I’m prepping for a test, and rearranging formulas is where I keep losing time. I’m fine with the easy ones, but as soon as there are fractions, exponents, or the variable shows up in more than one spot, I start second-guessing and make sign mistakes.
Examples I’m stuck on:
– Make t the subject in s = ut + (1/2) a t^2. When do you stop trying to peel terms and just treat it as a quadratic?
– Make r the subject in A = P(1 + r/n)^(nt). Are logs always the right move, or is there a tidier route?
– Make x the subject in y = (a x + b) / (c x + d). What’s the reliable order of moves so I don’t divide by zero by accident or drop a minus sign?
– Make h the subject in V = π r^2 h + k h. Feels like there should be a one-line trick here instead of dancing around.
I’m after a dead-simple checklist: what to do first, what to do next, when to factor, when to take roots/logs, and a couple of fast sanity checks so I know I didn’t introduce extra or lose solutions. Please keep it practical.
Follow-up: when is it actually okay to cross-multiply, and when should I multiply through by the entire denominator first (especially if there are sums in there)? Also, if the thing I’m tempted to divide by could be zero, what’s the safe way to handle that on test day without writing a full case breakdown?
Ready to make maths more enjoyable, accessible, and fun? Join a friendly community where you can explore puzzles, ask questions, track your progress, and learn at your own pace.
By becoming a member, you unlock:
3 Responses
A simple order of play: (1) Move everything involving the target variable to one side; (2) Clear denominators by multiplying both sides by the entire denominator(s); (3) Factor common factors; (4) Apply the matching inverse operation: if the variable sits in a single linear factor under a power, take the corresponding root/log; if it appears in more than one term or power, treat it as a polynomial (often quadratic) and solve; (5) Only divide by a factor after noting it is assumed nonzero; if it could be zero, handle that case briefly by checking it in the original equation. Now the examples: s = ut + (1/2) a t^2 is quadratic in t, so write (1/2) a t^2 + ut − s = 0 and get t = [−u ± sqrt(u^2 + 2as)]/a (if a ≠ 0); if a = 0 then t = s/u (if u ≠ 0). A = P(1 + r/n)^(nt): isolate the power A/P = (1 + r/n)^(nt), then take the nt-th root (tidier than logs): r = n[(A/P)^(1/(nt)) − 1], with A/P > 0; if t = 0 there’s no information about r. y = (ax + b)/(cx + d): multiply both sides by cx + d to get y(cx + d) = ax + b, collect x-terms, (yc − a)x = b − yd, so x = (b − yd)/(yc − a), with yc − a ≠ 0 and cx + d ≠ 0 (the latter is already required by the original fraction). V = πr^2 h + kh factors as h(πr^2 + k), so h = V/(πr^2 + k), provided πr^2 + k ≠ 0. Cross-multiplication is just “multiply both sides by the entire denominator(s)”; it’s fine when each side is a rational expression, but always include the nonzero conditions for those denominators, and never multiply term-by-term-use the whole parenthesized denominator. If a factor you plan to divide by could be zero, the safe test-day move is to say “assume X ≠ 0, divide and solve,” then quickly check X = 0 in the original equation to see if it yields solutions or is disallowed; that’s one short extra line, not a full case write-up. Fast sanity checks: plug the answer back mentally in a simplified form, watch units/signs, note domain restrictions you introduced (denominators nonzero, arguments of logs/roots positive), and for quadratics ensure the discriminant is nonnegative when real solutions are expected. Hope this helps!
