Define the practice target.
- Choose a task that exposes the target.
- Attempt without unnecessary support.
- Observe the error or limitation.
- Receive or generate feedback.
- Adjust immediately enough to matter.
- Repeat with variation.
- Return after delay.
- Integrate into larger tasks.
- Pressure-test.
- Remove support.
- Verify transfer.
- Diagnose again.
That is practice as a control loop.
Practice Must Have a Job
The central rule of high performance practice is simple: practice should target the current constraint.
If the current problem is retrieval, practise retrieval.
If the problem is method selection, practise discrimination among methods.
If the problem is execution, practise the unstable component.
If the problem is pacing, practise under measured time.
If the problem is transfer, vary the surface while preserving the underlying structure.
If the problem is checking, practise detecting seeded errors rather than merely creating more work to check.
If the problem is performance under pressure, pressure-test only after the underlying capability is stable enough that stress exposure produces information instead of random collapse.
The diagnosis decides the job. The job decides the practice design.
Repetition Is Not the Same as Practice
Repetition makes something easier to repeat.
That is useful only if the repeated thing deserves to become easier.
A learner can repeat the wrong algebraic move until it becomes fluent. A writer can repeat weak paragraph structures. A team can repeat an inefficient approval process until everybody becomes excellent at navigating inefficiency.
High performance practice therefore contains correction inside repetition.
This is why How Error Correction Works | Turning Mistakes Into New Defaults belongs at the centre of a practice system. Repetition without correction can stabilise error. Correction without repetition can produce insight that disappears before it becomes available under real conditions.
The two have to work together.
The Practice Target
A practice target should be more precise than the performance goal.
The performance goal may be “score 85 in Mathematics.” The practice target might be “select the correct algebraic method from four similar problem types without chapter labels and complete the first setup step correctly within ninety seconds.”
The performance goal may be “write stronger essays.” The practice target might be “produce a clear causal topic sentence and two logically connected supporting sentences from an unfamiliar prompt within four minutes.”
The performance goal may be “reduce team turnaround time.” The practice target might be “make ownership explicit at intake so no work enters the queue without a named decision-maker.”
High performance practice gets leverage by making the target small enough to train and important enough to matter.
Deliberate Practice: Practice With a Target
Deliberate practice is one of the core architectures of high performance because it refuses vague repetition.
A deliberate practice cycle usually contains a defined component, a task near the performer’s current edge, observable feedback and repeated adjustment.
The performer is not merely accumulating hours. They are accumulating corrected attempts around a known limitation.
That distinction is important because the hardest part of practice is often not doing the work. It is repeatedly entering the zone where the performer’s current model is exposed as insufficient.
Comfortable practice protects competence. Deliberate practice reveals its edge.
The edge must still be learnable
Practice should not be confused with constant failure.
If the task is so far beyond current ability that the performer cannot understand why attempts fail, the practice produces noise rather than useful adaptation.
The ideal difficulty depends on the skill, the learner and the purpose of the session, but the general principle is stable: the performer should have enough capability to make a meaningful attempt and enough difficulty to expose the next limitation.
Desirable Difficulty
Some practice structures feel worse while producing better learning.
Spacing makes recall less fluent in the moment. Interleaving makes method selection harder. Retrieval removes the reassurance of visible notes. Delayed feedback sometimes forces more independent checking. Mixed practice makes the session feel less smooth than repeating one method twenty times.
This is the logic behind desirable difficulty.
The important word is desirable.
Difficulty is valuable when it increases the cognitive work that the future performance actually requires.
Random inconvenience is not training.
High performance practice asks: what kind of difficulty will make the relevant process stronger?
Retrieval Practice: Train Access, Not Recognition
A learner can recognise an answer without being able to produce it.
This creates one of the most common illusions in revision. Notes look familiar, examples make sense, highlighted pages feel known, and confidence rises. Then the exam removes the cues.
Retrieval practice trains the operation that the examination actually requires: pull the knowledge back into active use.
