Fast Isn't Always Smart: What 35,000 Math Answers Taught Us About Speed

Fast Isn't Always Smart: What 35,000 Math Answers Taught Us About Speed

Oct 13, 2026 · @Subro

When a child answers a math question in under five seconds, they're right about one time in three.

That's not a hunch. It's what we found in more than 35,000 answers on North Star's practice platform.

Answers given in under five seconds were correct 32% of the time.

Answers that took longer were correct 59% of the time.

Same platform. Same pool of kids and questions. The biggest visible difference was the clock.

We find that gap fascinating, because almost everything children see about math tells them the opposite. Timed drills. Speed rounds. "Pencils down!" The kid whose hand shoots up first gets called quick — and quick slowly turns into smart.

Speed is what math looks like from the outside. Thinking is how it works on the inside.

This post is about that gap: what our data shows, what the research says about why, when fast genuinely is smart (sometimes it is), and what we changed because of it.

The short version

• Under 5 seconds: 32% correct.
• 5 seconds or more: 59% correct.
• Under 10 seconds, difficulty barely mattered: about 41% right on easy, medium and hard questions alike.
• Instant answers rise with age: 5% of answers in Grade 2, 35% in Grade 8.
• In contests, the earliest finishers scored lowest: 37% on average.
• A second try after a miss: 59% right (from our earlier analysis).
• What we did about it: North Star gives no points for speed. None.

---

What we found in our data

Every time a student answers a practice question on North Star, we record two things: whether the answer was right, and how many seconds passed between the question appearing and the answer arriving.

For this post we looked at 35,067 of those practice answers, plus 5,656 contest papers.

The five-second cliff

[embedded content: practice_bands_by_grade.csv · 35,067 practice answers]

Two things jump out.

First, instant answers are common. About 1 in 5 practice answers arrived in under five seconds, and the typical one of those took just two seconds. That isn't enough time to read most questions, let alone solve them.

Second, the drop isn't a slope. It's a cliff. Accuracy goes from 32% to about 58% within the first few seconds, then barely moves for the next two minutes.

There's no prize for being slow. There's a big penalty for being instant.

Slow isn't the problem, either. The small number of answers that took five minutes or more were the most accurate of all. Some of those probably include a break, or a grown-up helping, so we don't read too much into them.

Earlier this year, our first look at 19,000 answers found the same pattern: 47% right under five seconds, 71% for slower answers. With almost twice the data, the gap didn't shrink. It got wider.

A fast answer doesn't know how hard the question was

[embedded content: practice_bands_by_difficulty.csv · 35,073 practice answers by question difficulty]

If fast answers came from kids who simply knew the answer, they'd be far more accurate on easy questions than on hard ones.

They weren't. Under ten seconds, easy, medium and hard questions were all answered correctly about 41% of the time.

Knowledge cares how hard the question is. A guess doesn't.

There's a second lesson in this chart. On hard questions, students who stayed with the problem for five minutes or more got it right 67% of the time, against 42% for the quickest answers. Patience paid off most exactly where the questions were toughest.

The older the student, the faster the click

[embedded content: practice_bands_by_grade.csv · grades with 1,000+ practice answers]

This one surprised us.

In Grade 2, about 1 in 20 practice answers came in under five seconds. By Grade 8, it was more than 1 in 3.

And in every grade, those instant answers were the weak ones: between 25% and 39% right, against 53% to 67% for answers that took longer.

Our data can't tell us why older students rush more. But testing researchers have seen the same thing at scale: an NWEA research brief reports that rapid guessing increases with grade level across K–12.

The youngest group, K–1 (JK, SK and Grade 1), also had plenty of instant answers (21%). We suspect many of those are taps before the question has been fully read.

In contests, the early finishers scored lowest

[embedded content: contest_accuracy_by_time_used.csv · 5,656 contest papers]

Practice is low-stakes, so we checked contests too, where scores count and every paper has a set time limit.

Students who handed in their paper after using less than a quarter of the allowed time averaged 37%. Those who used half to three-quarters of it averaged 54%. Beyond that, extra time added little.

The fastest finishers also skipped more. In K–1 and Grade 2, students who used under a quarter of the time left more than 1 in 10 questions unanswered.

So the instant-answer habit doesn't stay in practice. It follows kids into the contests that count.

Two findings from our first look

Our earlier analysis of 19,000 answers recorded two things this export doesn't:

• When students flagged an answer as a guess: 34% right.
• When students tried a question again right after missing it: 59% right on the second go. (More on that below.)

