Tool Corner

Daily Mental Maths Trainer

Five ways to drill arithmetic: the Daily 20 everyone gets the same set of, a 60-second sprint, survival where one slip ends the run, times-tables practice in the shape of the Year 4 check, and a custom drill for one weak operation. Add your age and every run is measured against the reference pace for your age group — and against your own last five.

Built and verified by Jogeswar, MSc, PMP — Tool CornerTrains arithmetic speed only — no brain-training claims

{{ modeBlurb }}

{{ setLabel }}
{{ progressTxt }}

Twenty questions: addition and subtraction first, then multiplication, then percentages and division, then two-step problems. The clock starts when you press start.
Operations
{{ questionTxt }}
{{ clockTxt }}

Whole numbers only. Press Enter to move on — a skipped question counts as wrong.
{{ doneHeadline }}
{{ doneSub }}
{{ r.q }}{{ r.note }}
{{ resultKick }}
{{ scoreTxt }}
{{ scoreSub }}
Accuracy
{{ accTxt }}
Average per question
{{ avgTxt }}
Streak
{{ streakTxt }}
Compared with your age group
{{ indexTxt }}
{{ bandTxt }}

{{ benchLine }}

{{ coachLine }}

{{ improveLine }}

{{ histTxt }}

100 means you matched the reference pace for your age with nothing wrong. It is a reference model, not a percentile or a test score — how it is worked out.

{{ verdict }}
What next?
Percentages without a calculator
The shortcuts the tier-three questions reward.
Check your working
Percentage calculator, for the ones you missed.

What your result means

Two numbers matter, and they trade off. Accuracy is how many you got right; average time is how long each answer took, including the ones you got wrong. Rushing pushes the average down and the error count up, so a fast set with four mistakes is not better than a slower clean one.

If you have entered your age, the panel also shows an index. 100 means you matched the reference pace for your age with nothing wrong; 130 means you were about 30% quicker per correct answer than the reference; 70 means about 30% slower. It is a single number precisely because it refuses to let speed and accuracy be traded off — it divides your total time by correct answers, so a wrong answer costs you twice. The comparison is with the reference model set out below, not with other users, and it is not a measure of anything but arithmetic pace.

Why this one is different

Most arithmetic drills hand you a score with nothing to measure it against. This one puts the reference model on the page: the anchor is the Department for Education’s six-second-per-fact standard from the statutory Year 4 tables check, the adult curve comes from published processing-speed research, and this site’s own calibration is labelled as such so you can discount it. Five modes, and nothing leaves your browser.

How it works

For the Daily 20, today’s date is turned into a number and that number seeds a small deterministic generator. Because the seed is the date, the same twenty questions are produced for everyone, everywhere, without a server or an account — and tomorrow’s date produces a different set. They escalate in four tiers of five: two-digit addition and subtraction, then multiplication, then percentages and exact division, then two-step problems combining them.

  • Sprint 60 — sixty seconds, unlimited questions, difficulty climbing every six. Your score is how many you got right.
  • Survival — no clock and no limit, but the first wrong answer ends the run. There is no skip, and difficulty climbs every five.
  • Tables Attack — twenty-five single multiplication facts from the 2 to 12 tables, the same shape as the statutory Year 4 check, which is where the age model’s six-second anchor comes from.
  • Custom drill — choose which operations appear and how many questions, for the one type the coaching line points at.

Every answer in every mode is a whole number and never negative. Your age, streak, best times and index history are kept in this browser’s local storage only; nothing is uploaded.

How the age comparison works

Enter your age and each finished run is compared with a reference pace for that age. The whole model is three lines of arithmetic, and it is set out here so you can judge it rather than trust it.

