Walking out, Priya grabbed his arm. “Question 4? The one with the diffraction grating and the two wavelengths? Did you use nλ = d sinθ or the small angle approximation?”
By hour 54, Daniel had developed a system. He’d created a spreadsheet: columns for paper code, question number, marks lost, error type (calculation, conceptual, units, careless), and a final column titled “Examiner’s Actual Thought.” For the geostationary orbit question, he wrote: “We want to see if you know E = -GMm/2r, not just ‘energy decreases.’” For the alpha decay: “We want to see if you check units automatically. 1 MeV = 1.6e-13 J. Memorize it.” For the diffraction experiment: “We want creativity, not a textbook method. Show you understand the physics, not the recipe.” past papers a level physics
Daniel thought for a second. The small angle approximation would have lost a mark—the angles weren’t small. He’d caught that because the 2019 paper had the exact same trick. Walking out, Priya grabbed his arm
He flipped to 2021 Paper 1, multiple choice. Question 17: A particle moves in a circle with constant angular speed. Which graph shows the variation of its acceleration with time? The obvious answer—a sine wave—was wrong. Centripetal acceleration for uniform circular motion is constant in magnitude, only direction changes. The graph should be a straight line. He’d chosen the sine wave in his first attempt. The mark scheme said: B (straight line). Common distractor: C (sinusoidal). He drew a star next to it. Did you use nλ = d sinθ or the small angle approximation
On the fifth day, he did the 2025 specimen paper—the one that hadn’t been used yet, just released for practice. Question 8 was a beast: a particle in an infinite square well, then a perturbation, then a probability calculation involving Fourier series. He’d never seen anything like it. His first attempt: 34 out of 70. He spent three hours on that single question, reading the mark scheme, then the textbook, then the examiner’s commentary. By the end, he understood not just the answer but why they’d asked it: because the old questions on quantum tunnelling had become too predictable, and they wanted to separate the memorizers from the thinkers.
His phone buzzed. His friend Priya: “u done 2022 P5? the planning q is brutal.”