Originally published on Edu Global Institute.
Key takeaways
- The IPhO syllabus explicitly tests 'creativity and understanding of physics rather than mathematical virtuosity or speed of working'.
- Theory is 30 marks over three problems; the experimental paper is 20 marks - so 40% of the result is decided in a laboratory.
- Most students spend almost no preparation time on experimental skills, preparing for 60% of the examination.
- Uncertainty and error analysis is where most experimental marks are lost, and the marking rules require stating how the uncertainty was derived.
- Setting your uncertainty to cover the gap with the literature value is reasoning from the answer - error analysis comes first.
- Problems may use concepts outside the syllabus if the text explains them, so reading an unfamiliar definition under time pressure is itself a trainable skill.
- Problem texts are capped at 12,000 characters and numerical work must be manageable with standard high-school mathematics.
There is a line in the official IPhO syllabus that most physics olympiad preparation quietly ignores.
The statutes state that problems should be designed to focus on "testing creativity and understanding of physics rather than testing mathematical virtuosity or speed of working." They go further: mathematical manipulation should carry only a small share of the marks, and students who reach a sound approximate solution to a mathematically demanding task should receive partial credit.
Now compare that with how most students prepare — grinding through problem sets, optimising algebraic speed, treating physics as applied mathematics with heavier notation. The syllabus says, in writing, that this is not what is being rewarded.
The forty percent almost nobody trains for
The IPhO examination has two papers.
The theoretical examination is three problems over five hours, worth 30 marks, spanning at least four areas of physics. The experimental examination is one or two problems over five hours, worth 20 marks.
That is 40% of the total on experimental work — two fifths of a student's result decided in a laboratory.
Ask a student preparing for a physics olympiad how much of their preparation time goes on experimental skills. In our experience the honest answer is usually close to none. They are preparing for 60% of the examination and hoping the rest resolves itself.
This is the largest and most correctable inefficiency in physics olympiad preparation, and it is hiding in plain sight in the official rules.
Where experimental marks actually go
Not, as students assume, in getting the measurement wrong.
The dominant loss is uncertainty and error analysis. The IPhO marking rules require data analysis and uncertainty estimation, and require the student to state how the uncertainty was derived, whether graphically or theoretically. Marks go when a student produces a beautiful measurement and a bare number with no error attached, or an error quoted with no account of where it came from.
Three further patterns recur:
Presentation penalties. Rubrics commonly include explicit marks for indicating uncertainties in tables and keeping only meaningful digits. These feel like formatting. They are marks.
Backwards error analysis. Students compare their result with the accepted literature value and then set their uncertainty to cover the difference. Error analysis must be done before that comparison; doing it afterwards is reasoning from the answer.
Designing the experiment. The statutes say some experimental procedures are deliberately left for students to design, and that the experimental paper should test creativity rather than speed at technical tasks. A student who has only ever followed printed lab instructions has never practised the thing being assessed.
What the syllabus says about unfamiliar topics
A detail that changes how you should prepare: problems may use concepts outside the syllabus, provided the problem text explains them and they connect to syllabus topics.
So a student can meet, in the examination, a concept they have genuinely never studied — with a paragraph of explanation and an expectation that they will use it immediately.
That is a skill in itself, and it is trainable: reading an unfamiliar definition under time pressure and applying it correctly. It is also the clearest possible statement that memorised coverage is not the goal. You cannot memorise your way past a concept introduced on the paper.
The statutes add two practical constraints worth knowing: problem texts are capped at 12,000 characters, and numerical work must be manageable with standard high-school mathematics without extensive calculation. If a problem appears to require heavy computation, the student has almost certainly missed the shorter route — which is usually the physics being tested.
Why strong theorists lose
The pattern we see most often is a student who is genuinely excellent at theory and finishes outside the medals.
It is rarely mysterious. They scored well on 30 marks and poorly on 20. Because their theory is strong, every practice session feels productive — they are improving at something real, just not at the thing costing them the result.
This is why a diagnostic matters more in physics than in almost any other subject we teach. A single theory paper will tell you a student is strong and tell you nothing about why they are not medalling.
What good experimental preparation looks like
Four things, in order of return.
Uncertainty as a taught subject. Propagation through formulae, combining independent sources, deciding which dominates, and stating the derivation. Taught explicitly, not absorbed from lab work.
Linearisation and fitting. Choosing what to plot so a relationship becomes a straight line, extracting constants from gradient and intercept, and reading uncertainty off a graph. This is a large share of experimental marks.
Experiment design. Given apparatus and a quantity to measure, decide the method yourself. Uncomfortable at first for students trained on printed instructions, which is exactly why it must be practised.
Writing up under time. Tables with uncertainties, appropriate significant figures, a clearly stated final result. The marks exist for this whether or not a student believes they should.
The wider point
IPhO 2027 — the 57th — runs from 4 to 12 July 2027 in Riyadh.
The students who do well there will not be the ones who solved the most problems. They will be the ones who can meet an unfamiliar concept, reason about it carefully, measure something honestly and say precisely how confident they are in the number.
That is what the syllabus says it is testing. It has said so for years. The gap between that statement and how physics olympiad preparation is usually run is, in our view, the largest unexploited opportunity in the subject.
Related reading
Frequently asked questions
How much of a physics olympiad result is experimental?
At IPhO, the theoretical examination is 30 marks over three problems in five hours, and the experimental examination is 20 marks in five hours. That makes the experimental paper 40% of the total - two fifths of the result decided in a laboratory, and the part most students barely prepare for.
Where do students actually lose experimental marks?
Predominantly in uncertainty and error analysis rather than in taking the measurement. The marking rules require data analysis and uncertainty estimation, and require students to state how they derived the uncertainty. Presentation marks also go regularly - indicating uncertainties in tables and keeping only meaningful digits are explicit rubric items.
What is 'backwards error analysis'?
Comparing your result with the accepted literature value and then setting your uncertainty so it covers the difference. It is reasoning from the answer. Error analysis must be completed before any comparison with a published value, and examiners can see when it was not.
Can the exam include topics outside the syllabus?
Yes. The statutes allow problems to use concepts outside the syllabus provided the problem text explains them and they connect to syllabus topics. So a student may meet a genuinely unfamiliar concept with a paragraph of explanation and be expected to apply it immediately - which makes reading an unfamiliar definition under time pressure a skill worth training.
My child is strong at theory but not medalling. Why?
The most common explanation is a strong score on the 30 theory marks and a weak one on the 20 experimental marks. Because their theory is genuinely good, every practice session feels productive - they are improving at something real, just not at the thing costing them the result. A theory-only diagnostic will not reveal this.
How do you prepare experimental skills without a full laboratory?
The highest-return components are not equipment-dependent. Uncertainty propagation, choosing what to plot so a relationship linearises, extracting constants from gradient and intercept, and writing up with correct significant figures are all taught and practised on paper and with data sets.
Does heavy calculation mean I am doing it right?
Usually the opposite. The statutes require numerical work to be manageable with standard high-school mathematics and without extensive calculation. If a problem appears to demand heavy computation, the likeliest explanation is that a shorter route was missed - and that route is usually the physics being tested.
Sources
Read the original, with the full formatting and any later updates, on Edu Global Institute.