Logarithms for NDA Mathematics — Properties, PYQs, and Shortcuts

beginner 18 min read

Concept

A logarithm answers one simple question: to what power must this base be raised to get this number?

When you write log_b(a) = x, it means b^x = a. That's the entire definition. Everything else — all the rules, all the tricks — is a consequence of this one equivalence.

Here's an analogy that sticks. Suppose you know that 10^3 = 1000. Now someone asks: "what power of 10 gives 1000?" The answer is 3. That answer — 3 — is log₁₀(1000). Logarithm is just the inverse operation to exponentiation, the same way subtraction is the inverse of addition.

Two bases you must be comfortable with:

Domain restrictions — this is where most marks are lost, not gained. log_b(a) is defined only when:

  1. a > 0 (argument must be strictly positive)
  2. b > 0 and b ≠ 1 (base must be positive and not equal to 1)

Violate either condition and the expression is undefined. In equation-solving questions, any solution that makes an argument non-positive or zero must be rejected. This is not a technicality — NDA has tested this directly.

Antilog is simply the reverse: antilog_b(x) = b^x. If log₁₀(x) = 2, then antilog₁₀(2) = 10^2 = 100 = x.

Think of logs and antilogs as a pair of doors — one takes you from the exponent world into the log world, the other brings you back.


Deep Dive

The Core Properties

These four identities are non-negotiable. Derive them once from the definition, then use them without hesitation.

1. Product Rule: log_b(mn) = log_b(m) + log_b(n)

Why it works: If b^p = m and b^q = n, then mn = b^p · b^q = b^(p+q). So log_b(mn) = p + q = log_b(m) + log_b(n).

2. Quotient Rule: log_b(m/n) = log_b(m) − log_b(n)

Same logic: m/n = b^p / b^q = b^(p−q).

3. Power Rule: log_b(m^k) = k · log_b(m)

If b^p = m, then m^k = b^(kp), so log_b(m^k) = kp = k · log_b(m).

4. Change of Base: log_b(a) = log_c(a) / log_c(b)

This is the most underused property in exam halls. It lets you convert any log to base 10 (or any convenient base) using a calculator — or in NDA's case, lets you manipulate expressions algebraically.

A critical corollary: log_b(a) · log_a(b) = 1, which means log_b(a) = 1 / log_a(b). This appears constantly in NDA problems — memorise this reciprocal relationship.

Special Values to Internalize

| Expression | Value | Why | |---|---|---| | log_b(1) | 0 | because b^0 = 1 always | | log_b(b) | 1 | because b^1 = b always | | log_b(b^k) | k | direct from power rule | | b^(log_b(a)) | a | logs and exponentials cancel |

Logarithmic Equations — The Standard Approach

When you see a log equation, your first two moves are:

  1. Simplify both sides using properties until you have log_b(LHS) = log_b(RHS) with the same base, then cancel the logs to get LHS = RHS.
  2. Check domain — substitute back and verify every argument is strictly positive.

For equations like log₄(x−1) = log₂(x−3), the bases differ. The move is change of base: log₄(x−1) = log₂(x−1) / log₂(4) = log₂(x−1) / 2. Now both sides share base 2 and you can proceed.

Logarithmic Inequalities — The Sign Flip

Here's where students drop marks silently. When you remove the log from an inequality:

Think of it this way: log_{0.5} is a decreasing function, so larger log value means smaller argument.

The t + 1/t Pattern

Look at the PYQ on log_{sin x}(cos x) + log_{cos x}(sin x) = 2. If you let t = log_{sin x}(cos x), then log_{cos x}(sin x) = 1/t (by the reciprocal property). The equation becomes t + 1/t = 2.

Now, t + 1/t = 2 has exactly one solution: t = 1. This is because t + 1/t − 2 = (t−1)²/t = 0 implies t = 1.

This t + 1/t pattern — where log and its reciprocal appear together — is a recurring NDA structure. Recognise it and you save 90 seconds.

AM-GM Bound on t + 1/t

For t > 0: by AM-GM inequality, (t + 1/t)/2 ≥ √(t · 1/t) = 1, so t + 1/t ≥ 2.

This means t + 1/t can never be less than 2 for positive t. For t < 0: t + 1/t ≤ −2.

The PYQ on k = 2 − t − 1/t exploits exactly this. Since t + 1/t ≥ 2, we get k ≤ 0. So k = 1 is impossible.


Memory Tricks & Shortcuts

patternLog Product = Sum, Log Quotient = Difference

When you see multiplication inside a log, it becomes addition outside. When you see division inside, it becomes subtraction outside. The trick: logs "hate" multiplication and division — they break them apart into addition and subtraction.

