Systems of two linear equations

Skip this block if you can already say, without hesitating, what it means for a pair of numbers to solve a system and what the three possible outcomes are. Nothing later assumes more than that. Re-reading what you already know is the most reliable way to feel productive while learning nothing.

Start with one equation in two variables: 2x + y = 10. It is not a puzzle with one answer. The pair (0, 10) works. So does (1, 8), (5, 0), (−3, 16), and (2.5, 5). A single linear equation in two variables has infinitely many solutions, and if you plot every one of them you get a straight line. That line is the equation — the picture and the algebra are two views of the same object.

Now add a second equation: x − y = 2. On its own it also has infinitely many solutions, and they also form a line. A solution to the system is a pair that appears on both lists at once — a point sitting on both lines. Here that pair is (4, 2): 2(4) + 2 = 10 and 4 − 2 = 2. The word doing all the work is at once. Solving each equation separately and reporting two answers is not solving a system; it is solving two unrelated problems.

Because two straight lines in a plane can only relate to each other in three ways, a system of two linear equations has exactly three possible outcomes and no others. They cross at one point — exactly one solution. They are parallel and distinct — no solution, and the algebra will end in an impossible statement like 0 = 7. They are the same line drawn twice — infinitely many solutions, and the algebra will end in a statement that is always true, like 0 = 0. There is no arrangement of two straight lines that produces exactly two crossing points, so "two solutions" is never an answer to this question type.

One notation habit worth fixing now: the answer is an ordered pair (x, y), x first, y second, and the SAT will happily offer you the same two numbers in the other order as a distractor. And one checking habit: once you have a candidate pair, substitute it into both equations. A pair that satisfies one equation and not the other is not a partial success — it is simply not a solution.

Substitution. Take one equation, get a single variable alone on one side, and put the expression you get in place of that variable in the other equation. You now have one equation in one variable — solve it, then go back for the second variable. Use substitution when a variable is already isolated ("y = 3x − 7") or has a coefficient of 1 or −1, because isolating it costs you nothing and introduces no fractions.

Elimination. Line the two equations up in the same arrangement, multiply one or both through by a constant until one variable's coefficients are opposites, then add the equations. That variable disappears and you are left with one equation in one variable. Use elimination when neither variable is isolated and neither has a coefficient of 1 — which describes most SAT systems written in Ax + By = C form.

Two refinements turn elimination from a procedure into a strategy. First: eliminate the variable you were not asked about. If the question wants y, kill x, and you land on y directly without ever computing x or back-substituting. Weak path — solve for both variables every time, then look for the one you need. Strong path — read the stem first, then choose which variable to destroy.

Second: prefer adding to subtracting. Subtracting one equation from another requires negating every term on the right-hand side as well as the left, and a dropped negative there is the most common arithmetic loss on this skill. If the coefficients are equal rather than opposite, multiply one whole equation by −1 first and then add. One negation applied to a whole line is far safer than three applied term by term.

There is also a third move that is not a method so much as a piece of pattern recognition, and it is worth more time than either method on the hardest items. When the question asks for a combination — x + y, x − y, 2x + 2y — check whether adding or subtracting the two equations produces that combination directly. Systems with mirrored coefficients (5x + 7y = 34 and 7x + 5y = 26) are built for it: adding gives 12x + 12y = 60, so x + y = 5, and you never find out that x is 0.5. That is not a trick; it is noticing that the question asked for less than a full solution.

A large share of SAT systems questions never ask you to solve anything. They give you a system with a letter in it — "for what value of k does this system have no solution?" — and the entire item is the relationship between coefficients. There is one rule and one precondition, and the precondition is where the points go.

The precondition: both equations must be written in the same arrangement before you compare anything. Ax + By = C for both, x-term first, y-term second, constant alone on the right. "8y = 3x + 5" has coefficients (−3, 8) with constant 5, not (3, 8) — the 3x has to cross the equals sign, and it changes sign when it does. Reading coefficients straight off an equation in the wrong form is the single most reliable way to get one of these items wrong while doing every subsequent step perfectly.

The rule, in the form that never breaks: for a₁x + b₁y = c₁ and a₂x + b₂y = c₂, the two lines are parallel exactly when a₁b₂ = a₂b₁. If that product equality fails, the lines cross once and the system has exactly one solution — full stop, whatever the constants are. Use the cross-multiplied form rather than the ratio form a₁/a₂ = b₁/b₂, because the ratio form breaks when a coefficient is zero, and the SAT does write systems where one equation is 5y = 20 or 3x = 12.

Once you know the lines are parallel, the constants decide which parallel case you have. Same line, infinitely many solutions, when the whole triple is proportional — a₁ : b₁ : c₁ equals a₂ : b₂ : c₂. Two distinct parallel lines, no solution, when the coefficients are proportional and the constants are not.

In practice, do it as one motion. Find the number that multiplies equation one's coefficients into equation two's coefficients. Apply that same number to equation one's constant. If the result matches equation two's constant, the equations are the same equation and the system has infinitely many solutions. If it does not match, the system has none. If no single number works on both coefficients, the lines are not parallel at all and there is exactly one solution. Worked on 6x + 15y = 24 and 10x + ky = 40: the multiplier taking 6 to 10 is 5/3, applied to 24 it gives 40, which matches — so at k = (5/3)(15) = 25 the equations coincide.

