Optics alone is a fifth of the paper — more than Electrostatics, more than any other unit — and it is chapters 9 and 10, which is exactly where people run out of year. Electrostatics and Magnetism together are 21, and they are at the front, which is why everyone is comfortable with the first quarter of the syllabus and shaky on the part that carries the most marks.

The two cheapest units are Electromagnetic Waves at 4 and Electronic Devices at 7. Between them that is 11 marks for a very short syllabus, and Electromagnetic Waves in particular is essentially one table of the spectrum and its uses. If you are triaging, those two are the highest return per hour in the subject.

This is the thing my script makes visible that nothing else does. The evaluator writes the sum on the page: a five-mark answer of mine reads "31a : 1 + 0.5 + 0.5 + 0.5 + 0.5 + 0.5 = 3.5". Every fragment is scored separately, and you can watch the mark being assembled.

What that means is that a Physics long answer is not one answer, it is six to ten scoring slots. The formula quoted correctly is a slot. The diagram is a slot. The substitution is a slot. The final value with units is a slot. You can get the final answer wrong and still take most of the marks — and you can get it right, skip the derivation, and take very few.

It also means the reverse, and this is the part that cost me. Q26 on my paper reads "0.5 + 0 + 0 + 0 + 0 + 0 = 0.5". Six slots, five empty, on a question I had genuinely half understood. Being nearly right in your head and putting almost nothing on the page is the specific way Physics marks disappear.

Question 32 on my paper was a five-mark electrostatics question. I filled a page: field due to a dipole, the algebra expanded and simplified, a boxed final expression. It reads as a confident answer.

It scored 0 + 0 + 0 + 0 + 0 = 0.

I had answered a different question from the one asked. Nothing about my working was rewarded, because the marking scheme had five specific things in it and none of them were what I wrote. That is the single most expensive mistake available in a Physics paper, and it is not a knowledge failure — it is a reading failure.

The alternative I also attempted, 32b, scraped 1 out of 5. Between them I put nearly two pages of work into ten marks and came away with one.

This is why the fifteen minutes of reading time matter more in Physics than anywhere else. You cannot recover a misread question with effort. The page is published if you want to see what a confident zero looks like.

Section D is two case-based questions, four marks each — eight marks of the paper. On my script both are full marks, 4/4 and 4/4, and they were the easiest eight marks I got.

They are built to be answerable from the passage plus one standard idea. The photoelectric one gave me the setup and asked four short questions about it; the galvanometer one did the same. There is no derivation, no long numerical, and no risk of misreading a five-marker.

People leave these until last because they look long on the page. They are not long. Do them early while you are fresh and bank eight marks.