Most wrong answers from a calculator are not the calculator’s fault. The machine does exactly what it is told. The trouble is that “what it is told” depends on how it reads your keystrokes, how it rounds, which percentage you meant, and whether a comma is a thousands separator or a decimal point. Two people can type the same numbers into two calculators and get two different results, and both calculators can be working perfectly.
This guide explains how calculators actually work, where the traps are, and how to check a result before you rely on it. It is written for students, small-business owners, freelancers, landlords and anyone who uses a VAT, loan, percentage or unit calculator in the UK, USA, Canada, the Netherlands or elsewhere.
Key takeaways
- Test any new calculator with 2 + 3 × 4. It should show 14, not 20.
- To remove VAT, divide by 1.20 (UK) or 1.21 (Netherlands). Never multiply by 0.80.
- Canadian fixed-rate mortgages compound semi-annually, so US calculators overstate the payment.
- Round once, at the end, and check whether a comma is a decimal point.
Why two calculators can disagree
Type 6 ÷ 2(1 + 2) into a few devices and you will see both 9 and 1. Type 100 + 10% and some calculators show 110, others show 100.1, and a spreadsheet shows an error or 100.1 depending on how you wrote it. Neither result means a device is broken. Each one follows a different, documented rule.
In practice, nearly every calculator disagreement comes from one of five causes:
| Cause | What happens | Where you meet it |
|---|---|---|
| Entry logic | A basic calculator works each step as you press it; a scientific one waits for the whole expression | Phone calculator vs. scientific calculator |
| Implied multiplication | 2(1+2) is treated as tighter than ÷ on some devices, not on others |
School exams, social-media puzzles |
| Percentage convention | The % key adds, divides or does nothing, depending on the model |
Shop discounts, VAT, tips |
| Rounding | Values are rounded at different stages or with different rules | Invoices, payroll, tax returns |
| Locale | 1,250 means one thousand two hundred and fifty in the UK and US, but one and a quarter in the Netherlands |
CSV imports, cross-border invoices |
The rest of this guide takes these one at a time.
How a calculator reads what you type
There are two main kinds of entry logic.
Immediate execution (chain) logic. Most basic desk and pocket calculators, and the simple view of many phone apps, calculate each operation the moment you press the next operator. Typing 2 + 3 × 4 = gives 20, because the calculator works out 2 + 3 = 5 first, then 5 × 4.
Algebraic or formula logic. Scientific calculators, graphing calculators, spreadsheets and well-built online calculators read the whole expression and apply the order of operations. The same keystrokes give 14, because 3 × 4 is done before the addition.
Pro-tip: before trusting an unfamiliar calculator, type 2 + 3 × 4 =. If it shows 20, it uses chain logic. Use brackets or work in separate steps on that device.
Some older financial and engineering calculators use Reverse Polish Notation (RPN), where you enter numbers first and the operator after (3 ENTER 4 ×). RPN removes the need for brackets and is still popular with surveyors and engineers, but it gives nonsense if you enter it like a normal calculator.
Order of operations: PEMDAS, BODMAS, BEDMAS and Dutch “Meneer van Dale”
Different countries teach the same rule with different mnemonics:
| Country | Mnemonic | Stands for |
|---|---|---|
| USA | PEMDAS | Parentheses, Exponents, Multiplication/Division, Addition/Subtraction |
| UK | BODMAS or BIDMAS | Brackets, Orders (or Indices), Division/Multiplication, Addition/Subtraction |
| Canada | BEDMAS | Brackets, Exponents, Division/Multiplication, Addition/Subtraction |
| Netherlands | Meneer van Dale wacht op antwoord | Machtsverheffen, Vermenigvuldigen, Delen, Worteltrekken, Optellen, Aftrekken |
The order of letters misleads people in every country.
Multiplication and division have equal priority and are worked left to right. So do addition and subtraction.
PEMDAS does not mean multiplication before division, and BODMAS does not mean division before multiplication. The old Dutch mnemonic is known for the same problem, and Dutch schools now teach multiplication and division, and addition and subtraction, as equal pairs worked from left to right.
