What is Beat the Average?+
Beat the Average is a free browser financial literacy test in which every question is weighted by how many people actually get its knowledge point right. Ten questions are dealt from a bank of sixteen, each worth between 60 and 200 points depending on how rare the knowledge is, and a benchmark line advances alongside you at the published correct rates. A round takes two to two and a half minutes and needs no account.
Where does the average line come from?+
It is composed from three published surveys, not from other players and not from this site. The rates come from the OECD/INFE 2023 international survey of adult financial literacy across 39 countries and economies, the TIAA Institute–GFLEC Personal Finance Index 2026 covering 3,602 US adults, and the People's Bank of China consumer financial literacy survey of 2021 with 118,775 valid responses. The line is the sum of each question's weight multiplied by its published rate, which makes it a model built from those figures rather than any one survey's average respondent. No site-wide player average exists anywhere in this game.
The line mixes three different groups of people — is that a fair comparison?+
It is a composite, and the page never pretends otherwise. One round can hold a rate measured on adults across 39 countries, one measured on US adults and one measured on Chinese consumers, so the line is a benchmark assembled from three populations rather than a figure any single group produced. Every question prints its own population and year next to the rate, which is what lets you read the line for what it is. Where a knowledge point has a rate from the original instrument, that one is used in both languages; where only one survey published a figure, that survey's population is shown to every reader.
What does the percentile on the result page actually mean?+
It is an estimated position inside a modelled distribution, not a rank among people. The model treats each of your ten questions as an independent chance of success at its published rate and works out every score that combination can produce, then reports how much of that distribution your score is above. Real respondents are not independent — someone who understands compound interest tends to understand real return too — so the true spread is wider than the model's. The figure is labelled as an estimate and as a position in the benchmark distribution everywhere it appears.
Why do some questions offer 'Don't know' and others do not?+
Because the surveys did. The eight P-Fin questions were asked with a 'Don't know' option on the card, and their published correct rates were measured under those conditions, so removing it would change what the figure describes. Those questions keep it, as a separate button that scores nothing and counts as not correct — the same way the survey scored it. The OECD true-or-false questions and the questions written for a knowledge point the survey published without its wording do not offer one.
Are these the real survey questions?+
Nine of the sixteen are, and the game says which. A question marked as the original keeps the instrument's wording, its option form and its option count. Two carry a note saying what changed: the loan-interest question was open-ended in the survey and is multiple choice here, and the time-value question pins the inflation rate so that only one answer holds. The remaining five come from a survey that published a correct rate for the knowledge point but not the question, so the question is ours — one concept, four choices — and the rate is only ever described as belonging to the knowledge point.
How is the grade worked out?+
From the ratio of your score to the line: 1.70 or more is an S, 1.40 an A, 1.10 a B, 0.85 a C and anything below that a D. Ten out of ten is always an S regardless of the ratio, because a perfect round's ratio depends on which ten questions were dealt — over the 8,008 possible sets it runs from about 1.55 to 2.01, so on an easier set a flawless round would otherwise be graded down for the deal rather than the play.
Can I send someone the same ten questions?+
Yes. The share text carries a deal link and opening it deals exactly the same ten questions, in the same order, with the same choice order. It works across languages too: the English and Chinese banks hold the same sixteen ids with the same option counts, so one link is one round in either. Pressing play again drops the link and deals a fresh round, so the round whose answers you have just read is not the one you play next. The link is not spent by being opened, though: reloading it deals the same round over again, which is what makes it a link to one particular round rather than a one-time ticket.