That can take many forms:
- Answer without notes.
- Write the method from memory.
- Explain the concept aloud.
- Sketch the process from a blank page.
- Reconstruct the argument.
- Recall key vocabulary before checking.
- Solve a question before viewing the worked example.
The purpose is not to prove that the learner can remember. It is to make remembering easier next time.
Spacing: Train the Return
Performance often fails not because the learner never understood but because the learner could not return to the knowledge later.
This is why spaced practice matters.
Immediate repetition produces fluency because the previous attempt is still active. Delayed return is more demanding because the route has to be rebuilt.
High performance practice therefore contains intentional return points.
Today.
Again after a short delay.
Again after a longer delay.
Again after other topics have intervened.
Again under mixed conditions.
The skill is becoming durable when it can survive absence.
Interleaving: Train the Decision Before the Method
Blocked practice trains execution. Interleaving trains selection.
If the entire worksheet contains one method, the learner does not have to decide which method fits. The page has already told them.
But high performance often depends on choosing among plausible options.
Interleaving inserts that decision back into practice.
The learner must recognise structure before executing procedure.
This is particularly important after individual methods have become reasonably stable. Mixing too early can create confusion. Mixing too late can produce brittle expertise that works only when the method is announced.
High performance practice therefore changes structure over time: first make a method learnable, then make selection necessary.
Feedback: Information for the Next Attempt
Feedback earns its place in practice when it changes the next attempt.
“Wrong” is feedback, but it is low-resolution feedback.
“Your method is appropriate, but the sign changes incorrectly when you expand the second bracket” is more actionable.
“Good essay” is pleasant but weak feedback.
“Your argument is clear, but the second paragraph introduces evidence without explaining how it supports the claim” gives the next attempt a job.
The strongest practice systems shorten the loop between attempt, information and correction without making the performer dependent on constant external judgement.
At first the teacher may identify the problem.
Later the teacher may only point to the location.
Later still the learner may have to locate the issue independently.
Eventually the learner detects the divergence while performing.
That progression converts feedback from an external service into an internal capability.
Immediate Feedback and Delayed Feedback
Not all feedback should arrive at the same speed.
When a learner is building a completely new procedure, immediate correction can prevent repeated error from becoming familiar.
When a skill is more mature, delaying feedback can force the performer to monitor and check independently before external information arrives.
The timing of feedback therefore depends on the practice target.
If the target is acquisition, faster feedback may be useful.
If the target is self-correction, too-fast feedback can remove the very work we are trying to train.
High performance practice is designed around the future behaviour, not around the convenience of the present session.
Error Correction: Change the Default
Recognising a mistake is not the same as correcting a mistake.
A learner can understand why an answer is wrong and still repeat the same error next week.
The practice system therefore has to make the corrected route easier to access than the old route.
A useful correction sequence is:
- Identify the exact point of divergence.
- Explain why the incorrect move was plausible.
- State the cue that should trigger the better move.
- Execute the corrected move immediately.
- Repeat on a near-neighbour example.
- Return after delay.
- Test inside a mixed task.
This is how insight becomes a new default.
Practice the Cue, Not Just the Response
Many failures occur before execution.
The performer does not recognise the cue that should activate the correct response.
A student knows how to solve simultaneous equations but fails to notice that the word problem creates two linked unknowns. A reader knows how to infer a character’s motive but fails to recognise which textual clues should trigger the inference. A team knows the escalation procedure but does not recognise when the situation has crossed the threshold that requires escalation.
High performance practice therefore pairs cue and response.
What did you notice?
What feature told you this method was appropriate?
What signal should make you stop and check?
What condition should trigger help-seeking?
Expert performance often looks fast because the expert recognises relevant cues earlier.
Practice should train that recognition directly.
Decomposition: Break the Task Apart
Complex performance can hide the weak component.
An essay may fail because of idea generation, structure, paragraph development, evidence, sentence control, vocabulary, timing or revision. A Mathematics problem may fail because of representation, method selection, retrieval, symbolic execution, checking or answer form.
