How Many Times Should You Repeat a Passage Before Repetition Becomes Practice?

Building a Piano Piece · Article Seven

A bar fails on the first attempt, just survives the second, and appears correct on the third. Should it be repeated three more times for consolidation, or left alone? Some teachers use “three correct repetitions”; others ask for ten or twenty. Some learners move on after one lucky success.

These numbers confuse three questions: can the passage succeed once, can the result be repeated, and can it be recovered after time has passed? Consecutive correctness mainly measures immediate performance. It can depend on the residue of the previous attempt, a rapid compensation for the last error, or luck. Learning must also be tested after an interval, from another starting point, in longer context or on another day.

No universal repetition count exists. A more defensible rule is:

Stop or continue according to whether each repetition generates useful feedback and whether the result can be recovered under changed conditions—not according to a magic number.

Repetition is the material of practice, not its definition. Twenty attempts without a target may contain less learning than three attempts with explicit correction.

Why a recent success is not yet evidence of learning

Motor performance often improves during a session, but immediate improvement and delayed retention are distinct measures. Learning research uses retention after time or transfer to a changed condition precisely because fluent practice can be temporary.

Spacing in piano learning is also less simple than general memory slogans imply. One experiment taught musically untrained participants a seventeen-note keyboard sequence under different practice lags and found no expected spacing advantage. Short phrases from five songs also showed no spacing effect on reproduction accuracy. This null-result study matters: “spacing is always superior” cannot be copied mechanically into every piano task.

At the same time, motor-sequence research indicates that time and sleep participate in consolidation, although debate continues over whether sleep creates additional improvement or stabilises and restores performance already attained. One study found evidence more consistent with stabilisation than extra enhancement. These experiments usually involve finger tapping, not complex repertoire. The safe practical inference is not “sleep will improve the passage automatically”, but that a post-rest test supplies evidence unavailable during continuous repetition.

Every repetition needs a comparison

Productive repetition contains four operations:

define the target
→ perform once
→ compare outcome with target
→ change one controllable factor

A target might be “the left hand prepares its landing before the beat”, “inner notes do not obscure the melody”, or “the pedal changes after the new harmony sounds”. “Make it better” is too vague to guide the next action.

If the second trial uses the same tempo, starting point and focus as the first, and the player cannot name a difference, a third trial risks becoming activity without diagnosis. Pause, listen to a recording, shorten the material, change rate or redefine the problem.

The value and limit of “three times correctly”

Consecutive successes can rule out a single accident, but research has not established three as a universal threshold. It can be a low-cost same-day check if “correct” is defined and attention remains active.

A more informative three-trial method changes the conditions:

1. Perform once in the current practice condition.
2. After a rest or work elsewhere, enter from another point.
3. Restore longer context or use a slightly different tempo.

This tests recovery rather than only immediate continuity. Failure on the third trial shows that the solution has not yet returned to the complete music.

When should repetition continue?

Continue while a defined variable is becoming more stable and attention remains available. A leap may change from random to predictable; rhythm may survive with progressively fewer external cues; fingering may become consistent from different entries; or a failure may identify exactly which preparatory movement should change next.

The reason for continuation should be “this variable is learning”, not anxiety that an arbitrary quota remains incomplete. If success is rare and causes vary from attempt to attempt, the task is probably too large. Change it before increasing the count.

When should the task stop or change?

More identical repetition has low value when:

• the player can no longer describe the difference between attempts;
• errors migrate unpredictably among several locations;
• the same error recurs with no change of rate, grouping or movement;
• “do not make a mistake” has replaced the sound goal;
• the body becomes rigid, fatigued or uncomfortable;
• consecutive success disappears from a different starting point;
• the player is waiting for one lucky passage.

Changing task is not surrender. Rest for thirty seconds, inspect the score, practise a neighbouring section, rehearse silently, or return later for retrieval. An interval allows the next attempt to become reconstruction rather than the afterimage of the previous one.

Record outcomes, not only minutes

“Practised for forty-five minutes” says little about where learning occurred. A concise record might be:

location: passage X
problem: leap prepared too late
change: include preceding beat and following beat; slower rate; look early
outcome: stable twice in three trials; failed from alternate entry
next session: one cold test, then reconnect

This need not turn the piano into office work. Thirty seconds of notes prevents the following day from becoming another guess and reveals whether yesterday’s final fluency survives in today’s first attempt.

Example: repetition and sound in Schumann’s “Träumerei”

Schumann’s “Träumerei”, Op.15 No.7, is not primarily difficult because of fast notes. The score is available through IMSLP. Its demands include melodic connection, balance of inner parts and bass, harmonic motion and flexible time.

If “correct” means only pitch accuracy, a phrase may succeed three times while producing a different melodic direction and accompaniment weight on each attempt. The target must change. On one trial, listen only to whether the melodic line crosses a substitution; on another, hear whether the bass supports without projecting; on a third, record and judge the balance from a listener’s distance.

Leave the phrase and return several minutes later from inside the preceding one. If the line continues naturally, the sound decision is acquiring recoverability. If it appears only from the habitual beginning, performance still depends on the immediate repetition chain.

Three kinds of repetition

Exploratory repetition compares candidate fingering, touch, pedal and phrase solutions; its outcomes need not yet be identical. Stabilising repetition reduces variation once a solution is chosen. Retrieval repetition occurs after a delay, from another starting point or in performance-like conditions, testing whether the solution can be found again.

Errors have different meanings in each phase. Exploration tolerates them as boundary information. Stabilisation reduces the task so that known errors are not continually rehearsed. Retrieval trials should not be repaired during the run; they are completed and diagnosed afterwards.

If exploration is treated as stabilisation, the first idea freezes too early. If exploration never ends, fingering and interpretation change daily. Without retrieval, the familiarity of the practice room is mistaken for reliability.

The answer is a stopping rule, not a number

One can set a small cap—perhaps three to five attempts—before mandatory evaluation. That is an attentional device, not a law of learning. If the target does not improve, change the task. If it improves but remains unstable, make another small round. If it can be recovered after a delay and in larger context, allocate time elsewhere and test it later.

The stopping condition is one of three things: the target survives a more demanding condition; identical repetition no longer produces distinguishable feedback; or body and attention require rest.

Practice is not the extension of one successful attempt into many more. It is the ability to make success reappear without depending on the immediately preceding attempt.

Primary sources and further reading

• An experiment finding no expected spacing effect in piano learning
• Sleep and the stabilisation of motor-sequence memory
• A conceptual framework for musicians’ error management
• Self-regulated strategies in tertiary piano practice

Continue reading: Building a Piano Piece series page

If you would like to bring these ideas about listening, understanding, and practice to the keyboard, you might try ScoreFlow, an app I developed to make score reading and daily practice flow more naturally together.