Most data journalism asks you to read a finished chart. You look at the line, feel that you have understood something, and move on. The You Draw It format asks you to do the harder thing first: commit to a shape before the real one appears.
The format became widely known through The New York Times' Upshot section, which began running “You Draw It” interactives in 2015. The first well-known example asked readers to guess the relationship between family income and a child's chance of attending college; readers drew their guess, then saw the real line. The surprise was the point. Drawing forced a prediction, and the prediction made the data stick.
Why drawing first actually works
Reading is fast and forgiving. You can look at a chart, nod, and retain almost nothing. Drawing is slow and specific: to draw a line you have to decide a direction, a magnitude and a shape for every year. Those decisions are exactly the beliefs you did not know you had.
There is a second effect. Once you have committed a guess, the reveal becomes feedback rather than information. A number you guessed and got wrong is far more memorable than a number you were shown. That is the reason the format survived beyond a single article: it is a teaching device disguised as an interactive.
What the format is not
Drawing a line is not a test of artistic skill, and it is not a trivia quiz. You do not need to know the exact number — you need to know the kind of curve. A cautious player who understands saturation, mean reversion and shock years will beat a player who knows a lot of facts but insists every line goes up.
It also should not pretend that every chart is predictable. Some series are genuinely unknowable in advance: a platform launch, a pandemic, a war, a viral fad. In Extraplot those charts are labelled bold call, and the reveal is allowed to be surprising. A game where every chart could be deduced would be a spreadsheet exercise, not a game.
From newsroom feature to daily game
The original You Draw It pieces were editorial events: a few charts, published occasionally, each with a full article attached. They worked because they were rare and hand-made. Turning the idea into a daily game changes the constraints entirely.
A daily game needs a chart bank large enough that nobody sees the same chart twice in a season; a scoring rule that is fair across wildly different units; a difficulty ladder so the first chart of the day is not a wall; a data pipeline that rejects charts whose answer is guessable by a flat line; and an archive so late arrivals and missed days are not punished. Extraplot is built to those constraints.
How Extraplot differs from a one-off interactive
| A one-off You Draw It feature | Extraplot | |
|---|---|---|
| Frequency | Every few months | Five charts every day |
| Scoring | Usually none, or a rough comparison | 1,000-point curve per chart, 5,000 per day |
| Difficulty | Hand-picked for a story | Warm-up to bold call, filtered to reject easy charts |
| Feedback | The reveal's article | Score, average miss, house lines and a short what-happened note |
| Replay | The article stays online | A full archive plus practice mode |
| Commitment | A few minutes, once | About five minutes a day, optionally |
The most important difference is the house lines. A one-off feature only needs a reveal. A game needs a way to tell whether your draw was actually good, and a standard to beat. Extraplot shows the flat line and the recent-trend line at reveal, scored with the same curve as you. Beating them on all five charts of a day is a badge; beating them regularly is the mark of someone who reads data well.
An honest note on originality
Extraplot is an independent game. It is not affiliated with The New York Times, with GuessDash, or with any of the data providers whose public records it plots. The You Draw It format is widely imitated and the underlying idea — predict first, reveal second — is older than any single implementation.
What is original here is the packaging: the chart bank, the scoring curve, the house-line filter, the season structure and the warm, rounded, friendly look. The goal is not to copy a feature. It is to make the pleasure of drawing a guess and being told the truth available every morning, for free, in about five minutes.
What makes a good draw in this genre
- Start with the mechanism: is this a share, a count, a price or a rate? Each behaves differently near its limits.
- Look at the visible slope, then ask whether it is sustainable. A line that doubled every year will not keep doubling for twelve more.
- Check for known turning points around the cut-off. A series cut off at 2019 is a different question from one cut off at 2014.
- Aim for the average, not the anecdote. A single dramatic year inside an otherwise ordinary window is usually a mistake.
- Treat bold-call charts as permission to surprise yourself. They are labelled that way because the honest answer was surprising.