College Board provides a built-in Desmos graphing calculator inside Bluebook for the entire Math section, on both modules; you may also bring an approved handheld. That much is official. Every claim about how many points the calculator is worth, including anything in this lesson, is a prep-industry estimate and should be held as one.

The mechanics that matter: type each equation on its own line, exactly as printed. Desmos does not need y isolated and does not need standard form — 3x + 4y = 22 goes in as written. Then click where the two lines cross; Desmos labels the point with its coordinates. For a concrete system with no letters in it, that is a complete solution in about fifteen seconds, and it does not care whether the coefficients are ugly.

Where Desmos wins: concrete systems with coefficients that make elimination messy; verifying an answer you already produced algebraically, which costs ten seconds and catches sign errors that re-checking your own arithmetic will not; and the second half of a word problem, once you have written the two equations yourself. Note what is excluded there — translating the words into equations is still entirely yours, and it is where word-problem points are actually lost.

Where Desmos fails, and these are the items that decide 1600s. First, parameter questions: "for what value of k does the system have no solution" has nothing to graph until k is fixed. You can define k as a slider and drag it, but sliding toward a value you could have computed exactly is guessing at a number the coefficient rule hands you in one step. Second, and worse, no-solution and infinitely-many questions are visually ambiguous by construction: two identical lines look like one line, and two parallel lines a small distance apart also look like one line at default zoom. The picture cannot distinguish "one line drawn twice" from "two lines that never meet" — only the constants can. Third, the intersection readout is a decimal, so an exact answer of 2/3 appears as 0.6667 and a grid-in typed as 0.667 may not be what the question wanted.

One more failure that costs whole questions rather than partial credit: the default window is roughly −10 to 10 on both axes, and a system whose solution is (−40, 120) shows two lines that appear not to cross at all. If you see no intersection, zoom out before you conclude the system has no solution.

The summary that fits on one line: Desmos is an excellent solver of concrete systems and an excellent checker. It is not a reasoner about coefficients, and the coefficient items are the hard ones.

Everything above gets a careful worker through nearly every systems item on the test. The points that separate 1500 from 1600 on this skill are almost never lost to not knowing how substitution works — they are lost after the mathematics is already correct. Six specific things take them.

Restate the asked quantity before you compute it. The habit is two-part and both parts matter: read the stem and mark the requested expression before starting, then, once you have x and y, read the stem again and compute exactly what it names. Weak path — solve, then scan the choices for a number you produced. Strong path — solve, reread, compute, then look. On grid-ins that discipline is the entire defence, because there are no choices at all: College Board's published spec is that roughly a quarter of the 44 Math questions are student-produced responses, and a system item that arrives as a grid-in has nothing to catch a substituted variable.

Take the combination shortcut whenever it exists. If the stem asks for x + y and the two equations have mirrored coefficients, adding them is the whole solution and the individual values are frequently non-integers designed to cost you two minutes. The test writes these deliberately: 5x + 7y = 34 and 7x + 5y = 26 gives x + y = 5 in one line, while solving it properly gives x = 0.5 and y = 4.5. Before committing to a full solve, spend three seconds asking whether the stem wants less than a full solution.

Use the cross-multiplied form of the parallel test, and check which degenerate case is even available. a₁b₂ = a₂b₁ survives a zero coefficient; the ratio form a₁/a₂ = b₁/b₂ does not, and systems containing an equation like 5y = 20 do appear. Then the subtler point: on some parameter items only one of the two degenerate cases is reachable. If the two constants are already in the same ratio as the coefficients you are not allowed to change, then any parameter value that makes the lines parallel makes them identical — so "no solution" is impossible for every value of the parameter, and an item asking which statement is true is testing exactly that. Confirm the case you have concluded rather than assuming both were on the table.

Substitute back into the equation you did not use. Finding y by substituting your x into the same equation you rearranged confirms only your own rearrangement; it will happily validate a value that fails the other equation entirely. Use the other equation, always. Ten seconds, and it catches the sign error that the rest of your working cannot.

Then the pure execution losses, which cost the most at this level precisely because each one produces a confident answer and confidence is why they survive review. Reversing a comparison in a word problem — "the number of student tickets was 30 greater than the number of adult tickets" is s = a + 30, and writing a = s + 30 produces a fully self-consistent wrong solution that checks out against the total. Losing a unit: computing a count when the stem asked for a dollar amount, or vice versa. Swapping the coordinates in an ordered pair. Dropping a negative when a coefficient crosses the equals sign. None of these is a gap in understanding; all of them are answers to a question slightly different from the one printed.

Time is part of the margin, not separate from it. Each Math module is 35 minutes for 22 questions — about 95 seconds per question. That structure is College Board's published design; how anyone recommends spending it, including everything in this paragraph, is prep-industry advice. Concrete systems are among the fastest points on the section, especially with Desmos, and the right way to treat them is as time banked for the Advanced Math items later in the module. Parameter systems are the opposite: they look hard and take fifteen seconds once the coefficient rule is automatic, which is the specific reason it is worth making automatic.