Headcount · rules
How to play Headcount
Headcount gives you one population target per day — say, 10,000,000 people — and one drawing. Trace a loop anywhere on a glowing world population map, lock it in, and the game counts the real number of people living inside your shape. Score 0–100 on how close you got. Everyone in the world plays the same target.
The rules
- One target a day. Targets rotate from small towns (250k) to whole-continent numbers (250M+), always clean, memorable figures.
- Draw one loop. Any shape, any size, anywhere — ring a metro, hug a coastline, lasso an island chain. Drag inside your shape to move it, or redraw as often as you like before committing.
- Commit blind. Your official attempt shows the enclosed area but never the live population — reading the dot map is the skill.
- Then practice with live numbers. After your official lock, unlimited replays show a live people counter while you draw.
Scoring
Score = 100 × (1 − |enclosed − target| ÷ target), floored at zero and shown to one decimal. Misses are symmetric: 20% over scores the same as 20% under.
| Result vs target | Score |
|---|---|
| Exactly on target | 100.0 |
| 2% off (9.8M or 10.2M vs 10M) | 98.0 |
| 10% off | 90.0 |
| 25% off | 75.0 |
| 50% off (5M or 15M vs 10M) | 50.0 |
| Double the target or more | 0.0 |
Strategy tips
- Learn a few anchors.Greater Tokyo is ~37M, Cairo's metro ~22M, the Seoul capital region ~26M, greater London ~14M. One anchor city per continent turns guessing into interpolation.
- Brightness beats size. A tight loop around the Nile delta holds more people than a giant lasso across the Sahara. Density is the game.
- Use the area readout. The enclosed km² is always visible. Pair it with a density guess (a dense metro runs 2,000–10,000 people/km²) for a sanity check before you lock.
- Practice after committing. The live counter in practice mode is the fastest population-intuition trainer there is — tomorrow-you will thank today-you.
The data
Population comes from GHS-POP (EC Joint Research Centre, 2025 epoch), a global ~1 km grid modeled from census and satellite data. Counts are estimates by nature; the scoring curve is deliberately forgiving inside a few percent. Official scores are recomputed server-side on a finer grid than the in-browser preview, so everyone is judged by the same numbers.