What a rough estimate is for
A back-of-the-envelope estimate is a calculation done quickly, with round numbers, to find out roughly how big something is. The aim is not an exact answer. It is to learn whether the answer is closer to 10, 1,000 or 100,000, and to catch a result that is wildly off.
Engineers use these estimates at the start of a job to decide whether an idea is feasible, and at the end to check that a detailed calculation, simulation or spreadsheet has not gone astray. A model can produce a confident-looking number that is wrong by a factor of ten because of a single misplaced decimal; a rough estimate is a cheap way to notice.
The basic method
- State the question precisely. “How much energy?” is vague. “How much energy to lift this load to that height?” is workable.
- Break it into factors you can guess or look up: sizes, masses, times, rates.
- Round aggressively. Use one significant figure, or powers of ten.
- Multiply and divide, keeping units.
- Report a range or an order of magnitude, not a long string of digits.
Working with powers of ten is easier if you are comfortable with scientific notation. Rounding to one digit is fine here, since the inputs are guesses anyway.
What “order of magnitude” means
An order of magnitude is a factor of ten. Saying two quantities are within the same order of magnitude means they differ by less than about a factor of ten; saying one is two orders of magnitude larger means about a hundred times. In practice, order-of-magnitude estimates are expected to be right to within a factor of a few, not within a few percent. That is enough to answer many early questions.
Example 1: seconds in a year
A year has 365 days, each with 24 hours, each with 3,600 seconds.
365 × 24 × 3,600 = 31,536,000 seconds.
Rounded, that is about 3 × 10^7 seconds. Remembering “pi times 10^7” is a common shortcut, since it is close to 3.14 × 10^7, but plain 3 × 10^7 is within about 5 percent of the full figure, which is plenty for a rough check.
Example 2: water in a bathtub
Suppose a bathtub is about 1.5 m long, 0.7 m wide and filled to a depth of 0.4 m. These are illustrative guesses.
Volume = 1.5 × 0.7 × 0.4 = 0.42 m³.
Since 1 m³ is 1,000 litres, that is about 420 litres, or “a few hundred litres”. If a calculation for a similar bath returned 42 litres or 4,200 litres, this estimate would flag it.
Example 3: power to lift a load
Suppose a hoist lifts a 100 kg load through 10 m in 20 s. These are illustrative numbers.
- Force to lift: roughly 100 kg × 9.8 m/s² ≈ 980 N, which we can round to about 1,000 N.
- Work done: 980 N × 10 m = 9,800 J, about 10 kJ.
- Average power: 9,800 J ÷ 20 s = 490 W, roughly 500 W.
This is the mechanical power needed at the load, ignoring friction, acceleration and motor losses. A real motor must supply more, since efficiency is below 100 percent. The estimate is a lower bound, which tells you at once that a 5 W motor is hopeless and a 50 kW motor is far oversized.
Estimating from what you already know
Good estimators build up a small stock of reference values: the mass of familiar objects, typical room dimensions, the size of everyday currents and voltages. Anchor an unknown against something you know. If you are unsure of a value, estimate an upper and a lower bound and take a middle value. Guessing between “no less than 50” and “no more than 200” gives a middle of about 100, and, because errors in a product often partly cancel, the final answer is often better than the individual guesses.
Be careful with anything where an order-of-magnitude error is unacceptable. For design, safety or regulatory work, an estimate is only a first pass; the real figures must come from proper calculation, data and review according to the standards that apply to your industry.
Common mistakes
- Dropping units. Carry them through the estimate. A wrong unit often shows up as an absurd result. The discipline is the same as in dimensional analysis.
- Mixing prefixes. Milli, micro and mega look alike on paper. Writing every value in powers of ten helps.
- Rounding in the same direction every time. If you always round up, the errors add rather than cancel.
- Treating the estimate as a measurement. Say “about” and give the range you believe.
- Forgetting efficiency and losses in energy or power estimates, which tends to make real needs larger than the ideal figure.
Using estimates to check detailed results
After a spreadsheet or simulation finishes, do a short mental version of the same problem. If the detailed result and the rough number disagree by less than a factor of two or three, they probably agree for practical purposes. If they disagree by ten, look for a unit conversion slip, a misplaced decimal or a wrong input before trusting either number. This is a core habit for engineering simulation and engineering spreadsheets, where the software hides the arithmetic. The wider set of checks, including limit cases, is covered under sanity checks.
Practice
Try a few on your own: how many seconds are in a working week, what volume of water a small room would hold if filled to a depth of 1 cm, or how much a sack of 25 kg bags weighs if you carry 4. Work them out on paper, then check them with a calculator. The goal is to get used to the feel of sizes, so that a wrong answer looks wrong at a glance.