Worked example: 12 ÷ 3 × 2
- Correct (left to right): 12 ÷ 3 = 4, then 4 × 2 = 8
- Wrong (multiplication first): 3 × 2 = 6, then 12 ÷ 6 = 2
The implied-multiplication problem
Expressions like 6 ÷ 2(1 + 2) are ambiguous on purpose. Some calculators, and many physics textbooks, treat implied multiplication (a number pressed directly against a bracket) as binding more tightly, which gives 1. Others treat it as ordinary multiplication, which gives 9. There is no universal standard.
The fix is not to pick a side. It is to never write an ambiguous expression. Write (6 ÷ 2) × (1 + 2) or 6 ÷ (2 × (1 + 2)). Our scientific calculator shows how it has read every expression. Use the fraction template on a scientific calculator, or a clear formula in a spreadsheet, so the intent is visible.
Percentages: the most misused key
Percentages cause more real-world money errors than any other operation. Learn these four patterns and you will avoid nearly all of them.
1. Adding or removing a percentage
To add 20%, multiply by 1.20. To remove 20% that was already added, divide by 1.20. Do not multiply by 0.80.
- £100 + 20% VAT = £100 × 1.20 = £120
- Removing VAT from £120: £120 ÷ 1.20 = £100 (correct)
- £120 × 0.80 = £96 (wrong — £4 short)
This mistake is common on invoices, expense claims and price-list updates. If you hand-calculate net prices from gross prices, check one line by adding the tax back on.
2. Percentage vs. percentage points
If a mortgage rate moves from 4% to 5%, it has risen by 1 percentage point, which is a 25% increase in the rate. Journalists and marketers often blur the two. When you compare loans, savings rates or polls, say which one you mean.
3. Successive changes do not add up
A price that rises 10% and then falls 10% does not return to where it started: 1.10 × 0.90 = 0.99, so it ends 1% lower. Two successive 20% discounts are a 36% discount (0.8 × 0.8 = 0.64), not 40%. Multiply the factors; never add the percentages.
4. “What percentage is X of Y?” and “percentage change”
- Percentage of: X ÷ Y × 100. 45 out of 60 is 45 ÷ 60 × 100 = 75%.
- Percentage change: (new − old) ÷ old × 100. From 80 to 92 is (92 − 80) ÷ 80 × 100 = 15%.
The base is always the old value. Using the new value as the base is the single most common error in sales reports.
What the % key actually does
On many basic calculators, 100 + 10 % gives 110 because the key is a shortcut for “add 10% of the first number”. On others, % simply divides by 100, so the same keystrokes give 100.1. In Excel and Google Sheets, =100+10% gives 100.1, because 10% is just the number 0.1. For anything that matters, avoid the key and type the multiplier yourself: 100 × 1.10, or use a percentage calculator.
VAT and sales tax by country
Tax calculators are simple multipliers, but only if you use the right rate and know whether the price shown includes tax.
| Country / region | Tax | Standard rate | Shown in shelf prices? | Notes |
|---|---|---|---|---|
| United Kingdom | VAT | 20% | Yes (consumer prices) | Reduced rate 5%; zero rate 0% on many foods and children’s clothes |
| Netherlands | BTW (VAT) | 21% | Yes | Reduced rate 9%; short-stay accommodation moved from 9% to 21% on 1 January 2026 |
| Canada — GST provinces | GST (+ provincial PST in some) | 5% federal | No, added at checkout | Alberta and the territories charge GST only |
| Canada — Ontario | HST | 13% | No | Combined federal and provincial |
| Canada — Nova Scotia | HST | 14% | No | Reduced from 15% on 1 April 2025 |
| Canada — New Brunswick, Newfoundland and Labrador, PEI | HST | 15% | No | |
| Canada — Quebec | GST + QST | 5% + 9.975% | No | Both applied to the pre-tax price |
| United States | State and local sales tax | No federal rate | No, added at checkout | Combined rates vary by state, county and city; a few states have no statewide sales tax |
Use our VAT calculator, GST/HST calculator or US sales tax calculator to apply these rates. Rates change, so check the official source before you file or invoice: GOV.UK for UK VAT, the Belastingdienst for Dutch BTW, the Canada Revenue Agency’s GST/HST rate page, and your state’s department of revenue in the US.
Worked examples:
- Netherlands: €250 excluding BTW × 1.21 = €302.50. Reverse: €302.50 ÷ 1.21 = €250.