Practising the whole task repeatedly can make diagnosis difficult because several mechanisms move at once.
Decomposition isolates the component.
Instead of writing another full essay, practise only introductions. Instead of completing another full paper, practise mixed method selection. Instead of running the whole workflow, rehearse the handoff where information repeatedly disappears.
Decomposition increases the number of useful repetitions around the constraint.
Recombination: Put the Task Back Together
Decomposition is powerful and dangerous.
A component can become excellent in isolation and still fail inside the full task.
The learner who writes strong standalone paragraphs may still fail to build a coherent essay. The student who manipulates algebra fluently may still fail to select the correct method in a mixed paper. The team that performs every process step correctly may still lose time at interfaces.
High performance practice therefore alternates decomposition with recombination.
Zoom in enough to repair.
Zoom out soon enough to test integration.
The component is not fully trained until it can survive the larger system.
Variation: Change the Surface, Preserve the Structure
Practice becomes brittle when every example looks alike.
A learner may not be learning the underlying mechanism. They may be learning a surface pattern.
Variation helps reveal whether the performer can recognise what stays the same beneath what changes.
Change the numbers.
Change the context.
Change the representation.
Change the order.
Add irrelevant information.
Remove a familiar cue.
Ask the same concept from a different direction.
The goal is not variety for entertainment. The goal is to train structure recognition.
Transfer: Can the Skill Survive a Different Problem?
High performance is not complete when the learner succeeds on the practice set.
The deeper test is transfer.
Can the capability survive a different question, a different context, a different representation, a different sequence or a different pressure profile?
Transfer is one reason practice should gradually become less helpful and less predictable.
The performer needs opportunities to make decisions that the original training used to make for them.
If the tutor always chooses the method, transfer is untested.
If the worksheet labels the chapter, transfer is partly untested.
If feedback arrives before the learner checks, self-correction is untested.
Transfer begins when practice removes the answer to the question, “What should I do next?”
Fluency: When Correct Becomes Cheap
High performance often requires routine operations to become cheap enough that attention can be spent elsewhere.
A fluent reader does not consciously decode every common word. A fluent algebra student does not deliberate over every sign. A strong writer does not rebuild basic sentence structures from first principles.
Practice creates fluency by reducing the cognitive cost of stable operations.
But fluency can hide error if the wrong route becomes automatic.
This is why speed should usually follow accuracy and correct selection.
Make the route right.
Then make the route fast.
Speed Practice
Speed is a legitimate practice target when the underlying method is stable.
The wrong way to train speed is to add a timer to unstable work and celebrate completion.
The better sequence is:
- Establish the correct route.
- Measure baseline time.
- Identify the slowest components.
- Train those components directly.
- Recombine.
- Introduce tighter but realistic time.
- Track whether accuracy remains stable.
High performance speed is not rushing.
It is reduced unnecessary delay.
Accuracy Practice
Accuracy practice is not merely “be more careful.”
That instruction is too vague to change behaviour.
Accuracy improves when the practice system knows where errors arise.
Do they occur at transcription? During sign changes? At unit conversion? When copying values? When interpreting a command word? During final answer formatting?
Then practice can create specific checking cues.
High performance accuracy is not produced by anxiety.
It is produced by a reliable quality-control route.
Checking Practice
Checking itself can be practised.
Give the learner work containing deliberate errors.
Ask which errors are easiest to miss.
Ask the learner to design a checking order.
Require a second method where appropriate.
Separate checking from re-reading.
A good checker does not simply repeat the original route with the same assumptions.
They look for known failure modes.
Pressure Practice
Pressure should be trained progressively.
Time limits, public evaluation, consequence and uncertainty can change attention, decision quality and memory access. The performer therefore needs exposure to those conditions before the real event.
But pressure practice should not begin before capability exists.
A useful progression is:
- Learn the skill with enough support to understand it.