Micro-example: log(50) = log(100/2) = log 100 − log 2 = 2 − 0.301 = 1.699.

Standard method (recalling log 50 from memory): 15-20 seconds if you remember it. This breakdown: 5 seconds. Saves you 10-15 seconds per problem.

patternReciprocal Swap

log_b(a) = 1 / log_a(b) — the base and argument swap, and the whole thing flips.

Use this whenever a problem has log_b(a) and log_a(b) appearing in the same expression. Let t = log_b(a), then log_a(b) = 1/t, and the expression becomes t + 1/t (or t − 1/t, or t · 1/t = 1).

Micro-example: log_{sin x}(cos x) + log_{cos x}(sin x) = 2 becomes t + 1/t = 2 in one step versus three steps of change-of-base manipulation. Saves 2 algebraic steps.

substitutionSubstitution to Avoid Guessing

When an equation has terms like 5^x + 4^x + 3^x + 2^x + 1 inside a log and the options include small integers (especially 0 or 1), test x = 0 first — every exponential term becomes 1.

Micro-example: 5^0 + 4^0 + 3^0 + 2^0 + 1 = 1+1+1+1+1 = 5 and you need the argument to equal 5. Done.

Standard approach (algebraic manipulation): 60-90 seconds. Substitution: 10 seconds. Use it whenever you see mixed exponentials inside a log.

patternChange of Base to Unify Bases

When two log terms have different bases where one base is a power of the other (like base 4 and base 2), convert to the common smaller base using log_{b^k}(x) = (1/k) log_b(x).

Micro-example: log₄(x−1) = (1/2)log₂(x−1). Now both sides have base 2. Apply the one-to-one property and solve algebraically.

This reduces a "different bases" problem to a "same base" problem in one step. Saves the mental overhead of full change-of-base formula.

estimationAM-GM Ceiling on log expressions

For any t > 0, t + 1/t ≥ 2. Use this to immediately determine the range of expressions of the form a − t − 1/t or a + t + 1/t without solving.

Micro-example: k = 2 − t − 1/t. Since t + 1/t ≥ 2, immediately k ≤ 0. Any answer option with k > 0 is ruled out in under 5 seconds — no algebra needed.

Standard approach (trying specific values of t for each option): 40-60 seconds. AM-GM observation: 5 seconds.


Fast-Solving Framework

When you see a logarithm question in the NDA exam, run through this decision tree:

Step 1 — Identify the structure.

Step 2 — Simplify both sides. Apply product, quotient, and power rules to reduce each side to a single log expression.

Step 3 — Solve the resulting equation. Once log_b(LHS) = log_b(RHS), equate LHS = RHS.

Step 4 — Check domain before marking your answer. Substitute back into every log argument. Any argument ≤ 0 disqualifies that solution. This step eliminates traps set specifically for careless solvers.

Step 5 — If the expression asks for a range or "which value is impossible": Use AM-GM or known inequalities on t + 1/t to bound the expression without computing specific values.


Solved PYQs

Why this question: Tests whether you can convert a log expression on the LHS into a recognizable log value before handling the complex RHS.

Previous Year Questionपिछले वर्ष का प्रश्न2026
If 1 − log₁₀ 2 = log₁₀(5^x + 4^x + 3^x + 2^x + 1), then what is a value of x?
  1. 10
  2. 5
  3. 1
  4. 0
Solutionसमाधान
1 − log₁₀ 2 = log₁₀ 10 − log₁₀ 2 = log₁₀ 5. So 5^x + 4^x + 3^x + 2^x + 1 = 5. Testing x = 0: 1 + 1 + 1 + 1 + 1 = 5. ✓

Solving path: Left side — 1 − log₁₀ 2 = log₁₀ 10 − log₁₀ 2 = log₁₀(10/2) = log₁₀ 5. So the argument of the log on the right must equal 5. Test x = 0: 5^0 + 4^0 + 3^0 + 2^0 + 1 = 5. Confirmed. No algebra beyond this is needed.


Why this question: Tests the reciprocal property of logs — a pattern that recurs across NDA sessions. Students who don't recognise log_{cos x}(sin x) = 1 / log_{sin x}(cos x) spend 2-3 minutes on algebra that takes 30 seconds with the right substitution.