base = 6 ÷ 1.9 = 3.16 s  (one times-table fact, age 20)
reference = base × difficulty weight × age factor
index = 100 × reference ÷ (total seconds ÷ correct answers)
  • The 6 seconds is not ours. It is the time the Department for Education allows per question in the statutory Year 4 multiplication tables check: 25 facts from the 2 to 12 tables, six seconds each. That is the anchor the whole model hangs from, at age 9.
  • The age factor is 1.9 at age 9, 1.0 at 20 and 1.4 at 70, straight lines between. The adult limb comes from processing-speed research: Salthouse puts the reduction in processing resource at roughly 40% between 20 and 70, and timed arithmetic is one of the tasks that tracks it.
  • There is no age-varying accuracy expectation, deliberately. Arithmetic knowledge is largely maintained across adult life — WAIS Arithmetic declines far less with age than fluid measures do, because it draws on stored quantitative knowledge. So age adjusts the expected pace and nothing else; your accuracy enters through your own hit rate, because the index divides by correct answers rather than by questions attempted.
  • The difficulty weights are ours. A times-table fact is 1.0; two-digit addition and subtraction 1.3; multiplication, percentages and division 1.5; two-step problems 2.2. The Daily 20 averages 1.625 across its four tiers. These are a calibration by this site, not a published figure, and they are the part of the model to be most sceptical about.

Because the index divides by correct answers rather than by questions asked, guessing quickly makes it worse, not better: half the answers wrong doubles your seconds-per-correct and halves the index.

How to use this trainer

  1. Pick a mode. Start with Daily 20; use Tables Attack if you are practising for the Year 4 check, and Custom drill to hammer one operation.
  2. Enter your age once, in the panel on the right. It is stored in this browser and used only to look up the reference pace — leave it blank and everything else still works.
  3. Press start, type each answer and press Enter. Whole numbers only. Press Escape to abandon a run.
  4. Skip if you are stuck — a skip scores as wrong, which beats burning thirty seconds. Survival has no skip.
  5. Read the coaching line under the index: it names the one question type that cost you the most, so the next Custom drill has a target.
  6. Come back tomorrow for a new Daily 20. Only today’s Daily 20 feeds the streak; every non-practice run feeds your index history.

How the score is worked out

accuracy = correct ÷ questions × 100
average = total seconds ÷ questions
index = 100 × reference seconds ÷ (total seconds ÷ correct)
grid: 🟩 right within 6 seconds   🟨 right, slower   🟥 wrong or skipped

Accuracy and average time are reported separately on purpose: collapsing them hides the trade-off you are managing. The index is the one place they combine, and it is built so the trade-off cannot be gamed — dividing by correct answers rather than questions asked means guessing quickly lowers it. The six-second grid threshold is a display choice for telling recall from working-out; the six seconds inside the age model is the DfE’s figure and a different thing entirely. The index is capped at 400 for display.

Example calculation

Daily 20, age 34: 18 right, 96 seconds
Accuracy = 18 ÷ 20 × 100 = 90%
Average = 96 ÷ 20 = 4.8 seconds a question
Age factor at 34 = 1 + 0.40 × 14 ÷ 50 = 1.112
Reference = 3.16 × 1.625 × 1.112 = 5.71 s
Per correct answer = 96 ÷ 18 = 5.33 s
Index = 100 × 5.71 ÷ 5.33 = 107 — around the reference for 34
Every figure here is produced by the same functions the trainer runs, and is re-checked numerically on each build. The cleanest check of the model is Tables Attack at age 9: the reference is 3.16 × 1.0 × 1.9 = 6.0 s, exactly the DfE’s six seconds per fact, so a nine-year-old who answers all twenty-five correctly at that pace scores an index of precisely 100.
Note what the average does not tell you: two questions almost certainly took twenty seconds each while a dozen took two, which is why the grid shows every question separately.

Frequently asked questions

How does the comparison with my age group work?

You enter your age and each finished run is compared with a reference pace for that age. The anchor is the Department for Education’s Year 4 multiplication tables check, which allows six seconds per single times-table fact; from there the model scales by age and by how much harder each question type is than one tables fact. An index of 100 means you matched the reference with nothing wrong. It is a reference model with its workings published on this page, not a percentile against other users, and not any kind of cognitive test.

Am I actually getting better, or just having a good day?

The panel keeps the index from your last five runs in each mode and tells you where the latest one sits against your own average across every run you have done. One run is noise: fatigue, the device you are on and which question types happened to come up all move it. A rising average over a couple of weeks is the thing to look at, and the coaching line names the one question type costing you the most so there is something specific to work on.

Which mode should I use?