- Perform independently without time pressure.
- Mix the skill with competing tasks.
- Add moderate time limits.
- Increase sequence length.
- Reduce external feedback during the attempt.
- Simulate realistic conditions.
- Review what changed under pressure.
Mock examinations are useful because they rehearse the whole system before the outcome counts.
The purpose of pressure practice is not to prove toughness.
It is to discover what the system does differently when pressure arrives.
Practice Under Fatigue
Some performances fail late rather than early.
A student is accurate for the first hour of a paper and deteriorates in the final twenty minutes. A team makes good decisions early in a long incident response and weaker decisions after repeated interruptions.
Fatigue can therefore be a practice condition.
But fatigue practice has to be dosed carefully. Deliberately exhausting the performer until quality collapses teaches little.
The useful target is to train the known late-performance vulnerability while preserving enough quality for correction.
This is why full-paper practice should often come after component stability, not before it.
Load: More Practice Is Not Always Better Practice
Student load matters because every practice system has a point where added volume stops producing useful adaptation.
Past that point, attention falls, error rises, feedback is ignored, recovery becomes incomplete and the next session begins from a worse state.
High performance practice therefore asks about dosage.
How many high-quality attempts are needed?
How much variation can the learner currently absorb?
How long can the task remain cognitively productive?
How much recovery is required before another difficult session?
The answer is not “as much as possible.”
The answer is enough to create adaptation without destroying the capacity to adapt again.
Recovery Is Part of Practice
Recovery is where a high-performance practice system proves that it is designed for repeated excellence rather than one heroic session.
A useful practice week contains different loads.
Hard sessions.
Lighter retrieval.
Review.
Return after delay.
Enough sleep and unstructured time that tomorrow’s attention is still available.
This is not softness.
It is system preservation.
Resilience is partly the capacity to return to useful practice without losing the route.
Practice and Motivation
Motivation often rises after practice becomes informative.
Vague practice can feel endless because the learner cannot see what is changing.
Targeted practice creates visible movement.
Yesterday I missed seven out of ten sign-change questions. Today I missed two.
Last week I needed prompts to identify the method. This week I can select independently.
My full-paper score has not moved yet, but the late-paper completion time has improved by twelve minutes.
These are leading indicators of improvement.
Visible adaptation gives effort a reason to continue.
Practice and Confidence
Confidence becomes useful when it is calibrated by evidence.
High performance practice therefore asks the performer to predict before checking.
How confident are you in this answer?
Which topic is least stable?
What score range do you expect?
Which question will probably consume the most time?
Then compare confidence with reality.
This strengthens metacognition.
A performer who knows the state of their own performance can allocate practice more intelligently.
Practice and Measurement
Practice needs enough measurement to know whether the intended mechanism is changing.
Useful measures can include:
- Accuracy by error type.
- Time per item.
- Retrieval success after delay.
- Method-selection accuracy.
- Prompt dependence.
- Transfer to unfamiliar examples.
- Checking success.
- Late-session decline.
- Full-task integration.
- Performance under realistic pressure.
The purpose is not to turn practice into surveillance.
It is to make adaptation visible enough to guide the next decision.
Formative assessment belongs inside practice because measurement matters most while the route can still be changed.
Blocked Practice
Blocked practice groups similar tasks together.
It is useful during early acquisition because it reduces the number of decisions the learner has to make while a new technique is still unstable.
Ten similar examples can help the learner see the common structure.
But blocked practice becomes dangerous when it is mistaken for readiness.
The learner may become excellent at executing a method after being told to use it.
That is not the same as recognising when it applies.
High performance practice uses blocked practice as a stage, not a destination.
Mixed Practice
Mixed practice increases decision load.
It asks the performer to classify before acting.
This is closer to many real performances, where the correct method is not announced in advance.
Mixed practice is particularly useful once individual methods are reasonably stable and the bottleneck has moved from execution to selection.
The transition from blocked to mixed practice should be intentional.