What this data can and can't tell you

We'd rather be honest about the limits than oversell the headline:

• It's a correlation, not proof. We've split the answers by question difficulty, but we haven't yet compared fast and slow answers on the very same question.
• Practice is low-stakes. Researchers find rapid guessing is more common when stakes are low. That's why the contest chart matters: the same pattern shows up where scores count.
• Very long times are messy. An answer that took 15 minutes probably includes a break, or help.
• Grades are grouped as the platform records them. K–1 combines JK, SK and Grade 1, and Grades 9 and 10 (fewer than 1,000 practice answers each) are left out of the grade chart.
• The guess figure comes from a button we've since retired (more on that below).
• Rushing and not knowing look the same from here. Skipping is free, so a fast wrong answer isn't a scoring strategy — but it could still be a corner cut on something solvable, or a guess at something that wasn't. Time alone can't tell us which.

None of those caveats flip the finding. They just mean the 32% is a strong signal, not a law of nature.

---

Why fast answers go wrong

Try this before you read on. Don't overthink it.

A bat and a ball cost $1.10 in total. The bat costs $1.00 more than the ball. How much does the ball cost?

If "10 cents" just popped into your head, you're in very good company.

It's wrong. The ball costs 5 cents. (A 10-cent ball would make the bat $1.10, and the total $1.20.)

This puzzle comes from Shane Frederick's Cognitive Reflection Test, published in 2005. He gave it to thousands of people, and large numbers of students at MIT, Princeton and Harvard gave the fast, wrong answer.

They weren't short on ability. They were short on a few seconds of doubt.

Frederick's point was that catching the error is almost the whole job. Nearly everyone who stops to check gets it right.

Your brain has autocomplete

Think about the predictive text on your phone.

It's brilliant at common phrases. Type "See you" and it offers "soon." But give it anything unusual and it confidently suggests nonsense.

A child's fast answer works the same way. It isn't calculation. It's pattern-matching:

• Two numbers and the word "altogether"? Add them.
• Looks like last week's question? Same answer shape.
• One option looks familiar? Pick it.

For routine questions, autocomplete is often right. For anything with a twist, it's confidently wrong — and math contest questions are mostly twists. (Our post on non-routine problems vs. worksheets explains why that's the point of them.)

Rapid guessing is a known pattern

Testing researchers have studied this for two decades. They call it rapid guessing: answering far faster than anyone could actually read and solve the question.

NWEA, which runs the MAP Growth tests used in many schools, found that rapid guesses were right about 25% of the time — exactly what random clicking would score on four options. Other answers were right about 50% of the time. NWEA now invalidates tests with too much of it.

Steven Wise, who led much of that work, describes a rapid guess as a choice to step out of being measured for a moment. The answer doesn't tell you what the child knows. It tells you they weren't really there.

A rapid guess isn't a wrong answer to the question. It's an answer to a different question: "How do I make this go away?"

That's why our 32% matters. Some of those fast answers are genuine — a child who simply knows 7 × 8. But our data says most aren't: an instant answer was right about as often on a hard question as on an easy one. And when students told us outright that they were guessing, they were right just 34% of the time.

One more piece of context. Skipping costs nothing on North Star — no penalty for a blank answer, and no bonus for a lucky fast guess. So a wrong answer in under five seconds isn't a rational bet against a free alternative: a child who genuinely didn't know the answer could skip it for exactly the same cost as guessing. Most didn't. That doesn't tell us which fast answers were rushed past something knowable and which were guesses at something unknown — time alone can't separate those. It does suggest the 32% is closer to reflex than strategy. Answering is the default. Skipping, even when it's free, takes a decision not to.

So why do kids rush? In our experience, a few reasons come up again and again:

• It looks familiar, so they don't read the whole thing.
• They want the next thing — the next question, the reward, the end.
• They've learned that fast means smart. Hands up first, praise first.
• Being stuck feels bad, and a guess ends the feeling instantly.
• Checking feels like admitting doubt.

None of those are about math ability. All of them are about habits — which means they can change.

---

When fast is smart

Here's the twist. Fast isn't always a problem either.

A child who knows 7 × 8 = 56 instantly is fast and right. That's not rushing. That's fluency — and it's genuinely valuable. The U.S. National Mathematics Advisory Panel concluded in 2008 that instant recall of arithmetic facts frees up working memory for the harder parts of a problem.

So the question isn't fast or slow. It's fast because of what?

Typing your own name quickly is skill. Typing a stranger's address quickly is how you get typos.

Same speed. Completely different thing going on.

