005Mathematics · Approximation and bounds

Bounds of the True Value from a Recorded Value

Bounds of the True Value from a Recorded Value

Bounds of a true value for rounding, significant figures at a power of ten, or values cut down.

  • 33 bytes
  • COMP mode
01

Worked examples

Example 1

A laboratory rounds the actual capacity v mL of a small container and records it as 42.6 mL, correct to 3 significant figures. Find the possible range of v.

  1. A.42.5≤v<42.7
  2. B.42.55<v≤42.65
  3. C.42.55≤v<42.65
  4. D.42.595≤v<42.605
Answer
C 42.55 ≤ v < 42.65
Standard solution (without a calculator)
  1. Last kept digit: the 3rd significant figure of 42.6 is in the tenths, so q = 0.1. The record is not a power of ten, so this is ordinary rounding.
  2. Half-step bounds: half a step is 0.05 mL: lower = 42.6 − 0.05 = 42.55; upper = 42.6 + 0.05 = 42.65.
  3. End points: 42.55 rounds to 42.6, so it is included; 42.65 rounds to 42.7, so it is not. The answer is C.

Ordinary rounding uses C = 1; the two numbers still need the correct inequality signs.

Do it on the calculator

Example 2

A chemistry group rounds the actual mass m g of a reagent to 3 significant figures and records it as 1.00 g. Find the possible range of m.

  1. A.0.995≤m<1.005
  2. B.0.9995≤m<1.005
  3. C.0.9995<m≤1.005
  4. D.0.999≤m<1.001
Answer
B 0.9995 ≤ m < 1.005
Standard solution (without a calculator)
  1. Place values on each side: 1.00 has 3 significant figures and is a power of ten; its last kept step is q = 0.01. A neighbour just below 1, to 3 significant figures, has a step of only 0.001.
  2. Each bound: below, 0.9995 rounds up to 1.00, so the lowest value is 0.9995. Above, the half-step is still 0.005, so the upper bound is 1.005.
  3. End points: 0.9995 gives 1.00 but 1.005 gives 1.01, so 0.9995 ≤ m < 1.005 and the answer is B. Enter q = 0.01 and C = 0.1.

If 1.00 were “correct to 2 decimal places”, C = 1 would apply and the answer would be 0.995 ≤ m < 1.005. Decide the kind of accuracy first.

Do it on the calculator

Example 3

For positive values, a distance display keeps only 1 decimal place and drops the digits after it. It shows an actual distance d m as 18.4 m. Find the possible range of d.

  1. A.18.35≤d<18.45
  2. B.18.4<d≤18.5
  3. C.18.4≤d<18.45
  4. D.18.4≤d<18.5
Answer
D 18.4 ≤ d < 18.5
Standard solution (without a calculator)
  1. Rule: this cuts a positive value down to the tenths; it is not rounding. The step is q = 0.1.
  2. Bounds: every value from 18.4 up to anything below 18.5 is shown as 18.4; 18.5 itself shows 18.5.
  3. End points: 18.4 is included and 18.5 is not, so the answer is D. The program uses C = 0.

Cutting down uses C = 0: the range is a whole step wide, and the rounding half-step does not apply.

Do it on the calculator
02

When to use it

Some DSE questions on approximation work backwards: from a recorded value, find the range of the true value. 005 covers three cases with one formula — ordinary rounding, rounding to significant figures when the record is a power of ten (such as 1.00), and positive values cut down. You decide q and the type on paper; the program shows the two bounds.

What it is for

  • Ordinary rounding to decimal places or significant figures (C = 1)
  • Significant figures when the recorded value is exactly 10^k, such as 0.1, 1, 10 or 1.00 (C = 0.1)
  • Positive values cut down to a step, such as a display that drops later digits (C = 0)

What you enter

  • A: the recorded positive value r (not the unknown true value)
  • B: the unit q of the last kept digit (6.24 to 2 decimal places → 0.01)
  • C: the type coefficient — only 1, 0.1 or 0

What it shows

  • The lower bound (included); write it down and press EXE
  • The upper bound (not included)
  • Both with 10 digits (Sci 0): 4.255000000E+01 is 42.55

Check before use

  • Inputs come in a fixed order: A = recorded positive value, B = unit q of the last kept digit, C = type coefficient. A and q must be positive, and q must match the recorded accuracy (6.24 to 2 decimal places → q = 0.01, not q = 2).
  • C may only be 1, 0.1 or 0. This short version has no input checks: it neither rejects other values of C nor checks that A fits q, so check the inputs against this page before trusting an output.
  • Use C = 0.1 only for significant-figure rounding when the recorded value is exactly 10^k (0.01, 0.1, 1, 10, 100 …). For fixed decimal places or a fixed measuring step use C = 1, even if the value is 1 or 10. For values clearly cut down, always use C = 0.
  • Positive values only. For ordinary rounding and for cutting down, the lower bound L is included and the upper bound U is not: L ≤ x < U. The program shows only the two numbers; write the inequality signs yourself.
  • Sci 0 means 10 significant digits on screen, not 0 significant figures: E−01 means × 10⁻¹ and E+01 means × 10, so 4.255000000E+01 is 42.55. The trailing zeros are display format, not the question’s accuracy.
  • Treat an end point as exact only if it fits in 10 significant digits. When A ≥ 1 and the bounds come from d decimal places, (digits in A’s whole part) + d ≤ 9 keeps them fully shown; otherwise keep the exact end point on paper. The screen does not round bounds outwards.
  • Checked for 1–9 significant figures, q from 10⁻¹² to 10⁶ and end points that fit the screen. Values near the calculator’s limits, or q many places smaller than A, are outside this guarantee.
  • Areas, sums, products and percentage errors of measured quantities are not solved by the program: find each measurement’s bounds first, then use the normal keys and the formulas in the appendix. When a variable appears more than once, check whether increasing it raises or lowers the result.
  • The last screen is the upper bound; the lower bound is not saved in X or Y, so write it down at the first pause. EXE after the final result runs the program again and AC leaves. Change back to Norm 1 yourself for ordinary display.

Sign in to unlock the full programme

Checking access…

The programme code and appendix unlock together.

Checking access…

Sign in to unlock the full programme

Checking access…

The programme code and appendix unlock together.