Daily 20 for a mixed workout and the streak. Sprint 60 when you want pressure. Survival to build the habit of checking before you commit, because one wrong answer ends it. Tables Attack if a child is preparing for the Year 4 check, since it uses the same twenty-five-fact format. Custom drill when you already know what you are bad at — pick that one operation and nothing else.

Does everyone get the same questions?

In Daily 20, yes: the set is generated from the date itself, so every visitor on the same day gets the same twenty questions in the same order. The other four modes generate fresh questions every run, which is the point of them. Nothing is stored on a server — the date is the seed, and the same seed always produces the same set, which is what makes comparing results with someone else meaningful.

Will this make me better at maths?

It trains one narrow thing: the speed and accuracy of arithmetic you already know. That is genuinely useful for mental checking — splitting a bill, sanity-testing a quote — and it is all this claims. There is no evidence that drilling sums improves general reasoning, memory or intelligence, and this page makes no such claim.

Can I do a previous day’s set?

Yes. Add ?d=2026-08-15 to the URL, with any date, and the trainer rebuilds that day’s twenty questions exactly. Only today’s set counts towards your streak.

Related calculators

Assumptions & limitations

  • The age comparison is a reference model, not a norm. It is not a percentile, not standardised on a sample, and not a cognitive assessment. It has one external anchor (the DfE’s six seconds per tables fact) and one external shape claim (processing speed falling roughly 40% between 20 and 70); the difficulty weights in between are this site’s calibration. Treat the band, not the exact number.
  • It measures typing as well as arithmetic. A keyboard-entry test on a phone will read slower than the same person on a laptop, and that has nothing to do with their maths.
  • This trains arithmetic speed, nothing more. No claim is made — or supported by evidence — that drilling sums improves memory, reasoning or intelligence generally.
  • The date is the seed, in your local time zone. Someone a day ahead of you sees tomorrow’s set. On 25 August 2026 the daily generator was corrected so that a percentage-minus question can no longer produce a negative answer, which it did on roughly one day in eight; sets for dates before that fix no longer reproduce byte-for-byte.
  • Progress is local. Age, streaks, times and index history live in this browser only: clearing site data, or switching device or browser, loses them.
  • Difficulty is banded, not calibrated. The tiers escalate, but no attempt is made to hold difficulty constant between days, so compare like-for-like over a week rather than day to day.
  • Every answer is a whole number. That constrains the question types — no decimals, no fractions, no negatives.
A drill, not a brain-training programme — and not a test

This is a timed arithmetic drill for practice and fun. It is not an assessment, not a diagnostic tool, and not evidence of cognitive ability. Nothing here is stored on a server or shared.

The age comparison is a published reference model, not a standardised norm: it is not a percentile, it has not been validated on a sample, and a low index means you were slower than one arithmetic model expects — nothing more. If you are worried about your own or someone else’s memory or thinking, that is a conversation for a GP, not a web page.

Sources & references

The trainer generates its own questions, so there is no question set to cite. The age reference model has two external inputs, and both are named here:

  • Multiplication tables check administration guidance, Standards & Testing Agency — the statutory Year 4 check: 25 facts from the 2 to 12 tables, six seconds each. This is the model’s anchor at age 9.
  • National curriculum in England: mathematics programmes of study — the Year 4 expectation of recalling multiplication and division facts up to 12 × 12, which is what Tables Attack drills.
  • Salthouse, T. A. (1996), “The processing-speed theory of adult age differences in cognition”, Psychological Review 103(3), 403–428 — the source of the adult limb of the age factor. Cited by name and reference rather than by link: the paper has no stable free canonical URL, and linking one of the mirrored PDFs would be citing a copy rather than the work.
  • Arithmetic and aging: impact of quantitative knowledge and processing speed, Learning and Individual Differences (2010) — why there is no age-varying accuracy expectation in the model: arithmetic performance in ageing is limited by processing speed, while the underlying quantitative knowledge is largely maintained.

The mental methods the questions reward are covered in percentages made simple, fractions, percentages and ratios and mean, median and mode. Deliberately not cited: research on transfer from arithmetic drills to general cognition. This page makes no such claim, so borrowing the literature to imply one would be dishonest.

Found an error? Report it →
Last updated
Found this useful? Share it
Help someone else find this free tool.