Systems check: studying subjects you calculate, not read

There is a moment on every problem sheet where you read the model solution, follow it line by line, and conclude you could have done that yourself. Usually you could not. Subjects you calculate rather than read punish that confusion harder than any other: the exam never asks whether the steps made sense to you, it asks you to produce them on an empty page.

Following a solution is not solving it

Reading a worked solution is comfortable work. Each line follows from the one above and nothing is missing. What you are actually doing is checking that the author made no mistakes, which is a real skill and not the one being examined.

The counter test is cheap. Close the solution, take a blank page, and redo the problem from the first line: the actual arithmetic, the actual code, the actual derivation, down to the last sign. If you stall within thirty seconds, you had not solved it, you had watched someone solve it. Most people run that test once, find the gap, and go back to reading solutions. The gap was the information.

Redo the errors, not the sheets

The standard revision plan in a calculating subject is to work through every problem sheet again. It looks thorough and it is mostly waste: a large share of those problems you can already do, and doing them twice only feels busy.

The alternative costs about a page a week. Keep an error log, one line per problem you got wrong: what was asked, what you did, what the correct move was, and why you did not see it. Longer than three sentences and you will stop.

The value arrives two weeks later, when the same few situations turn up again and again. That short list is the real syllabus of your revision; redoing the sheets would have buried it under everything you already get right.

Four kinds of mistake

Wrong is not a category. Before an error goes into the log, decide which of four things happened, because they lead in four different directions.

A conceptual gap means you did not know what the object was or what the theorem said. More practice will not touch it, since practice only rearranges what you already understand. That one goes back to the book, the lecture, or a person.

A wrong method means you understood the material and reached for the wrong tool. The cure is a short list of trigger conditions rather than more problems: you are training the decision, not the execution.

An arithmetic slip or a sign error means the method was right and your hand was not. This is the only one of the four that genuinely responds to volume, and it responds faster to one step per line and a plausibility check at the end.

A misread question means you solved something nobody asked for. The fix sits before the calculation: read twice, mark what is given and what is wanted, write both down before starting.

Work under the clock

Speed in the exam room is a separate ability and has to be trained separately. You can be able to do every problem in the course and still run out of time, because the version of you that solves one in forty relaxed minutes has never had to do it in twelve.

Practise it directly. Take four problems, allow yourself the ratio of minutes to marks the exam uses, and stop when the time is up. What you learn is mostly where the minutes go: the staring before you commit to an approach, the step recalculated although it was already right.

Timing everything from day one backfires; learn a topic slowly, then put a clock on it.

When to look at the solution

Sitting for an hour with no idea is not perseverance, and opening the solution after ninety seconds is not studying either. Fix a number in advance, ten or fifteen minutes for an ordinary exercise, and hold to it.

When the time is up, the rule matters more than the number: read only the next step. One line, then close the solution and carry on alone. Almost always that line was the whole gap, and the rest of the problem stays your own work. Reading the full solution instead turns a problem you nearly had into one you have now watched, and every problem where you needed the hint comes back in a few days, from scratch.

Build your own formula sheet

A downloaded formula sheet is someone else’s compression of the course. Everything is on it, which is the problem: it draws no line between what you know cold and what you never understood.

Write your own from memory, then check it against the material. The writing is the exercise and the sheet is the by-product. Whatever you cannot reproduce without looking is a gap for the log. Add the conditions under which each formula holds, since many written exam errors come from applying a correct formula where it does not apply. And if your exam permits a sheet, a page you wrote yourself can be navigated under stress, while someone else’s is a document you have to read first.

Mix the problems up

Problem sheets arrive sorted by chapter, and that sorting quietly does half the work. When every question on the page uses the same technique, you never have to decide which technique it is, and that choice is usually the hardest part of an exam question.

Two or three weeks out, stop practising by chapter. Pull problems from across the course, shuffle them, and work them without knowing what each one is about. It will feel worse and your accuracy will drop, and that drop is the honest reading: sorted practice measured your execution, mixed practice measures your choice.

Past exam papers are the easiest source, not for the questions but for the mixture.

Where drilling stops working

All of this assumes the course is mainly about producing correct results. Proof-heavy courses are not. In analysis or abstract algebra the product is an argument in words, and an argument can be logically wrong while every symbol on the page is right. Drilling problems there without learning to read and write proofs produces someone who computes everything and proves nothing. Those courses need the same blank page applied to statements: reconstruct in sentences why a theorem holds.

None of this is quick either: an error log says nothing useful for the first few weeks, and mixed practice makes your results look worse before it makes them better.

If you run your sessions on a timer that logs each one as a flight, note there which problems needed a hint. Over a few weeks that record says more about where you stand than any feeling on the evening before.