Fix your proportions

What it can name

The faults, drawn

A percentage on its own tells you nothing you did not already know. These are the separable problems the measurement identifies — each with a different fix, which is the whole reason for telling them apart. Every example below is a real stroke, generated by the same code that tests the engine.

• Leaning late

  Every mark past the middle. Whatever you look at first ends up the bigger half — one habit, and one habit can be corrected deliberately.

• Leaning early

  The same rule, the other way round: you stop short of half and the second piece comes out bigger.

• Scatter

  Misses in both directions with no pattern. Nothing here to correct — this is estimating instead of comparing.

• Uprights read high

  Flat bars right, standing ones marked above centre. The commonest bias there is, and the reason eyes get drawn too high on a head.

• Lost over distance

  The short bars are fine and the long ones drift. A distance too big to take in at a glance has to be halved and halved again.

• One careless mark

  Everything else near-exact and one well out. That is attention rather than eyesight, and it must not be read as a bias.

How the number is made

What the score actually measures

Every number here is a fraction of the bar it belongs to, never a pixel count. That is the whole reason this drill exists: a circle drawn with a finger on a phone is partly a measurement of the phone, but a tick costs nothing to place on any device, so what is left in the error is your judgement — and judgement is the part that carries over to paper. Every threshold, and the evidence behind it.

The technique

Divide a distance accurately by eye

1. 01

   Stop looking for the middle

   The middle is not a thing you can see. What you can see is whether two pieces match, and that comparison is far more reliable than the estimate.

2. 02

   Place a mark, then judge the leftover

   Put the tick roughly where it belongs and ask which of the two pieces is bigger. You will nearly always have an opinion, and that opinion is usually right.

3. 03

   Move it and ask again

   One correction gets most people inside two percent. It is the second look that does the work, not the first.

4. 04

   Check the answer sideways

   Turn the page — or your head — ninety degrees. If the mark suddenly looks wrong, nothing about the drawing changed, so the error was in how you were seeing it.

Proportion is a judgement, not a skill of the hand

Almost every drawing that looks wrong for no obvious reason is a proportion problem. The lines can be confident and the shading can be careful, and it will still read as off if the head is a fraction too large or the eyes sit slightly too high. It is also the error people find hardest to see in their own work, because the same judgement that made it is the one being asked to check it.

Which makes it worth separating from everything else. The circle and line drills measure how well your hand carries out an intention — real skill, and it improves with practice, but partly a measurement of what you are drawing with. Nobody draws from the shoulder on a phone; the arc is about sixty centimetres across and the phone is not. Placing a small tick is different. It costs almost nothing to execute on a mouse, a finger or a pen, so the error that is left is not about your hand at all.

That is why this is the one score on the site that means the same thing wherever you took it. Whatever this drill tells you about your eye is true on paper too.

Why six bars, and why they are different lengths

One bar tells you how accurate somebody is. It cannot tell you anything about why, and why is the part worth having.

So the six bars are arranged as two crossed comparisons: three lie flat and three stand upright, and every one is a different length so the long ones and the short ones also split three against three. That is enough to ask two questions a single bar cannot. Does an upright distance read differently to you than a flat one? And does your accuracy survive as the two ends get further apart?

Three per group is the minimum that can answer them honestly. An earlier version of this drill used four bars, two per group, and it was wrong often enough to matter: across sixty attempts with purely random misses, seventeen produced a gap between upright and flat large enough to look like a real bias. A drill confidently explaining how your eyes work on the strength of a coincidence is worse than one that says nothing, so the gap now has to clear the scatter behind it before it is named at all — and when the evidence is not there, this page will tell you your marks simply scattered.

The bias almost everybody has

Most people put the middle of an upright bar above where it really is. The top half comes out the smaller of the two, and it happens consistently enough that you can predict it before someone draws.

It shows up everywhere once you know to look. Eyes drawn too high on a skull is the most familiar case — the true midpoint of a head, from crown to chin, runs almost exactly through the eyes, and almost every beginner puts them well above it. The same bias raises waistlines, shortens foreheads and stacks the features of a face into the top half of it.

A bias is good news, because a bias is one habit rather than a general lack of accuracy, and one habit can be corrected deliberately. What this page gives you is the number and the direction — how far yours runs, and which way. Knowing your own is worth more than any single score.

Questions