Extraplot is a free daily game about real data. Each day, five charts from public records stop at a cut-off year. You draw where the line went next, the truth inks in over your line, and you find out how well you read the world. Five charts, 1,000 points each, 5,000 for a perfect day.
It is built as a static website with no accounts and no ads. Your game lives in your browser. The charts are generated ahead of time from published datasets and shipped as static files, so the whole thing costs almost nothing to run and loads fast on a phone. The only third-party tags are Google Analytics 4 (via Google Tag Manager) and Cloudflare Web Analytics, used for aggregate traffic measurement; the game never depends on them.
Why it exists
Data is everywhere and understanding it is not. Most of us read charts passively: we glance, we nod, we remember the headline and not the shape. The draw-first format fixes that by making you commit. Once you have drawn a line, the reveal becomes feedback instead of decoration — and feedback is what changes what you believe.
The format was popularised by newspaper interactives in the 2010s, and daily puzzle games in the 2020s proved that one shared challenge a day is a durable habit. Extraplot sits between the two: the prediction of a chart interactive, the ritual of a daily game, and a scoring rule that makes both comparable.
How the charts are made
Every chart starts as a real annual time series from a public source. The build normalises the series, chooses a cut-off year so that five to twelve real years are hidden, and checks that the y-axis is padded so the truth never touches the top or bottom edge. Then it applies the rule that defines the game: it scores two lazy guesses — a flat line holding the last visible value, and a trend line carrying the recent years forward. If either would score above 450 out of 1,000, the chart is rejected.
The surviving charts are sorted into difficulty bands and scheduled into 365 days of five slots, warm-up to bold call, with no repeated category in a day and no repeated chart in the season. The whole bank is 1,825 charts. Everything is generated offline and committed as static JSON, so the game is deterministic and reproducible.
Design and data philosophy
- Real data only. No synthetic curves, no invented points, no composite series pretending to be a single record.
- One source per chart, cited at reveal with a link and a licence. Definitions are kept consistent within a chart.
- Honest about uncertainty. Public data is revised; charts with ambiguous definitions are dropped rather than patched.
- No dark patterns. No timer, no ads, no paid streak repair, no notification begging. Play when you want.
- Accessible by default. Keyboard play, screen-reader readouts, reduced motion and contrast that survives both themes.
Sources we credit
- World Bank, World Development Indicators — CC BY 4.0.
- Our World in Data — CC BY 4.0 per dataset.
- US Social Security Administration baby names — public domain US government work.
- Wikimedia pageview statistics — CC0.
- Published industry and company figures, linked individually on each chart's reveal.
The site itself
The look is warm white and rounded: Outfit headings, soft taupe cards, one lime accent for actions and a pink line for the player's drawing. It is meant to feel like a friendly game you can pick up in a coffee break, not a dashboard. The type is open-licensed and served from this domain, the icons are original, and the only third-party requests are the analytics tags described in the privacy policy.
Extraplot is free and stays free. There are no accounts and no personal data to sell, so there is nothing to monetise beyond the game itself. If you find a data error, a bug or an accessibility problem, tell us — corrections are the most useful thing a player can send.
Read the privacy policy for exactly what is (and is not) stored, or the terms for the boring but necessary part.