Too early, and the learner cannot tell whether failure comes from not knowing the method or not recognising it.
Too late, and the learner may build brittle fluency.
Open Practice and Closed Practice
Open practice allows access to notes, models, checklists or worked examples.
Closed practice removes those supports.
Both are useful.
Open practice is often better during acquisition because it reduces pointless struggle while a route is still being constructed.
Closed practice is necessary when the future performance requires independent retrieval.
The mistake is to leave the support in place so long that the performer never discovers what disappears without it.
A high performance programme removes support deliberately.
Guided Practice and Independent Practice
Guided practice lets the teacher shape attention.
What do you notice?
Which information matters?
What method are you considering?
What tells you that method fits?
Independent practice removes those prompts.
The learner must generate the questions internally.
This is why independent learning is not separate from practice. It is the end state of successful scaffolding.
The teacher gradually disappears from the decisions the learner can now own.
Practice Logs
A useful practice log records less than most students expect.
It does not need every question.
It needs the information that guides future practice.
- What was the target?
- What failed?
- What kind of error was it?
- What correction was made?
- What cue should trigger the correction next time?
- When should this be revisited?
- Did the repair survive after delay?
A practice log becomes valuable when it changes the next session.
Otherwise it is record-keeping without control.
The Mistake Bank
A mistake bank stores recurring error patterns rather than random wrong answers.
For each mistake, record:
- The task.
- The wrong move.
- The cause.
- The correct cue.
- The corrected move.
- A near-neighbour example.
- A later return date.
Over time, the mistake bank becomes a map of the performer’s previous weak links.
The goal is not to keep mistakes forever.
The goal is to retire them.
The Comfort Zone Problem
People naturally prefer the parts of practice that confirm competence.
Strong chapters get revised because they feel satisfying.
Familiar question types get repeated because fluency feels like progress.
Difficult tasks are postponed because they generate uncertainty.
This is why diagnosis has to control practice selection.
High performance practice spends disproportionate time near the constraint, not near the performer’s favourite evidence of competence.
This does not mean every session should feel unpleasant.
It means the programme cannot be designed by comfort alone.
The Plateau
Performance plateaus when the current training no longer changes the limiting mechanism.
The natural response is often more volume.
But if the constraint has moved, more of the old practice produces diminishing returns.
A learner may no longer need more chapter questions. They may need mixed selection.
A team may no longer need more training. It may need clearer decision rights.
A writer may no longer need grammar exercises. They may need argument development under time.
The plateau is often a signal to diagnose again.
Practice After Success
Success creates a new practice question: was that performance stable or accidental?
One good result is not enough evidence.
Return after delay.
Change the surface.
Remove the support.
Introduce another plausible method.
Add time.
Then see whether the success survives.
High performance practice does not chase endless perfection after success.
It verifies reliability.
Practice After Failure
Failure should not automatically trigger more repetition.
First ask what kind of failure occurred.
Was the knowledge missing?
Was retrieval weak?
Was the method misselected?
Did execution break?
Did checking fail?
Did time run out?
Did pressure change the route?
The answer determines the next practice structure.
Failure is useful only if the system extracts a better next attempt from it.
Alicia and the Practice Redesign
Alicia has been doing two full Mathematics papers every week.
Her score has stayed between 72 and 76 for a month.
The obvious response is a third paper.
Diagnosis shows something else.
Her conceptual knowledge is mostly adequate. She loses marks in three recurring places: she chooses the wrong algebraic route when two methods look similar, she makes sign errors after expansion, and she leaves the last section rushed because she spends too long restarting earlier questions.
The full papers are revealing the constraint but not producing enough repetitions around it.
So practice changes.
First, twelve short mixed classification problems. Alicia must name the method before solving. The target is method selection, not arithmetic.
Second, a ten-minute sign-control drill using carefully varied expansion problems. The target is technical stability.
Third, timed half-paper sections with planned decision points. The target is pacing and restart control.
Only after those components improve does Alicia return to full-paper practice.
