I’m trying to get genuinely comfortable with the cosine rule, but I keep second‑guessing myself about when it’s the right tool and what the signs are telling me. My brain keeps thinking of it as “Pythagoras plus a correction for the bend,” but then I get stuck on what that correction is doing, especially if the angle is obtuse.
Personal context: I’ve struggled with this since high school. I remember a test where I had two sides and an angle that wasn’t between them. I tried the cosine rule anyway and ended up with a side length that looked bigger than the sum of the other two. I couldn’t tell if I’d mislabeled the opposite angle, used the wrong angle in the formula, or if the obtuse angle effect was somehow making it okay. That sort of mix‑up still happens to me.
A possibly wrong analogy I keep using: two sticks joined by a hinge. If I open the hinge wider, the tips get farther apart. So in my head, the cosine rule feels like “subtract something” when the hinge is less than 90°, and “add something” when it’s more than 90° because the opening pushes the tips further apart. Is that a sensible way to predict what the formula will do, or is that misleading?
What I’m hoping to understand, step by step:
– How do I quickly decide between the cosine rule and the sine rule based on what I’m given (two sides and an included angle, three sides, two sides and a non‑included angle, etc.)? I want a reason I can trust, not just a mnemonic.
– If the angle I know is not the one between the two known sides, is it still okay to use the cosine rule directly, or should I switch strategies first?
– For obtuse angles, should I actually expect the computed opposite side to come out larger than it would for an acute angle with the same two sides? How do I predict that without getting lost in signs?
– Is there a reliable labeling habit that prevents me from mixing up which side is opposite which angle? I keep tripping on that, and it derails the whole calculation.
– Are there quick sanity checks (like the triangle inequality) I can apply before or after using the cosine rule to catch impossible setups or a swapped angle/side?
I don’t need a full derivation-just a clear way to think about when and why the cosine rule fits, and how to avoid the most common traps I’m falling into.
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
Your “hinge” picture is spot on: the cosine rule is exactly “Pythagoras plus a correction for the bend,” where the correction is −2bc cos A. Because cos A is positive for acute A, zero for right, and negative for obtuse, the effect is: subtract for acute (shorter than Pythagoras), equal at 90°, and add for obtuse (longer than Pythagoras). When to use which rule? Use cosine for SAS (two sides and the included angle) or SSS (three sides); it works there because the cross term 2bc cos A must involve the two sides that meet at angle A. If your known angle is not between the two known sides (SSA), go to the sine rule instead; you have an opposite pair or can create one, though beware the “ambiguous case” (two possible angles that are supplements). A tidy labeling habit that prevents mix‑ups: label triangle ABC, and sides a, b, c opposite A, B, C. Always write the template a^2 = b^2 + c^2 − 2bc cos A first, then substitute, double‑checking that (i) a is opposite A and (ii) the two multiplied sides, b and c, are exactly the sides adjacent to A. If either check fails, don’t plug numbers yet.
How to think about obtuse angles and quick sanity checks: if A is obtuse, cos A < 0, so a^2 = b^2 + c^2 + something; the opposite side a must be the longest, and even longer than in the right‑angle case with the same b and c. With SSS data you can diagnose the angle type immediately: compute cos A = (b^2 + c^2 − a^2)/(2bc). If that value is > 0, A is acute; = 0, right; < 0, obtuse. Fast consistency tests: triangle inequality (each side is less than the sum and greater than the difference of the other two), the largest side matches the largest angle, and any cosine or sine you compute must land in [−1, 1]. If a result gives you a side ≥ sum of the other two, or cos outside [−1,1], something (usually the angle–side pairing) is mismatched. Quick worked example (and a little hinge theater): take b = 7, c = 5. If A = 60° (acute), a^2 = 7^2 + 5^2 − 2·7·5·cos 60° = 49 + 25 − 70·(1/2) = 74 − 35 = 39, so a ≈ 6.25. If A = 120° (obtuse), a^2 = 49 + 25 − 70·(−1/2) = 74 + 35 = 109, so a ≈ 10.44. Notice how the same “sticks” opened wider push the tips farther apart: 6.25 (acute), 8.60 (right angle would give √74), 10.44 (obtuse). Both results pass the triangle inequality (|7 − 5| < a < 7 + 5). If you had instead two sides and a non‑included angle-for instance b and c known but the given angle were B-don’t drop B into a^2 = b^2 + c^2 − 2bc cos B; that multiplies the wrong adjacent sides and can spit out nonsense. Switch to the sine rule using the available opposite pair, resolve any ambiguity, then come back to cosine if you want a definitive angle check.