Fast is a result of knowing. It's a terrible substitute for it.

Here's a simple way to read what you're seeing at home:

• Fast, right, and they can explain it → fluency. Celebrate it.
• Fast and right on facts, but fast and wrong on word problems → autocomplete is running the show on anything unfamiliar.
• Fast and wrong, again and again → rushing. The fix is a habit, not more content.
• Slow and right → thinking. Celebrate this one even more.
• Very slow and wrong → stuck. Time for a hint, or a smaller version of the problem.

One honest note. You'll read online that timed tests cause math anxiety. We'd be careful. A 2017 review of the research on time pressure, stress and children's arithmetic strategies found too few studies to draw firm conclusions. Timing can change how kids approach problems. Exactly how much it harms them is less settled than the headlines suggest.

What we can say from our own data is narrower, and more useful: on problems that need thinking, the fastest answers are the least reliable ones.

The mathematician who thought he was stupid

Laurent Schwartz won the Fields Medal — mathematics' most famous prize — in 1950.

In his autobiography, he describes spending his school years quietly convinced he wasn't intelligent. Classmates who were weaker at math often answered new questions before he did. He was slow, he wrote, because he needed to understand things completely.

His conclusion, reached near the end of school: "The fact of being quick or slow isn't really relevant." What mattered was understanding how ideas connect.

If a Fields medallist can spend years mistaking slowness for stupidity, so can a nine-year-old. It's worth telling them they don't have to.

---

The second try is where the learning lives

Here's the number in our data that we like most.

In our earlier analysis of 19,000 practice answers, there were 3,354 occasions when a student got a question wrong and then tried the same question again. 59% of the time, they got it right on the second go.

In that same analysis, a considered first answer was right 71% of the time, an instant one 47%, and a self-declared guess 34%. So a second try fell short of a careful first try, but comfortably beat a quick one.

In other words, a miss followed by a second look is worth more than a quick answer that skipped the first look.

Think of a GPS. When you miss a turn, it doesn't announce "You have failed." It says "Recalculating" — and the new route uses what just went wrong.

That's what a good second try is. The miss is information.

The first answer tells you what a child guessed. The second tells you what they learned.

Two honest caveats, because this number is easy to oversell:

• Retries are a self-selected group. The kids who come back to a question they missed are probably the more persistent ones.
• On a multiple-choice question, one wrong option is already ruled out, so part of the lift is simple arithmetic.

Even so, the pattern fits what learning research keeps finding. Struggling with a problem and then getting feedback tends to build deeper understanding than being handed the method up front — our post on non-routine problems vs. worksheets walks through that evidence.

There's one place a second try doesn't exist: a real contest. Nobody tells you an answer was wrong until the round is over.

Which is exactly why the habit matters. A child who's used to taking a second look in practice learns to take it before clicking — when it still counts.

---

What we changed because of this

Data is only interesting if it changes something. Here's what this data changed for us.

• No speed bonus. Anywhere. Plenty of quiz games award extra points for faster answers. Ours never does. A speed bonus would reward the exact behaviour that's right only about a third of the time.
• Coins reward the right things. Students earn coins for correct answers, for keeping a streak going, and for mastering concepts — not for finishing first.
• Instant answers don't earn the extras. An answer that arrives too fast to have been read doesn't collect the bonus rewards a thoughtful answer does.
• A second try still pays — just not more than the first. Getting a missed question right on the retry earns a reward, but never more than getting it right first time. Otherwise the smart strategy would be to miss on purpose.
• We stopped asking kids if they're guessing. We used to offer an "I'm guessing" button; that's where the 34% figure comes from. We retired it, because how long an answer takes already tells us the same thing — without asking a child to label themselves.

There's one more thing the data pointed us towards.

The strongest positive signal in our practice data isn't speed. It's coming back. The length of a student's day streak correlates with how far they progress (r ≈ 0.42). That's a correlation, not proof — but it's a much stronger one than anything to do with speed.

We'd rather reward the kid who comes back tomorrow than the kid who finishes first today.

---

What parents and teachers can do

You don't need software to change a speed habit. You need a few small swaps in what you say and what you notice.

