Picture a dartboard. The bullseye is the true answer; your darts are your measurements. Two very different misses hide behind one little word — "wrong."
Step up to the lineAccuracy is hitting the true value — landing on the bullseye. Precision is your repeated tries landing close to each other — a tight little cluster, wherever it sits.
Here's the surprise: those two things are completely separate. You can be precise but wrong — four darts in a tight clump, all stuck in the corner, miles from the centre. You can also be accurate on average but scattered — darts flung all over, yet they happen to balance around the bullseye. By the end of this page you'll have thrown your own darts, watched a live readout judge them, met all four combinations, and learned why the tidy-looking miss is the most dangerous one in science.
Click anywhere on the board to throw a dart (tap three or more). Or press a classic case to scatter a ready-made pattern. The dashed line shows how far your average dart sits from the bullseye — that's accuracy. The two meters read your cluster live.
A full meter is good: tight darts fill the precision bar, a centred average fills the accuracy bar. The two move independently — which is the whole point.
Play for a moment and feel the two ideas come apart in your hands. Bunch your darts up in one corner and the precision bar fills right up — but the accuracy bar stays low, because the dashed line to the bullseye is long. Now spray them evenly around the centre: the accuracy bar climbs even though the darts are everywhere, because their average lands near the middle. Tight and centred is the dream. Everything else is a different flavour of "off."
Before we mix them, let's pin each one down on its own. Same dartboard, two different questions you can ask about where the darts went.
Accuracy asks one thing: how close is your result to the true value — the real, correct answer? On the board, that's how close you are to the bullseye. These darts are scattered, but they average out right on the centre, so they're accurate.
Precision asks something else entirely: how close are your tries to each other? It doesn't care about the bullseye at all. These darts form a tight little knot — beautifully precise — even though that knot is parked up in the corner, nowhere near the centre.
A tiny memory trick: accurate and aim both start with A — accuracy is about your aim being true. Precise and repeat go together — precision is about your repeats agreeing. One is measured against the bullseye; the other is measured against your own other darts.
Accuracy can be yes or no. Precision can be yes or no. Two coins, two sides — that makes exactly four ways a set of darts can land. Here they all are.
The dream. Your darts agree with each other and with the truth. Repeat the measurement and you keep getting the same right answer.
The sneaky one. The darts agree perfectly — they just all agree on the wrong spot. Repeating won't help; you'll confidently get the same wrong answer every time.
Messy but honest. No single dart is great, yet they spread evenly around the bullseye, so their average is close to the truth. Take more shots and average them, and you close in.
The worst of both. The darts don't agree with each other and don't agree with the truth. There's nothing reliable to hold onto here — back to the drawing board.
Take a second on the top-right case, because it's the trap. Precise but not accurate looks fantastic — neat, tidy, repeatable — which is exactly why it fools people. Tight darts whisper "I'm trustworthy," even when they're all huddled in the wrong corner. Tidiness is not truth.
Imagine a kitchen scale that always reads 5 grams too heavy. Weigh the same apple ten times and you'll get ten almost-identical numbers — gorgeously precise. But every single one is wrong by the same 5 grams. The scale isn't being random; it's being reliably wrong. A steady error that pushes every reading the same way, in the same direction, is called a systematic error.
Compare that to the other kind of mistake. If your hand wobbles a little each time you read a ruler, or a draught nudges the balance, your readings jiggle up and down around the truth — sometimes a touch high, sometimes a touch low. That's a random error, and it's the cause of scatter: the spread that makes you imprecise.
Random errors hurt your precision — they spread your darts out, but they cancel out if you take enough shots and average. A systematic error hurts your accuracy — it shifts the whole cluster off-centre, and no amount of repeating will ever average it away. You'll just collect more and more precise wrong answers.
This is why the tidy miss is the dangerous one. Scatter is honest — it looks messy, so you know not to trust any single reading, and you instinctively take more. A systematic error hides. It hands you a clean, confident, repeatable number that feels rock-solid, and never once hints that it's been quietly lying the whole time. The classic giveaway is forgetting to zero an instrument: a scale that says "5 g" with nothing on it will add 5 g to everything you ever weigh.
You're timing how long a toy car takes to roll down a ramp. You always start the stopwatch a split second after you let go, because your thumb is slow. Every time, your recorded time comes out about 0.2 seconds short — same direction, same amount. Your five readings agree beautifully (precise!), but they're all 0.2 s too small (not accurate). Averaging them just gives you a very precise 0.2 s that's still wrong.