Previous Year Questionपिछले वर्ष का प्रश्न2026
What is the smallest positive x satisfying log_{sin x} cos x + log_{cos x} sin x = 2?
  1. π/2
  2. π/3
  3. π/4
  4. π/6
Solutionसमाधान
Let t = log_{sin x} cos x. Then t + 1/t = 2, giving t = 1. So log_{sin x} cos x = 1, meaning cos x = sin x, i.e., tan x = 1. Smallest positive x is π/4.

Solving path: Let t = log_{sin x}(cos x). Then log_{cos x}(sin x) = 1/t. Equation becomes t + 1/t = 2, i.e., t² − 2t + 1 = 0, i.e., (t−1)² = 0, so t = 1. This means cos x = sin x, so tan x = 1. Smallest positive solution: x = π/4.


Why this question: Tests whether you can convert log statements into exponential form and then reconstruct the combined expression. The key move is recognising ace = (bd²f³)^p.

Previous Year Questionपिछले वर्ष का प्रश्न2024
If log_b a = p, log_d c = 2p and log_f e = 3p, then what is (ace)^(1/p) equal to?
  1. bd²f³
  2. bdf
  3. b³d²f
  4. b²d²f²
Solutionसमाधान
From log_b a = p: a = b^p. From log_d c = 2p: c = d^(2p). From log_f e = 3p: e = f^(3p). So ace = b^p · d^(2p) · f^(3p) = (bd²f³)^p. Therefore (ace)^(1/p) = bd²f³.

Solving path: From log_b a = p, write a = b^p. From log_d c = 2p, write c = d^{2p}. From log_f e = 3p, write e = f^{3p}. Multiply: ace = b^p \cdot d^{2p} \cdot f^{3p} = (bd^2f^3)^p. Raise both sides to 1/p: (ace)^{1/p} = bd^2f^3.


Why this question: Two logs with different but related bases, leading to a quadratic after change of base — and then a domain-check trap. Missing the domain step costs the question.

Previous Year Questionपिछले वर्ष का प्रश्न2024
What is the number of solutions of log₄(x – 1) = log₂(x – 3)?
  1. Zero
  2. One
  3. Two
  4. Three
Solutionसमाधान
log₄(x–1) = log₂(x–3) gives (1/2)log₂(x–1) = log₂(x–3), so log₂(x–1) = 2 log₂(x–3) = log₂(x–3)². Hence x–1 = (x–3)², giving x–1 = x² – 6x + 9, so x² – 7x + 10 = 0, yielding x = 2 or x = 5. For x = 2, x – 3 = –1 < 0 (invalid for log). Only x = 5 is valid. So there is exactly one solution.

Solving path: log_4(x−1) = (1/2)log_2(x−1). Equation becomes (1/2)log_2(x−1) = log_2(x−3), so log_2(x−1) = 2\log_2(x−3) = log_2(x−3)^2. Equate arguments: x−1 = (x−3)^2 = x^2 − 6x + 9, giving x^2 − 7x + 10 = 0, so (x−2)(x−5) = 0. Solutions: x = 2 or x = 5. Check x = 2: argument of log_2(x−3) = log_2(−1) — undefined. Reject. Only x = 5 is valid. Answer: one solution.


Why this question: Tests the AM-GM bound on log expressions — a purely reasoning-based question where no equation needs to be solved, only bounded.

Previous Year Questionपिछले वर्ष का प्रश्न2024
For x ≥ y > 1, let log_x(x/y) + log_y(y/x) = k, then the value of k can never be equal to
  1. –1
  2. –1/2
  3. 0
  4. 1
Solutionसमाधान
Let log_x y = t. Since x ≥ y > 1, we have 0 < t ≤ 1. Then log_x(x/y) = 1 – t and log_y(y/x) = 1 – log_y x = 1 – 1/t. So k = (1–t) + (1 – 1/t) = 2 – t – 1/t. By AM-GM, t + 1/t ≥ 2 with equality when t = 1, so k ≤ 0. The maximum value of k is 0 (when t = 1, i.e., x = y). For t ∈ (0, 1], k ranges from –∞ to 0. Therefore k can equal –1, –1/2, 0 but never 1.

Solving path: Let log_x y = t. Since x ≥ y > 1, we have 0 < t ≤ 1. Then log_x(x/y) = 1 − t and log_y(y/x) = 1 − \log_y x = 1 − 1/t. So k = (1−t) + (1−1/t) = 2 − t − 1/t. For t \in (0, 1], by AM-GM: t + 1/t \geq 2, so k \leq 0. The value k = 1 > 0 is impossible.


Common Mistakes


Related Topics


Practice on SarkariRise

Sign up + get 3 free mocks →