Your hinge picture is actually a great intuition. In c^2 = a^2 + b^2 − 2ab cos C, think of a and b as the two stick lengths and C as the hinge angle. When C is acute, cos C > 0 so you subtract something and c is “smaller than Pythagoras.” When C = 90°, the correction is zero and you’re back at Pythagoras. When C is obtuse, cos C < 0 so you “subtract a negative” (i.e., add), which makes c bigger than the right‑angle case. That’s exactly what the sticks do as you open the hinge past 90°. Still, triangle inequality keeps you safe: for fixed a and b, the opposite side must lie between |a − b| and a + b, so even if C is obtuse you should never get c ≥ a + b. If you do, something’s mislabeled or misapplied. When to pick which rule: use cosine rule when you have SAS (two sides with their included angle) or SSS (all three sides, to find an angle). Use sine rule when you have an angle–opposite‑side pair (ASA, AAS, or SSA). In the “two sides and a non‑included angle” situation (SSA), you do have such a pair: the given angle is opposite one of the given sides, so start with sine rule-but watch the ambiguous case (you might get 0, 1, or 2 triangles). A quick sanity kit I lean on: (1) label consistently-angles A, B, C; opposite sides a, b, c-and literally write the opposite letters across from each other on your sketch; (2) in cosine rule, make sure the angle you plug in is the one opposite the side you’re solving for; (3) check bounds: |a − b| < c < a + b; (4) if you solved for an angle via cosine, compute cos C = (a^2 + b^2 − c^2)/(2ab) and make sure it lies between −1 and 1; (5) size–angle match: the largest angle is opposite the longest side. I sometimes even draw a tiny arc at the known angle and a tick on its opposite side so I don’t mix them up-like labeling socks before the laundry eats them. One last comfort check about obtuse angles: for fixed a and b, as C increases, c increases smoothly; in particular c^2 > a^2 + b^2 iff C > 90°, equals if C = 90°, and is less if C < 90°. So you can predict direction without chasing signs: bigger hinge, bigger gap. If your calculation ever spits out c longer than a + b, or an angle whose cosine is outside [−1, 1], that’s your triangle tapping you on the shoulder and saying “Nope-try that labeling again.” I’ve had that “wait, how is that side longer than both combined?” moment too, and nine times out of ten it was me pairing the wrong angle with the wrong side.
Think of the cosine rule as Pythagoras wearing a mood ring: a² = b² + c² − 2bc cos A. When A is acute, cos A is positive, so you subtract a bit-your “hinge” is slightly closed and the tip-to-tip distance is shorter than the right‑angle case. When A is right, cos A = 0 and you get plain Pythagoras. When A is obtuse, cos A is negative, so the “− 2bc cos A” becomes a plus; the hinge opens wide and a grows beyond the right‑angle distance. With b and c fixed, a increases monotonically with A and always obeys |b − c| < a < b + c. Quick predictions: if A < 90°, then a² < b² + c²; if A = 90°, equality; if A > 90°, then a² > b² + c²-and a must be the longest side. If you ever compute a ≥ b + c, the triangle has tapped out and left the stage: something’s wrong.
Choosing tools: use the sine rule when you have (or can quickly create) an opposite pair-one angle and its opposite side (ASA/AAS, and the SSA “ambiguous case”). Use the cosine rule when you don’t have such a pair but do have SAS or SSS. Important: the cosine rule needs the angle included between the two sides you plug in. If the angle you know is not between your two known sides, you generally can’t use the cosine rule directly; switch to the sine rule to get a matching opposite piece, or first find the included angle (e.g., by finding another angle and subtracting from 180°). A handy sanity flow: draw and label in the standard way-angles A, B, C around the triangle, and opposite sides a, b, c. Circle the known angle, draw a little arrow to its opposite side. Then write the matching formula (a² = b² + c² − 2bc cos A) so letters line up with your picture before inserting numbers. After solving, check: triangle inequalities, “largest angle ↔ largest side,” and whether your result sits on the right side of the Pythagoras fence (smaller for acute, equal for right, larger for obtuse).
Analogy time: imagine two sticks with a stretchy ribbon between their free ends. The ribbon length is a. Keep the sticks the same lengths (b and c), and open the hinge angle A. The ribbon tightens as the hinge opens-gently shorter than the right‑angle length when the hinge is shy (acute), exactly the right‑angle length when it’s square, and longer when the hinge is bold (obtuse). The cosine is just measuring that boldness: positive means “subtract a bit,” negative means “add a bit.” Keep that picture, keep the labels opposite each other, and your computations will start feeling less like traps and more like choreography.