- Ontario: a $49.99 item × 1.13 = $56.4887, which rounds to $56.49 at the till.
- Quebec: $100 × 0.05 = $5.00 GST, $100 × 0.09975 = $9.975 QST (rounds to $9.98). Total $114.98. Both taxes are calculated on the $100, not one on top of the other.
Pro-tip for freelancers invoicing across borders: an EU business invoicing another VAT-registered EU business often applies the reverse-charge mechanism, so no Dutch BTW appears on the invoice. A simple VAT calculator will not know this. Confirm the rule for your situation with your tax adviser or the tax authority before relying on a calculator.
Interest, loans and mortgages
Money calculators look simple, but they hide assumptions about compounding, timing and fees. Two loans with the same headline rate can cost different amounts.
APR, AER, APY and the nominal rate
| Term | Used mainly in | What it means |
|---|---|---|
| Nominal rate | Everywhere | The stated yearly rate before compounding is considered |
| APR | UK, US, Canada (loans and cards) | The yearly cost of borrowing. In the UK and the US it includes certain required fees, so it can be higher than the interest rate |
| AER | UK (savings) | The effective yearly return after compounding |
| APY | US (savings) | Same idea as AER |
| JKP (jaarlijks kostenpercentage) | Netherlands | The yearly cost of consumer credit, including costs — the Dutch equivalent of APR |
Compounding turns a nominal rate into an effective one with this formula:
Effective rate = (1 + r ÷ n)ⁿ − 1
Here r is the nominal rate and n is the number of compounding periods per year.
Example: a savings account at 4.5% compounded monthly has an effective rate of (1 + 0.045 ÷ 12)¹² − 1 = 4.594%. $10,000 left for 10 years grows to about $15,669.93.
The standard loan payment formula
Almost every online loan, car-finance and mortgage calculator uses the same amortisation formula:
M = P × r ÷ (1 − (1 + r)⁻ⁿ)
- M = monthly payment
- P = amount borrowed (principal)
- r = monthly interest rate (yearly rate ÷ 12, for a US- or UK-style loan)
- n = number of monthly payments
How to calculate a loan payment step by step:
- Write down the principal: £15,000.
- Convert the yearly rate to a monthly decimal: 7% ÷ 12 = 0.0058333.
- Count the payments: 5 years × 12 = 60.
- Work out (1 + r)⁻ⁿ: 1.0058333⁻⁶⁰ = 0.70531.
- Apply the formula: 15,000 × 0.0058333 ÷ (1 − 0.70531) = £297.02 a month.
- Check the total cost: £297.02 × 60 = £17,821.08, so the interest is about £2,821.
Our loan calculator uses this formula. If your calculator’s answer is far from this for the same inputs, it is using a different compounding rule or adding fees. Look for a note that says so.
Canadian mortgages compound differently
Under Canada’s Interest Act, fixed-rate mortgage interest must be stated as compounded semi-annually, not monthly. A US-style calculator will therefore overstate a Canadian payment.
Example: a $400,000 mortgage at 5% over 25 years.
- Canadian semi-annual compounding: monthly rate = 1.025^(1/6) − 1 = 0.41239%, payment ≈ $2,326.42
- US-style monthly compounding: monthly rate = 5% ÷ 12 = 0.41667%, payment ≈ $2,338.36
The gap is about $12 a month, or more than $3,500 over the term. Use a calculator built for the Canadian market, or one that lets you set the compounding frequency, such as our mortgage calculator. Variable-rate Canadian mortgages can be compounded differently, so read the lender’s disclosure.
Limitations of loan calculators: they usually assume a fixed rate, equal payments and no early repayments, missed payments, fees or insurance. Treat the result as an estimate and compare it with the lender’s official disclosure (the UK’s ESIS, the US Loan Estimate, or the cost-of-borrowing disclosure in Canada). This guide is not financial advice; for a major decision, speak to a regulated adviser.
Rounding: half-up, half-even and when to round
Rounding looks trivial until two systems disagree by a penny on 10,000 invoices.
| Rule | 2.345 → 2 d.p. | 0.125 → 2 d.p. | Where it is used |
|---|---|---|---|
| Round half up | 2.35 | 0.13 | Schools, most shop tills, most calculators |
| Round half to even (“banker’s rounding”) | 2.34 | 0.12 | Many accounting systems, Python’s Decimal default, IEEE 754 floating point |
| Round down (truncate) | 2.34 | 0.12 | Some tax and payroll rules |
Banker’s rounding sends exact halves to the nearest even digit, so rounding errors cancel out over large data sets instead of always pushing upwards.