Now the full paper has a different job.
It is testing whether the repaired capabilities integrate under examination conditions.
This is the core shift in high performance practice: do not ask practice to diagnose, repair, stabilise and verify everything at the same time.
The Practice Ladder
A useful practice ladder moves from support toward independence.
- Observe. Watch a correct model.
- Imitate. Reproduce with guidance.
- Execute. Perform on a familiar task.
- Retrieve. Perform without visible support.
- Select. Choose the method.
- Vary. Handle changed surfaces.
- Integrate. Use the skill inside a larger task.
- Pressure-test. Perform under realistic constraints.
- Self-correct. Detect and repair during performance.
- Transfer. Use the capability in a novel situation.
Not every skill needs every rung in exactly this order.
The ladder is useful because it shows that “can do it with help” and “can perform independently under pressure” are different states.
The 20-Minute Practice Block
Short practice can be powerful when the target is precise.
- 2 minutes: state the target.
- 3 minutes: retrieve the rule, method or cue without notes.
- 8 minutes: complete several focused attempts.
- 3 minutes: analyse errors.
- 2 minutes: repeat the corrected move.
- 2 minutes: schedule the next return.
Twenty minutes of high-resolution practice can be more valuable than an hour of undirected repetition.
The 60-Minute Practice Block
A longer block can combine repair and integration.
- 10 minutes: retrieval warm-up.
- 20 minutes: targeted deliberate practice on the constraint.
- 10 minutes: error correction and feedback.
- 15 minutes: mixed integration.
- 5 minutes: reflection, confidence estimate and return plan.
The structure keeps the session from becoming one undifferentiated hour.
The Weekly Practice Architecture
A high-performance week should not repeat the same practice format every day.
A useful pattern might be:
- Day 1: diagnose and repair.
- Day 2: retrieve and repeat with variation.
- Day 3: lighter review or another subject.
- Day 4: mixed selection and transfer.
- Day 5: timed integration.
- Day 6: full-task or mock performance where appropriate.
- Day 7: review, recovery and return planning.
The exact schedule depends on the performer, the subject and the season.
The principle is what matters: high performance needs acquisition, return, variation, integration and recovery.
Practice Across a School Year
Practice design should change as the school year changes.
Build phase
New knowledge and technique dominate. More guided and blocked practice is appropriate.
Consolidation phase
Retrieval and spacing become more important.
Integration phase
Topics mix. Selection and transfer become major targets.
Performance phase
Timed work, past papers and realistic sequencing become more useful.
Sharpening phase
Near major examinations, practice becomes selective. Repair the highest-value remaining weaknesses, keep retrieval active and avoid exhausting the system before it has to perform.
Exam preparation is strongest when the practice architecture changes with the calendar.
Past Papers: Whole-System Practice
Past papers are useful because they reproduce the examination architecture more closely than isolated chapter exercises.
But past papers are not always the best repair tool.
A full paper produces relatively few repetitions of any one weak mechanism.
If the student repeatedly loses marks through one narrow algebra error, a full paper may contain only two opportunities to correct it.
That makes past papers excellent for integration, pacing, selection and realistic performance, but often inefficient for early repair.
High performance practice uses whole papers and component practice for different jobs.
Model Answers
Model answers are useful when they make invisible decisions visible.
The weak use of a model answer is imitation without analysis.
The stronger use asks:
- What decision did the writer make here?
- What information was selected?
- What was omitted?
- How is the answer structured?
- Where does explanation add value?
- Which features are transferable to a different question?
The goal is not to copy the surface.
It is to extract the operating logic.
Mark Schemes
Mark schemes convert knowledge into credited performance.
High performance practice uses them to reveal answer form.
What does the question reward?
What evidence is sufficient?
What precision is required?
Where do students produce technically relevant information that does not earn credit?
Practice should eventually train not only knowing the subject but expressing the knowledge in a form the task can recognise.
Practice for Teams
Teams practise too, even when they do not call it practice.
