A long project means drawing the same few characters hundreds of times, and something predictable happens. Each drawing is made partly from memory of the last one, small simplifications creep in, and after a few hundred iterations the character has quietly become a shorthand of herself. From inside the process nothing appears to have changed.

This is one of the most common quality complaints about long-running work, and it is entirely mechanical rather than a matter of effort or care. Knowing the mechanism means it can be interrupted.

Copying your own last drawing

Each new drawing is easiest to make by referring to the previous one, and each copy loses a little information. Structural detail that was observed originally becomes a remembered mark, then a simplified mark, and eventually a symbol that no longer describes anything.

Faces suffer worst because they are drawn most and because their construction is subtle. The nose becomes two dots, the jaw loses its planes, the eye becomes a shape rather than a form sitting in a socket, and none of it is visible when comparing today's drawing with yesterday's.

Drift is invisible day to day

Comparing consecutive drawings shows nothing, because the change per drawing is tiny. Comparing this month's against six months ago shows it immediately, which means the check has to be deliberate rather than incidental.

Keeping a printed or pinned copy of an early drawing where you can see it is the cheapest version of this. It is also the version most people abandon, because looking at it is mildly uncomfortable, which is precisely why it works.

Refresh from reference

The reset is to go back to observation: draw the character from life-derived reference again, rebuild the construction, and notice which parts had become symbols. That does not mean redesigning her, only re-learning what the design was describing.

A session of this every few months is enough. It is the same argument as the practice discipline in the piece on what improves figure drawing: without external checks, repetition confirms existing errors rather than correcting them.

The model sheet is not enough

Checking against a sheet catches proportion drift and not simplification, because the sheet is itself a drawing and can be copied in the same degraded way. It also cannot catch the loss of structure underneath a face, since the sheet may not show that structure at all.

That is why the sheet check described in the piece on working from a model sheet should be paired with occasional observation. One catches drift, the other catches decay, and they are different problems.

Boredom produces its own errors

Drawing the same thing repeatedly is genuinely dull, and dullness produces speed, and speed produces shortcuts. The shortcuts then become the standard version. This is a different mechanism from copying and it compounds with it.

Varying what you draw within a project helps, as does the batching approach in the piece on working faster, since a session spent entirely on one stage is less monotonous than one spent grinding through complete pages.

What readers actually notice

Readers rarely identify simplification directly. They report that later chapters look rushed, or that a character has changed, or that the early work had something the later work does not. Those are all descriptions of the same underlying decay.

They also notice consistency of energy more than consistency of proportion, which is the gesture argument in the piece on the line of action. A character who has kept her structure and lost her energy reads as worse than one who has drifted slightly and stayed alive.

There is a compensating benefit worth naming, because the picture is not all decline. Drawing a character hundreds of times produces genuine fluency: the hand learns her, and drawings that would once have taken an hour arrive in ten minutes. The trade is that fluency and decay travel together, and the same automaticity that makes the work fast is what allows the shortcuts to become invisible.

Keeping a character alive over a long project

Do not draw from the last drawing. Keep an early one visible. Return to observation every few months. Check against the sheet for proportion and against reference for structure. And watch the energy as carefully as the measurements.

Characters here are fictional and unmistakably adult, as set out in the reader agreement.