Rules that prevent most rounding errors:
- Round once, at the end. Keep full precision through every step of a multi-step calculation.
- Round to the right unit. Two decimals for pounds, dollars and euros; none for Japanese yen.
- Decide line-level or total-level tax rounding and use it consistently. Tax authorities often allow either method but expect you to apply one consistently. Mixing them is how invoice totals end up a cent out.
- Match precision to your input. A tape measure read to the nearest millimetre does not justify an answer to six decimal places.
Floating-point error: why 0.1 + 0.2 is not 0.3
Ask JavaScript, Python or a spreadsheet’s hidden working value for 0.1 + 0.2 and you get 0.30000000000000004. This is not a bug in the software. Most computers store numbers in binary using the IEEE 754 standard, and just as 1/3 cannot be written exactly in decimal, 0.1 cannot be written exactly in binary. A standard “double” holds roughly 15 to 17 significant digits, and the tiny leftover shows up in the last place.
You will normally only see it in three situations:
- Equality checks in code or spreadsheets (
IF(A1=0.3, …)can unexpectedly be false). - Rounding exact halves. In Python,
round(2.675, 2)gives 2.67, because 2.675 is actually stored as 2.67499999…. - Long chains of money calculations that are summed thousands of times.
How good calculators handle it: they round the displayed result to a sensible number of digits, and money software uses decimal arithmetic (integer pennies or cents, or a decimal type) instead of binary floating point. If you build your own spreadsheet calculator, wrap money results in ROUND(…, 2) and compare values with a small tolerance instead of =.
Decimal commas, thousands separators and locale
| Written value | UK, USA, English Canada | Netherlands, French Canada, much of Europe |
|---|---|---|
| 1,250 | one thousand two hundred and fifty | one and a quarter (1.25) |
| 1.250 | one point two five | one thousand two hundred and fifty |
| 1,234,567.89 | standard | not used |
| 1.234.567,89 | not used | standard in the Netherlands |
| 1 234 567,89 | not used | standard in French Canada (space as thousands separator) |
This is not a cosmetic difference. A spreadsheet opened on a computer set to Dutch can read 1.250 as 1,250, or treat it as text and silently leave it out of a total. A CSV exported from a US system and opened in a Dutch locale can split columns in the wrong places, because the comma is also the field separator.
Pro-tips:
- When sharing figures across borders, write them unambiguously (for example
1250.00 EUR, or ISO 4217 codes likeCAD). - When importing a CSV, use the import wizard and set the decimal and thousands separators explicitly rather than double-clicking the file.
- Check one known value after any import before trusting the totals.
Ten-second sanity checks
Before you copy any result into an invoice, an exam answer or a purchase order, run at least one of these:
| Check | Example |
|---|---|
| Estimate first | 19.8 × 51 is about 20 × 50 = 1,000. An answer of 10,098 or 100.98 is a typo |
| Reverse the operation | If £120 ÷ 1.20 = £100, then £100 × 1.20 should give £120 again |
| Check the direction | Adding tax or interest must increase the number; a discount must decrease it |
| Check the size | With any positive interest rate, the monthly payment must be more than the loan divided by the number of months |
| Check the units | Is the rate yearly or monthly? Is the term in months or years? Is the length in metres or feet? |
| Try a known benchmark | 10% of 250 is 25; 2 + 3 × 4 is 14 |
How to choose a calculator you can trust
Use this checklist for any online calculator, including ours.
- Shows the formula and the inputs it used, so that you can repeat the calculation by hand.
- States its assumptions, such as the compounding frequency, the tax rate and its effective date, or the rounding rule.
- Cites primary sources for rates and constants: a tax authority, a central bank, or a standards body such as NIST or BIPM.
- Timestamps live data such as exchange rates or interest rates, and names the provider.
- Handles your locale (decimal comma, currency, date format) or tells you which one it expects.
- Works without sign-up and does not collect personal data it does not need. A payment calculator should not need your name or email address.
- Has a clear owner: an about page, a contact route, and a privacy policy.