The fix isn't more repeats. It's catching the bias: use a release gate, or get a friend to start the timer, so the error stops being baked into every single reading.
Here's the practical part. Precision and accuracy aren't just ideas — there's a real way to measure each, and they need different tools.
To check your precision, you simply repeat the measurement several times and see how much the answers wander. Tight, similar numbers mean high precision; numbers all over the place mean low precision. Notice you don't need to know the right answer for this — precision is purely about whether your darts agree with each other. That's why repeating a measurement is the single most useful habit in a school lab: it shows you, instantly, how shaky your results are.
To check your accuracy, repeating isn't enough — you need something to compare against: the true value, a known, trusted answer. Scientists call a known object a standard. Weigh a certified 100 g mass on your scale; if it reads 105 g, you've just caught a systematic error and you know your scale is 5 g heavy. Without a trusted truth to check against, you can be perfectly precise and never realise you're off.
Want to know if your darts agree with each other? Throw more darts. Want to know if they agree with the bullseye? You have to be able to see the bullseye. Precision you can test alone; accuracy needs a truth to lean on.
A dartboard is 2D, but most measurements are just one number. Here you're weighing the same object — its true mass is 100.0 g (the green line). Each press takes twelve fresh readings and drops them on the line. Switch between the four cases and watch where the cloud lands.
Twelve readings, hugging 100 g. The purple line (your average) sits right on top of the green line (the truth). This is what trustworthy data looks like.
Watch the gap between the green line and the purple line — that gap is the systematic error. In the "precise, not accurate" case the readings still huddle into a narrow tower, but the whole tower has slid sideways off the truth. Hit "measure again" all you like: the tower stays narrow and stays shifted. That's the lesson made visible — repeating sharpens precision, but it can never drag a biased cluster back onto the bullseye.
Once you can feel the difference, you'll spot it everywhere — in sport, in the lab, on your phone, in a doctor's office. Same dartboard, much bigger stakes.
An archer whose arrows cluster tightly is precise. If that cluster is off to one side, they're not accurate — but it's an easy fix: adjust the sight, and the tight group slides onto the gold. Precision first, then nudge it to the centre.
Timing a swinging pendulum by hand, your reflexes add random scatter to every reading. So you time ten swings and divide by ten — letting the random errors cancel out and pulling a precise average from messy single tries.
Before any careful weighing, scientists put a known standard mass on the balance and zero it. They're checking accuracy against a trusted truth — catching the "always 5 g heavy" trap before it poisons every result.
Your map app draws a blue dot and a circle around it. A small circle means high precision — it's confident about where you are. But if the whole map is mis-aligned, that confident little dot can still be on the wrong street: precise, not accurate.
Notice the pattern under all four. The careful move is almost never just "try harder." It's to ask the two questions separately: Do my repeats agree? (precision) and Have I checked against something I trust? (accuracy). Get tight first, then hunt down the bias that's shifting you off-centre.
In everyday talk we use "precise" to mean "exactly right." In science it means something narrower — and forgetting that leads people straight into the trap.
"My measurement is super precise, so it must be correct."
Precise only means your repeats agree with each other — it says nothing about whether they agree with the truth. A scale that's 5 g heavy gives you wonderfully precise readings that are all wrong. Precision is confidence; accuracy is correctness. They are not the same thing, and confidence without correctness is exactly how a tidy mistake hides.
"More decimal places means a better measurement."
Writing 4.728361 cm looks impressive, but if your ruler only has millimetre marks, those extra digits are made-up. Lots of decimals can dress up a wrong number in a fancy suit. A reading is only as trustworthy as the instrument behind it — not as the number of digits you can squeeze out of it.
"If I just repeat the measurement enough times, I'll get the right answer."
Repeating cancels random error, so it sharpens precision and improves an honest average. But it does nothing to a systematic error — that bias is baked into every reading equally, so averaging biased readings just gives you a very confident wrong answer. To fix accuracy you have to find and remove the bias itself.
Five clusters land on the board, one at a time. For each, decide which of the four cases it is. Friendly nudge if you miss one.
Look at where these darts landed. How would you describe this cluster?
Pick the label that fits the pattern.
Whenever you measure anything — in a lab, on a field, with an app — run it past these three.
Accuracy asks "am I near the truth?" Precision asks "do my repeats agree?" They're separate.
Precise-but-wrong is a systematic error — a steady bias. It hides because it looks so clean.
Repeats reveal precision. A known true value reveals accuracy. Get tight, then remove the bias.