Runbooks, simulations, drills, pre-mortems, incident exercises, rehearsals, tabletop scenarios and post-project reviews all create opportunities to strengthen team performance before the real event.
A high-performing team can practise:
- Decision escalation.
- Handoffs.
- Communication under pressure.
- Role substitution.
- Quality checks.
- Failure recovery.
- Operating playbooks.
- Response to ambiguous information.
The same principle applies as in education: practise the interface that is likely to break.
A team does not become high-performing merely because each individual has practised their own task.
The edges between people must be practised too.
Practice for Institutions
Institutions practise through procedures, exercises, audits, simulations, drills, training cycles and review systems.
The challenge is scale.
One person can correct immediately. An institution has to preserve the correction across teams, shifts, locations and personnel changes.
That requires turning useful learning into standards, documentation, training and interfaces without making the system too rigid to adapt.
Institutional high performance therefore includes practice at the level of memory.
Can the organisation retain what one team learned?
Can a new person inherit a better route rather than repeat the old failure?
This is one of the bridges between high performance and civilisation: capability becomes more powerful when it survives the individual.
The Practice Audit
A useful practice audit asks:
- What exactly is this practice trying to improve?
- Is that still the current constraint?
- Does the task expose the target clearly?
- Is the difficulty high enough to create useful error?
- Is the difficulty low enough that correction is possible?
- Does feedback arrive in time to change the next attempt?
- Is the learner repeating the corrected route?
- Is there enough delay to test retention?
- Is there enough variation to test transfer?
- Is support being removed over time?
- Does practice eventually resemble the real performance?
- Is the workload absorbable?
- Is recovery adequate?
- How will we know when the practice target has been achieved?
If the practice system cannot answer those questions, volume may be hiding design weakness.
The Green, Amber, Red Practice Check
A simple practice status system can help decide the next move.
Red: the skill is unstable even with support. Stay close to acquisition and correction.
Amber: the skill works under familiar conditions but fails after delay, variation, mixing or mild pressure. Stabilise and transfer.
Green: the skill is reliable enough under realistic conditions to stop dominating practice time. Maintain it while moving attention to the next constraint.
The purpose is not to make every skill permanently Green.
The purpose is to keep practice aligned with where the system actually needs work.
When to Stop Practising a Component
Practice should stop being intensive when the component is no longer the best use of scarce training time.
Useful exit criteria include:
- Accuracy is stable across repeated attempts.
- The skill survives delay.
- The skill survives mixed tasks.
- The learner can recognise when to use it.
- The skill survives realistic timing.
- The learner can detect common errors independently.
- The capability transfers into larger tasks.
At that point the component moves from repair to maintenance.
High performance practice knows how to stop as well as how to start.
The Independence Test
The final question is whether the performer can increasingly design useful practice for themselves.
Can they identify the weak area?
Can they choose a task that exposes it?
Can they distinguish recognition from retrieval?
Can they notice when practice has become too easy?
Can they detect when fatigue makes further work unproductive?
Can they use feedback without depending on the person giving it?
Can they decide when a component is stable enough to move on?
This is the deeper aim of high performance practice.
Not endless dependence on somebody else’s programme.
Increasing ownership of improvement.
The Practice Return
High performance practice is not about doing more.
It is about changing the right thing.
Diagnose the constraint.
Choose the target.
Design the attempt.
Expose the error.
Correct it.
Repeat the corrected route.
Return after delay.
Vary the surface.
Mix the choices.
Integrate the component.
Add pressure gradually.
Remove support.
Verify transfer.
Then diagnose again.
Do not practise what is comfortable merely because it is easy to repeat. Practise what changes the system.
Return to How High Performance Works | The System Behind Repeatable Excellence.
Previous article: How High Performance Diagnosis Works | Find the Constraint Before You Add More Effort.
Next in the series: How High Performance Feedback Works | Make the Next Attempt Better.
Properly Taught Kids Shine a Bright Light Into the Future.