How numbers
mislead you.
An interactive field guide. Ten traps, each with a toy you can play with, a famous real case, and a way to spot it next time.
Ten ways to be fooled
Simpson's paradox
A trend that appears in every group can vanish, or reverse, when the groups are combined.Switch between the whole university and the six departments.
Base-rate neglect
A positive test for a rare condition is usually a false alarm.Set how rare a condition is and see how many positives are real.
Survivorship bias
You only see the ones that made it back.Study the damage on returning planes, then see who never came home.
Misleading axes
Chop the bottom off a bar chart and a 5% gap looks like a landslide.Drag the baseline of the chart up and down.
p-hacking
Test enough things and something will look significant by luck.Run 20 experiments on pure noise and see what “works”.
Regression to the mean
Extreme results are followed by less extreme ones, with no cause required.Pick the top scorers from round one and watch them fall back in round two.
Correlation is not causation
Things that move together are not necessarily connected.Generate random data until two lines happen to match.
Relative vs absolute risk
“Doubles your risk” could mean 1 in 7,000 becoming 2 in 7,000.Choose a baseline risk and a headline percentage and see the people.
The law of small numbers
Small samples throw up extreme results far more often than big ones.Rank 100 identical schools and see who ends up on top and bottom.
Mean vs median
The “average” income can rise while almost nobody gets richer.Drag the top earner’s income and watch both averages.
Spot the trap
A new scenario every day. Which trap is hiding in it? The scenarios are invented to be typical, not taken from real news.
Loading today’s scenario…