Skewed data · median v mean
Half our games get 4% of the answers
Should we build more games, or help people find the ones we've got?
Real data. Real decisions.
We know. But every answer you give on this site lands in a spreadsheet, and we use it to decide what to fix, what to build and what to bin. Here's how, with the working shown.
I didn't enjoy statistics at university. Even with a maths degree, stats was the bit that never clicked. I disliked it so much that in my final year I took Quantum Mechanics just to avoid it. If you know anything about Quantum Mechanics, you'll know how bad that makes stats sound. Would I have chosen differently if my lecturers had used real numbers with real context? Probably.
Then I built this site, and it started producing data: thousands of answers, none of it identifying a student. I've had to work hard to tell the numbers that matter from the noise. When is a trend not a trend? When is a difference actually significant?
So this section is the version I wish I'd been taught: real numbers, real decisions, and a nudge to think before you punch the buttons on a calculator. Computers are better at calculating. It still takes a human (for now) to make sure the right numbers go in.
Jon
Skewed data · median v mean
Should we build more games, or help people find the ones we've got?
Inference from limited data · Core Maths
Our most-answered game is a times-tables game. Is that classes, or a few keen players at home? We never know who you are. We do know the game, the time and roughly which country. Can you work it out from that?
Outliers · mean v median
Our busiest month looked like growth. Was it?
Correlation · lurking variables · Core Maths
Our weekly answers rise with the share of positive flu tests in England. Does one cause the other?
More coming soon.
Every case comes with the raw data as a CSV and questions for your class, from Year 9 to Core Maths.
Counts only: which game, right or wrong, when. Never names, never accounts.