Geometry Proofs: How to Start When You Are Stuck
2026-05-22 · 8 min read
A concrete strategy for the blank-page moment in geometry proofs, focused on what to write down first rather than what theorem to recall.
The real problem is rarely the theorems
Most students who say they are "bad at proofs" actually know the relevant theorems perfectly well when asked directly. The real problem is a different skill: knowing what to do with a blank page and a diagram before any theorem has been chosen. Proof-writing is a search process, not a recall process, and it needs its own strategy.
Step one: mark up the diagram completely
Before writing a single line of proof, redraw or annotate the diagram with everything you are given, using tick marks for equal sides, arcs for equal angles, and right-angle boxes where relevant. Then add anything that follows immediately and automatically - vertical angles are equal, a straight line is 180 degrees, angles in a triangle sum to 180 degrees. This step alone often reveals the missing link, because the proof becomes a matter of connecting marks that are now sitting next to each other on the page, rather than living only in your head.
Step two: write down what you are trying to reach
Many students dive into the "given" information and start deriving things outward without ever writing down, explicitly, what the final statement needs to say. Write the goal at the bottom of your working area first. This turns the proof into a two-ended search: what do I know, and what does the goal actually require. If the goal is "prove triangle ABC is congruent to triangle DEF," write out the specific congruence criterion you would need - three sides, two sides and an included angle, and so on - because that tells you exactly which three facts you are hunting for.
Step three: look for the "connector" shape
Almost every geometry proof hinges on one shared feature that links the given information to the goal - a shared side, a shared angle, a pair of parallel lines cut by a transversal, or a circle's radius drawn to a specific point. Ask directly: is there a line or angle that appears in both the given information and the thing I need to prove? If nothing shared exists yet, consider whether adding an auxiliary line - drawing a diagonal, extending a side, dropping a perpendicular - would create one. Auxiliary lines feel like a magic trick until you realise they are almost always drawn for a specific, nameable reason: to create a shared side, a shared angle, or a pair of similar or congruent triangles.
Step four: work backward from the goal
Instead of only pushing forward from the given information, ask "what would let me conclude this?" and work one step backward. If the goal is that two angles are equal, ask what would guarantee that - are they corresponding angles on parallel lines, are they angles in congruent triangles, are they both equal to a third angle. Working backward narrows the search dramatically, because instead of exploring everything you could possibly derive, you are only looking for the specific bridge the goal needs.
Congruence and similarity criteria as a checklist
Keep a short mental checklist of the standard criteria and run through them explicitly when a proof stalls: side-side-side, side-angle-side, angle-side-angle, angle-angle-side for congruence, and angle-angle for similarity. Treat it as a checklist to test against your diagram rather than a fact to be recalled from nowhere - go through each one and ask whether the marked-up diagram already supplies what it needs.
Writing the final proof
Once you have found the logical chain by working through the steps above, the write-up itself should be almost mechanical: state each fact, state the reason it is true (a given fact, a definition, a previously proven theorem), and move to the next line only once the previous one is justified. A common mistake is writing the proof in the order you discovered it, which is often messy and roundabout - it is worth reordering the final written version into the cleanest logical path, even if your scratch work took a longer route.
When you are truly stuck
- Re-read the given information slowly and check you have used every single piece of it - unused information is a strong hint about what step you are missing. - Try a slightly different diagram or a specific numerical example to see which relationships seem to hold, then try to prove that specific relationship in general. - Ask whether a related, simpler proof you have seen before shares the same structure - many proofs are variations on a small number of recurring patterns.