Stuck, or consolidating?

Most people quit a plateau one week early.

The two plateaus look identical

Progress stops. It has been three weeks and you are no better at the thing than you were. From the inside this feels exactly one way, and it is in fact two different situations with opposite correct responses.

One is consolidation: the visible measure has stopped moving while the underlying thing is still being built, and it will move again. The other is a genuine ceiling for the current method, and continuing will produce more of the same nothing.

Telling them apart matters because the advice for one is wait and the advice for the other is change, and each is actively harmful applied to the other.

Four signs you are consolidating

None is conclusive alone. Three together usually is.

• The hard parts have stopped being hard, even though you are no better at the outcome. A phrase that took four attempts now takes one. The measure has not moved and the effort behind it has dropped, which is the signature of something being consolidated rather than stalled.
• Errors have changed character. You are making different mistakes than a month ago, even if there are the same number. New mistakes mean new territory.
• It comes back faster after a break. Skill that survives a week off is skill that has settled rather than been held in place by recent practice.
• You can do it in worse conditions. Tired, distracted, in front of somebody. Robustness grows before performance does, which is why the plateau is invisible.

Four signs you are actually stuck

Also not conclusive alone.

• The same mistake, in the same place, for weeks. Not a similar mistake. The same one. That is a method problem rather than a practice problem.
• More hours makes no difference. If doubling the time produces nothing, the constraint is not time, and adding more of it is the most common thing people do.
• You are avoiding one part of it. There is a component you have been practising around for a while, and it is almost always the component the ceiling is made of.
• You cannot say what you are working on. "Getting better at it" is not a target. Without a specific one there is nothing for practice to improve, and effort disperses.

How long to wait

A rough figure, worth having because the alternative is an arbitrary one. For most skills a consolidation plateau runs two to six weeks. Under two weeks is not a plateau, it is a bad fortnight. Past about eight weeks with no change in the signs above, the waiting has become the problem.

The most common error is quitting at four or five weeks, which is late enough to feel like a fair trial and early enough to be just before the thing moves. That is not bad luck. It is what a plateau of that length does to a person who is measuring the wrong variable.

What to change, if it is a ceiling

Not everything at once, because then you learn nothing about which part mattered.

Change the specific target first. Narrow to one component, the one being avoided for preference. Most apparent ceilings are an unpractised sub-skill wearing a general plateau.

Then change the feedback, not the effort. A ceiling usually means you cannot see what you are doing wrong, and the fix is a recording, a coach, a test, somebody else looking. More repetitions of an error you cannot perceive produce a more reliable error.