In short

You have two algebraic tools for a pair of linear equations — substitution and elimination. Picking the right one before you start saves a page of arithmetic. Use this two-question decision tree:

  1. Is one variable already isolated (e.g., x = 2y + 3 or y = \dots)? Yes → substitution. The work is half done.
  2. Are the coefficients of one variable equal, opposite, or simple multiples (like 2 and 4, or 3 and -3)? Yes → elimination. One small multiplication aligns them and that variable cancels.
  3. Otherwise — neither isolated, nor a clean coefficient match — either works. Pick whichever produces fewer fractions. By default, lean towards elimination for ax + by = c form.

You will never solve a system blindly again.

In a CBSE Class 10 board paper, you might face four or five systems of equations across the algebra section. Each one takes between 90 seconds and four minutes depending on which method you pick. Pick wrong, and you are inverting fractions you never needed to write. Pick right, and the answer falls out cleanly. This article is the one-page rule book that tells you which to pick — and why.

The decision tree

Decision tree: substitution versus elimination A flowchart with two diamond-shaped questions. The first asks if one variable is already isolated. If yes, the path leads to a box marked Substitution. If no, the path leads to a second diamond asking if the coefficients of one variable are equal or simple multiples. If yes, the path leads to Elimination. If no, the path leads to a box that says either method works, lean elimination. Q1: Is one variable already isolated? e.g., x = 2y + 3 YES NO SUBSTITUTION Inject the isolated form into the other equation. Q2: Coefficients of one variable equal or simple multiples? YES NO ELIMINATION Multiply, then add or subtract. EITHER (lean elim.) Pick the cleaner arithmetic.

Two questions, three answers. Memorise the shape — in the exam hall, you will run this tree without thinking.

Why these two questions, in this order?

Q1 first because isolation is rare and decisive. If a problem hands you x = 2y + 3 or y = \tfrac{5 - 2x}{3}, the substitution method is essentially already started — there is nothing to engineer. Why: substitution's whole cost is the isolation step. If that cost is zero, every other method is just doing extra work.

Q2 next because aligning coefficients is elimination's whole cost. If the coefficients of, say, x are 3 and 6, then one multiplication (\times 2 on the first equation) makes them match — and a single subtraction kills x entirely. Why: elimination's only friction is finding the right multipliers. When the multiplier is 1, 2, 3 — a number you can see by eye — elimination beats substitution by a wide margin, because substitution would force you to push fractions through both equations.

If neither question fires, the system has neither a freebie isolation nor a freebie alignment. Both methods will produce fractions. Pick the one whose fractions are smaller — usually elimination, because adding two equations often cancels signs that substitution would carry through.

Three worked examples — running the tree

Example 1 — Q1 fires: substitution wins

Solve:

x = 2y + 3, \qquad 5y - x = 7.

Run the tree. Q1: is one variable already isolated? Yes — the first equation literally reads x = \dots. Stop. Use substitution.

Inject x = 2y + 3 into the second equation:

5y - (2y + 3) = 7.

Expand the bracket: 5y - 2y - 3 = 7, so 3y = 10, giving y = \tfrac{10}{3}.

Back-substitute: x = 2 \cdot \tfrac{10}{3} + 3 = \tfrac{20}{3} + \tfrac{9}{3} = \tfrac{29}{3}.

Check: 5y - x = \tfrac{50}{3} - \tfrac{29}{3} = \tfrac{21}{3} = 7. ✓

Total work: one substitution, two lines of algebra. Elimination would have forced you to first rearrange the first equation into -x + 2y = -3, then multiply by 5 — pointless extra steps.

Example 2 — Q2 fires: elimination wins

Solve:

2x + 3y = 12, \qquad 4x - y = 7.

Run the tree. Q1: is one variable isolated? No — both equations are in ax + by = c form. Q2: are coefficients of one variable equal or simple multiples? Yes — the x-coefficients are 2 and 4, and 4 is exactly 2 \times 2. Use elimination.

Multiply the first equation by 2 to align the x-coefficients:

4x + 6y = 24, \qquad 4x - y = 7.

Subtract the second from the first:

(4x + 6y) - (4x - y) = 24 - 7 \implies 7y = 17 \implies y = \tfrac{17}{7}.

Back-substitute into 4x - y = 7: 4x = 7 + \tfrac{17}{7} = \tfrac{66}{7}, so x = \tfrac{66}{28} = \tfrac{33}{14}.