1. Praise the check, not the clock. "I liked how you went back and looked again" teaches more than "Wow, that was quick!"
2. Ask "How do you know?" before "What did you get?" It makes thinking the thing that gets attention.
3. Read it twice, then say it back. Before answering, have your child say in their own words what the question is asking. Autocomplete can't survive that step.
4. Estimate first. "About how big should the answer be?" catches most wild guesses before they happen. (Our Estimation Station game is built around exactly this.)
5. Keep speed where it belongs. Short, quick fact practice is fine. Problem solving is not a race.
6. Make the second try normal. After a miss: "What will you try differently?" — not "Come on, you know this."
7. Treat a rapid guess as a signal, not a sin. If answers suddenly speed up and accuracy drops, the problem is usually too hard, too boring, or too late in the day. Shrink it, switch it, or stop.
8. Teachers: retire "first one finished." Try "best explanation" or "most surprising method" instead. You'll see different kids shine.

If your child is heading into their first contest, our guide to the right age to start math competitions has a before-during-after checklist that builds the same habits.

In math, the pause before the answer is not wasted time. It's where the answer comes from.

---

Quick answers to questions parents ask

Is it bad if my child is fast at math? No. Fast, right, and able to explain it is fluency, and that's worth having. The trouble is speed on problems that need thinking, where fast usually means skipped.

My child makes "careless" mistakes. Is that a speed problem? Often, yes. Many careless mistakes are autocomplete: the brain matched the question to something familiar and answered that instead. Reading the question twice and estimating first catch most of them.

My child is slow at math. Should I worry? Slow and right is thinking — celebrate it. If your child is slow and often stuck, the problems may simply be too hard right now. Step back a level and build up again.

How fast is "too fast"? There's no universal number. In our data, accuracy fell to 32% for answers under five seconds. NWEA flags an answer as a rapid guess when it arrives in under 10% of the typical time for that question. If most kids need 20 seconds and yours answered in 3, it's worth a gentle "How did you get that?"

Should I use a timer for math practice? For short fact review, a light timer works for some kids. For problem solving, skip it. If a timer makes your child tense or sloppy, it's costing more than it gives.

Does North Star reward speed? No. Coins and progress come from correct answers, streaks and mastering concepts — never from answering quickly.

---

The bottom line

• Answers under five seconds were right 32% of the time. Answers that took longer were right 59% of the time.
• Fast is great when it comes from knowing. It's risky when it replaces thinking.
• Instant answers get more common with age, so the habit is worth tackling early.
• The second look matters. A retry after a miss was right 59% of the time.
• That's why North Star never pays for speed — and rewards coming back instead.

Don't try to raise a fast kid. Raise a kid who knows when to slow down.

Try it this week

• Watch the pause. Do a free practice test for your child's grade together and notice which answers come fast. Ask about those ones.
• Play a slow-thinking game. Ladder Logic punishes rushing and rewards planning ahead. The first level is free, no account needed.
• Enter a contest. The next North Star round runs November 10–11, 2026, and it's free to try. Register here →

Keep reading

• Grade 3 Math Problems: What a Good One Actually Looks Like — 18 sample problems, and where some of these numbers first appeared
• Non-Routine Problems vs. Worksheets — why problems that make kids pause matter so much
• At What Age Should a Child Start Math Competitions? — and how to make the first one go well
• Caribou Contest Alternatives: 9 Online Math Competitions (US & Canada) — every option, by grade

Sources

• North Star Contests, internal data (2026): 35,067 practice answers and 5,656 contest papers (October export); an earlier analysis of 19,000+ practice answers (September).
• Frederick, S. (2005). Cognitive reflection and decision making. Journal of Economic Perspectives, 19(4), 25–42.
• Toplak, M. E., West, R. F., & Stanovich, K. E. (2011). The Cognitive Reflection Test as a predictor of performance on heuristics-and-biases tasks. Memory & Cognition, 39(7). Article
• Wise, S. L. (2017). Rapid-guessing behavior: Its identification, interpretation, and implications. Educational Measurement: Issues and Practice. Summary
• Wise, S. L., & Kuhfeld, M. (2019). What happens when test takers disengage? Understanding and addressing rapid guessing. Collaborative for Student Growth at NWEA. Research brief
• NWEA (2018). Why NWEA invalidates MAP Growth tests for rapid guessing.
• Caviola, S., Carey, E., Mammarella, I. C., & Szucs, D. (2017). Stress, time pressure, strategy selection and math anxiety in mathematics: A review of the literature. Frontiers in Psychology, 8, 1488. Article
• National Mathematics Advisory Panel (2008). Foundations for Success: The Final Report of the National Mathematics Advisory Panel. U.S. Department of Education. News coverage
• Schwartz, L. (2001). A Mathematician Grappling with His Century. Birkhäuser.

This article was written by the team at North Star Contests, a registered Canadian not-for-profit that runs math contests, practice and math games for students in JK to Grade 8. It is general information, not educational advice.