Trade-offs worth knowing. An online calculator is fast and consistent, but it only knows what you tell it. It cannot tell you that you typed a monthly rate into a yearly field, or that your situation has a special tax rule. A spreadsheet is flexible and auditable, but it is easy to break a formula and not notice. Working by hand is slow but forces you to understand each step. For high-stakes figures, such as tax returns, medication doses, structural loads or mortgage decisions, use two methods and confirm the result with the official source or a qualified professional.
Calculators and unit conversion
Many calculator errors are really unit errors: a rate per month used as a rate per year, a length in feet typed as metres, or a US gallon confused with an imperial gallon. Convert every input into consistent units before you calculate. Our complete guide to unit conversion covers the method and the exact factors.
Frequently asked questions
Why does my phone calculator give a different answer from my scientific calculator?
Basic calculators often work each step as you press it (chain logic), so 2 + 3 × 4 gives 20. Scientific calculators apply the order of operations and give 14. Use brackets, or break the calculation into steps, on a basic calculator.
Is it PEMDAS or BODMAS?
They are the same rule with different names. PEMDAS is taught in the US, BODMAS or BIDMAS in the UK, and BEDMAS in Canada. In all of them, multiplication and division have equal priority and are worked left to right, and so are addition and subtraction.
How do I remove VAT from a price?
Divide the gross price by 1 plus the VAT rate. For UK VAT at 20%, divide by 1.20. For Dutch BTW at 21%, divide by 1.21. Do not multiply by 0.80 or 0.79, which removes too much.
What is the difference between APR and AER (or APY)?
APR is the yearly cost of borrowing and, in the UK and the US, includes certain required fees. AER (UK) and APY (US) show the effective yearly return on savings after compounding. Compare loans by APR and savings by AER or APY.
Why do Canadian mortgage calculators give lower payments than US ones?
Canada’s Interest Act requires fixed-rate mortgage interest to be stated as compounded semi-annually, which gives a slightly lower effective monthly rate than US-style monthly compounding at the same quoted rate.
Why does 0.1 + 0.2 show 0.30000000000000004?
Computers store most numbers in binary floating point (IEEE 754), which cannot represent 0.1 exactly. The difference is tiny. Good calculators round the display, and money software uses decimal arithmetic to avoid it.
Should I round each step or only at the end?
Round only at the end, unless a tax or payroll rule tells you to round at a specific stage. Rounding at every step builds up errors.
Are online calculators accurate enough for tax and legal purposes?
They are accurate for the inputs and rules they use, but they cannot know your full situation. Use them to estimate and check, then confirm the figures with the tax authority’s own guidance or a qualified adviser before filing.
Summary and next steps
- Test the entry logic of any unfamiliar calculator with 2 + 3 × 4 (it should give 14).
- Never write an ambiguous expression. Add brackets so that there is only one reading.
- Multiply to add a percentage and divide to remove one. Multiply successive changes; do not add them.
- Use the right tax rate for the right place and date, and check the official source.
- Know your compounding: APR vs. AER/APY, and semi-annual compounding for Canadian fixed-rate mortgages.
- Round once, at the end, using the rule your accounting system expects.
- Watch the decimal separator when figures cross borders.
- Sanity-check every result with an estimate or a reverse calculation before you rely on it.
Try the converters: Currency Length Weight and mass Temperature Volume Area Speed Time Pressure Power Data storage
Related reading: The complete guide to unit conversion
Sources
- IEEE 754-2019, IEEE Standard for Floating-Point Arithmetic
- HM Revenue & Customs, VAT rates (GOV.UK)
- Belastingdienst, Btw-tarieven and notice on the accommodation rate rising to 21% from 1 January 2026
- Canada Revenue Agency, GST/HST calculator (and rates) and Charge and collect the tax – Which rate to charge
- Revenu Québec, QST rate
- Interest Act (R.S.C., 1985, c. I-15), section 6, Canada
- Financial Conduct Authority, Consumer Credit sourcebook (CONC) on APR calculation (UK)
- Consumer Financial Protection Bureau, Regulation Z (Truth in Lending) and Regulation DD (Truth in Savings), USA
- BIPM, The International System of Units (SI Brochure), 9th edition
- NIST Special Publication 811, Guide for the Use of the International System of Units (SI)