The two skills, separated
| Skill | Trained by | Tested by |
|---|---|---|
| Execution — doing the algebra correctly | Repetition of similar problems | Marks lost to arithmetic slips |
| Selection — knowing which method applies | Mixed problem sets where the method isn't announced | Almost every exam question worth real marks |
| Interpretation — knowing what the answer means | Sketching, checking limits, sanity-checking signs | Applied and word problems, and part (c) of everything |
Most calculus study is 90% execution. Most calculus exams are 60% selection. Closing that mismatch is the highest-leverage change available in this subject, and it costs nothing but a change in how you order your practice.
Practise selection deliberately
Textbook exercise sets are organised by section, which quietly destroys their value: every problem in section 7.3 is an integration-by-parts problem, so you never decide anything. This is blocked practice, and its weakness is precisely the selection skill — see interleaving for the general case.
- Build a mixed problem set by taking three problems each from six different sections and shuffling them. Do them without looking at which section they came from.
- Practise identification without solving. Take twenty integrals and, for each, write only the first move — substitution, parts, partial fractions, trig identity — in under fifteen seconds. Twenty of these takes five minutes and trains the exact skill the exam wants.
- Keep a decision list. "Product of a polynomial and an exponential → parts." "Odd power of sine → save one for the substitution." This is your real revision material, not the worked examples.
- Do old problems again cold, weeks later, from a mixed pile. If you can only solve a problem when you remember it's the parts one, you can't solve it.
How to use worked examples properly
- 1
Read only to the first line you couldn't have written
Cover the solution and reveal one line at a time. The moment you hit a line you wouldn't have produced, stop. That line is the entire lesson in the example; the rest is algebra you can already do.
- 2
Ask why that line, not what that line
"Why substitute u = x² here?" is the question. "What is du?" is not. Write the answer to the why-question in one sentence — that sentence goes on your decision list.
- 3
Close it and redo the whole problem from the statement
Not from where you got stuck — from the start. If you can't, you haven't learned the example, you've read it. This step is unpleasant and is where the learning is.
- 4
Then do a different problem needing the same move
Same technique, different numbers and shape. If the technique only works on the exact example, you've memorised a solution rather than acquired a method.
- 5
Add the trigger to your review queue as a question
"When do you use trigonometric substitution, and which one for √(a²−x²)?" Facts like these decay and cost you whole questions when they go. They belong in spaced review, not in a notebook you'll never reopen.
The prerequisites that are actually the problem
A large fraction of "I can't do calculus" is algebra and trigonometry failing under load. The calculus step is right and the answer is wrong because of a sign, a fraction, an exponent rule or a half-remembered identity. This is worth diagnosing honestly, because the remedy is completely different.
Go through your last twenty marked errors and label each: calculus concept, algebra slip, arithmetic slip, or misread question. Most students are startled by the ratio. If two thirds are algebra, then two thirds of your practice should be algebra — and it will lift your calculus marks faster than any amount of studying limits. How to learn from mistakes has the error-log method in full.
Concepts you cannot get away without
Calculus has a small number of ideas that everything else is built from, and students who struggle usually have one of them as a memorised phrase rather than a picture.
- The derivative as a rate, and as the slope of the tangent. If you can't sketch a function and sketch its derivative underneath, this isn't in place.
- The integral as accumulation, and as area with sign. The sign part is the bit that costs marks.
- The fundamental theorem as the statement that these two are inverse operations. Most students can recite it and can't say why it's surprising.
- Limits as behaviour near a point, not value at it. Almost every continuity and differentiability question is testing this distinction.
Test each by explaining it out loud in ninety seconds, no notation. The Feynman technique stall will find the one you're faking within a minute.
The week before the exam
Stop learning new techniques four days out and switch entirely to timed mixed sets from past papers. At this stage your marks are decided by selection speed and error rate, not by knowledge, and both only improve under time pressure. Mark them yourself against the scheme, categorise every lost mark, and spend the next day on whichever category was largest.
Bring the decision list, not the textbook. One page of "when you see this, do that" is worth more on the morning of the exam than four hundred pages of exposition.