I think of rearranging like untangling earbuds: strip the outer wraps first, collect the strands you care about, then do the one clean “pull.” My practical checklist: 1) Clear wrappers (multiply through by common denominators, take roots to undo squares, etc.), 2) Get all target-variable terms on one side and everything else on the other, 3) Factor the target, 4) Apply the inverse move (divide, take logs, take a root), 5) Do a tiny domain/units sanity check. Now your examples: s = ut + (1/2)at^2 is a quadratic in t, so stop peeling and go standard form: (1/2)at^2 + ut − s = 0 → t = [−u ± sqrt(u^2 + 2as)]/a (if a ≠ 0; if a = 0 then t = s/u when u ≠ 0). A = P(1 + r/n)^(nt): isolate the power first, A/P = (1 + r/n)^(nt) → (A/P)^(1/(nt)) = 1 + r/n → r = n[(A/P)^(1/(nt)) − 1] (logs are optional; same result). y = (ax + b)/(cx + d) worked example: note cx + d ≠ 0; multiply by the whole denominator: y(cx + d) = ax + b → ycx + yd = ax + b → (yc − a)x = b − yd → x = (b − yd)/(yc − a) (also need yc ≠ a; if yc = a, check consistency b − yd = 0). V = πr^2h + kh: factor the target straight away, h(πr^2 + k) = V → h = V/(πr^2 + k) (provided πr^2 + k ≠ 0). Cross-multiplying is just “multiply both sides by both denominators” and is safe when you truly have a/b = c/d; otherwise multiply by the full common denominator and state the nonzero restrictions (quick note “assume … ≠ 0; handle zero separately” is enough on test day). Fast checks: plug a simple value back, check special cases (a = 0, t = 0), and watch operations that change domains (logs require positives). Nice refresher: https://www.khanacademy.org/math/algebra/formulas-and-expressions/rearranging-formulas/v/rearranging-formulas-intro.
Love this! Rearranging is like unstacking nested functions: undo the last operation first, keep the domain constraints in your peripheral vision, and pounce on common factors. My go-to sequence: (1) declare any “no-zero” constraints from denominators or logs/exponents (e.g., cx+d ≠ 0, A/P > 0), (2) clear fractions in one swoop by multiplying both sides by the full LCD (cross-multiply only when each side is a single fraction), (3) gather all terms with the target on one side and the rest on the other, (4) if the target appears in multiple terms, factor it; if it appears in an exponent, take logs after isolating the base; if it appears squared as well as linearly, rewrite in standard quadratic form and use the quadratic formula, (5) divide by the remaining coefficient/factor, then (6) sanity checks: plug an easy value, verify units/signs, and confirm any excluded values don’t sneak back in. If you’re tempted to divide by something that might be zero, write “assume … ≠ 0; check case … = 0 at the end.” That one line saves points and time. For a quick refresher with examples, see “Rearranging formulas to isolate a variable” on Khan Academy: https://www.khanacademy.org/math/algebra/one-variable-linear-equations/alg1-rearranging-formulas/a/rearranging-formulas-article
Simple worked example (your rational one): Solve y = (a x + b)/(c x + d) for x. Constraint first: cx + d ≠ 0. Because each side is a single fraction (y = …/…), cross-multiply cleanly: y(cx + d) = ax + b. Expand: ycx + yd = ax + b. Collect x-terms one side, constants the other: ycx − ax = b − yd. Factor x: x(yc − a) = b − yd. Divide (assuming yc − a ≠ 0): x = (b − yd)/(yc − a). Quick branch check: if yc − a = 0, then the equation reduces to 0·x = b − yd, so either no solution (if b − yd ≠ 0) or infinitely many x values that also satisfy cx + d ≠ 0 (if b − yd = 0). Fast and safe!
Lightning round on your others: s = ut + (1/2)at² → treat as a quadratic in t once both t and t² are present: (1/2)at² + ut − s = 0, so t = [−u ± sqrt(u² + 2as)]/a when a ≠ 0; if a = 0 it’s just linear t = s/u (and if also u = 0 then you need s = 0 to have any solution). A = P(1 + r/n)^(nt) → isolate the exponential then log: A/P = (1 + r/n)^(nt), so ln(A/P) = nt·ln(1 + r/n), hence 1 + r/n = (A/P)^(1/(nt)) and r = n[(A/P)^(1/(nt)) − 1], with A/P > 0 and t,n ≠ 0. V = πr²h + kh → factor the target: h(πr² + k) = V, so h = V/(πr² + k), and note the quick constraint πr² + k ≠ 0 (if it equals 0, then V must be 0 or there’s no solution). Cross-multiplying is fine only when both sides are single fractions; otherwise multiply through by the entire LCD to avoid piecemeal mistakes, and keep those “might-be-zero” notes so you can check edge cases in seconds at the end.