Why substitution would have been worse: isolating y from the second equation gives y = 4x - 7. Inject into the first: 2x + 3(4x - 7) = 12 \Rightarrow 14x = 33. Actually, this works out fine here — but try isolating x instead and you immediately get x = (12 - 3y)/2, dragging a fraction through the second equation. Elimination sidesteps the fraction trap.

Example 3 — neither fires: pick by inspection

Solve:

3x + 5y = 11, \qquad 7x + 11y = 25.

Run the tree. Q1: isolated? No. Q2: coefficients of one variable equal or simple multiples? 3 and 7 share no factor; 5 and 11 share no factor. No. You are in the "either works" branch.

Pick elimination, because the alternative — solving x = (11 - 5y)/3 and injecting — drags a denominator of 3 through everything.

To eliminate x: multiply the first equation by 7 and the second by 3:

21x + 35y = 77, \qquad 21x + 33y = 75.

Subtract: 2y = 2 \implies y = 1. Back-substitute into 3x + 5 = 11: x = 2.

Check: 7(2) + 11(1) = 14 + 11 = 25. ✓

Notice how clean this was — the cross-multiply trick (multiply each row by the other's leading coefficient) is the universal elimination move when no obvious match exists. Why: multiplying eq1 by a_2 and eq2 by a_1 always makes the x-coefficients equal to a_1 a_2. It is brute-force alignment.

When ratios matter

The "simple multiple" criterion in Q2 deserves a closer look. Examine the ratio of the coefficients of one variable across the two equations:

A useful sanity rule: if the smaller coefficient divides the larger one in either column, you have free elimination. Why: when a_2 = k \cdot a_1 for integer k, multiplying eq1 by k creates a perfect match, so subtraction kills the variable with zero fractions.

Common student traps

  1. "Substitution is easier" instinct. Many students default to substitution because it feels familiar from earlier classes. But when both equations are in ax + by = c form with no isolated variable, you create fractions the moment you isolate. Pause and check Q2 first.

  2. Mismatching signs during elimination. If you align coefficients as 4x and -4x (opposite signs), you must add, not subtract. If both are +4x, you subtract. Getting this backward inverts your answer's sign.

  3. Forgetting to back-substitute. You found y. The system asks for both. The back-substitution into either original equation is the cheapest step in the whole problem — never skip it.

  4. Over-multiplying. If the coefficients are already 3 and 3, you do not multiply at all — just subtract directly. Multiplying both sides by 1 is a habit that wastes board-paper time.

The exam-hall version

When you see a system, here is the five-second mental script:

  1. Glance at both equations. Is one of them already x = \dots or y = \dots? Yes → substitute.
  2. No? Look at the columns. Do the coefficients of x (or y) share a factor? Yes → eliminate.
  3. Still no? Cross-multiply (multiply each equation by the other's leading coefficient) and eliminate.

In a 40-mark board paper section, picking the right method on every system saves roughly two to four minutes — exactly enough time to revisit the geometry section before the bell.

Going deeper

For matrices and beyond

The decision tree above generalises. For a system A\mathbf{x} = \mathbf{b} with three or more variables, Gaussian elimination is essentially elimination scaled up — you systematically zero out columns. Substitution survives as back-substitution in the final triangular system. Numerical analysts choose pivots (which row to use as the active equation) using a rule that is the multi-variable analogue of Q2: pick the row whose leading coefficient is largest in magnitude, because it minimises the multipliers and keeps round-off error small.

For square systems, Cramer's rule offers a third path — but it costs you three determinants per system, so it is rarely cheaper than elimination unless you need the answer in symbolic form. The decision tree extends: if the determinants are visibly small integers, Cramer is fast; otherwise, elimination remains king.

The deepest reason elimination dominates in computation is that it scales as O(n^3) for an n \times n system, while iterating substitution naively can balloon symbolically. This is why every linear-algebra library on Earth — NumPy, MATLAB, Eigen — implements LU decomposition (a structured form of elimination) at its core, and never substitution.

References

  1. NCERT, Mathematics Textbook for Class X, Chapter 3 — Pair of Linear Equations in Two Variables — the source for the CBSE methods presented here.
  2. Wikipedia, System of linear equations — overview of solution methods including substitution, elimination, and matrix-based approaches.
  3. Khan Academy, Solving systems of equations with substitution — companion video walkthroughs aligned with the substitution and elimination methods.
  4. Trefethen and Bau, Numerical Linear Algebra (SIAM, 1997) — the standard reference for Gaussian elimination as the multi-variable generalisation of the elimination method.
  5. Strang, MIT 18.06 Linear Algebra Lecture Notes — Lectures 2 and 3 frame elimination as the universal computational engine.