Compare two shapes, not two numbers
Comparing two things is the part of measurement people are reliably good at. This format is built on that half: one shape held at a known scale, one shape free to resize, and a score for how close the pair came. Five comparisons a day, no tape measure involved.
Ask for the height of a door in centimetres and most people stall. Put a person beside it and the same people answer at once.
Comparing is not a shortcut around measuring. It is the measurement most of us actually perform, and it has rules worth knowing before the first round.
People are consistent at ratios and unreliable at absolutes. Shown two lines, almost anyone can say which is longer and roughly by how much; asked to draw a line of exactly nine centimetres, the same group scatters widely. A comparison puzzle leans entirely on the first skill and never tests the second. Nothing in a round asks you to name a figure, which is why it works on objects you have never held.
A comparison needs an anchor, and here the anchor is a blue silhouette locked to a known size. It does not move, scale or vary between rounds. Everything you decide is expressed against it: the target is twice the reference, or a third of it. Take the anchor away and the same round turns into a question about centimetres, which is a harder and much less interesting game.
Full marks sit inside 6% of the true size. On a two-metre door that leaves 12 cm of slack, wide enough for a rough estimate to survive. On a 30 cm object the same 6% is under 2 cm, and the margin has to be judged by eye against a shape that is already small. This is why tall objects feel kinder than short ones in the same set, and why a low total does not always mean a poor eye.
The same mechanic covers furniture, film props and coastlines, and the difficulty moves with the subject rather than the rule. Everyday objects are anchored in your own kitchen and body. Fictional props are only as good as your memory of the source, so a fan carries an advantage that has nothing to do with estimating. Geography shifts the scale by several orders of magnitude, where a country can end up smaller than a city you know well.
A set stops at five comparisons, and not because five is a round number. Precision drops once you have been staring at silhouettes for a few minutes; the tenth comparison starts being answered by fatigue instead of judgement. Short sets also keep the habit cheap, since two minutes is easy to repeat and twenty is not. The trade is that one careless round costs a fifth of the day, which is what the practice mode is for.
02
Properties of the format rather than a list of buttons.
The blue shape is pinned for the whole round. Your answer is always phrased against it, so the task stays a comparison instead of drifting into recall.
Points follow the proportional gap between your shape and the true one. Pixels, zoom level and monitor size cancel out of the calculation.
The target grows or shrinks with one gesture, without snap points or a shortlist of plausible answers. Halving and doubling are each a single move.
Revealing draws the real outline on top of your guess. One look tells you whether you overshot and by how much of the reference.
Objects, fiction and landmasses each keep their own daily set. Only the things being compared change; the fixed-plus-resizable layout stays identical.
Because scoring is relative, a small screen does not narrow your margin. A round on a phone is exactly as hard as that round on a desktop.
03
Four habits that separate a 400 day from a 250 day.
Say the true size before touching anything: a door is two metres, a mug is ten centimetres. A named anchor holds far steadier than a felt sense.
Decide whether the target looks like half, a third or twice the reference. Fractions survive a change of object; centimetres do not.
Get the pair near your fraction with one large drag, then correct in smaller steps. Overshooting in a single jump is the common way to lose a round.
Lock in once further dragging stops improving the match. Anything inside 6% pays the same, so chasing the last pixel buys nothing.
04
Three confessions from people who kept playing.
I overshoot anything small. Under twenty centimetres and my hand drags it to double size, every time.
We do one round as a warm-up: name the reference together, guess the ratio, then look at the overlay.
Africa looks manageable until you drag it over Europe. Past a thousand kilometres my sense of ratio just quits.
05
What the mechanic is, how it is scored, and where it stops being a fair test.
Drag until the pair looks right, then see the true outline laid over your guess. The full set takes about two minutes.
Start comparing