How to Actually Study Your Class Notes: Turning Raw Pages into Long-Term Memory

Decide the exam date first, then set the gap. If about five weeks will pass between your review session and the paper, the evidence points to leaving roughly eleven days between first learning the material and that review; for a ten-week stretch, about three weeks. Those numbers come from spacing experiments, not from the forgetting curve, and the difference matters because the curve cannot give you intervals at all. Erring long costs far less than erring short.

Anyone searching how to study notes meets the same picture within three clicks: a decaying curve with review points marked on it at one day, one week, one month. The curve is real. The review points were added by someone else.

The notes revision kit at the end has the 14-day countdown, a blurt sheet and a gap calculator, in Word and PDF.

The Ebbinghaus curve in action: exact review intervals for college courses

Start with what the curve is, because its provenance decides how much weight it can carry.

Hermann Ebbinghaus measured forgetting on himself in the 1880s, using lists of nonsense syllables and the method of savings, which records how much relearning effort a list needs after a delay rather than how much you can recall. Murre and Dros repeated the procedure faithfully more than a century later: one subject, about 70 hours of learning and relearning, with delays of 20 minutes, 1 hour, 9 hours, 1 day, 2 days and 31 days. The curve replicated. They also reported that it is not perfectly smooth, most probably showing a jump upwards starting at the 24-hour point (Murre & Dros, 2015, PLOS ONE, 10(7), e0120644).

Note what that study is and is not. It is a careful replication on a single subject learning syllables with no meaning. It describes the shape of forgetting. It does not tell you when to review your biochemistry notes, and no amount of redrawing it will make it do so.

For intervals you need a different experiment, one that varies the gap deliberately and measures what happens at the exam rather than during relearning. Cepeda and colleagues ran it at scale: 1,354 participants learned 32 obscure facts to criterion, returned after a gap of up to 3.5 months for a second session, then sat a final test after a further delay of up to a year (Cepeda, Vul, Rohrer, Wixted & Pashler, 2008, Psychological Science, 19(11), 1095–1102).

The results give you something the curve never could. One thing to get right before reading the table: in this design the delay is counted from the review session to the test, not from when you first learned the material. So a row reads as learn, wait the gap, review, then sit the test that far later.

Those are recall figures, measured against a zero-day gap with study time held constant. Read the third row again: for material you need in ten weeks, spacing the review correctly rather than restudying straight away roughly doubled what was recalled, for the same total minutes.

Two things to take from the shape of the data rather than the rows.

The optimal gap is a proportion of the delay, not a fixed number. It sits around 20% of the test delay for delays of a few weeks and falls to about 5% at a year. So the question you have to answer before scheduling anything is how long you need to hold the material, which for a cumulative final is longer than most students assume.

Errors are not symmetric. Accuracy climbs steeply as the gap approaches optimal, then declines gently past it. Overshooting the gap costs far less than undershooting it, which is the opposite of how most revision timetables are built.

Honest limits on all of this. The material was isolated trivia facts rather than connected course content, and the design had two study sessions rather than a term's worth. The intervals above are the best available starting points, not a prescription. What generalizes reliably is the direction: longer gaps than feel comfortable, chosen against the date of the exam.

Blurting method vs flashcards: which generates better retention for finals?

The honest answer first, since it is more useful than a winner.

No published study compares blurting with flashcards under those names. What has been compared is the underlying pair: free recall, where you write everything you can about a topic from a blank page, against targeted recall, where a specific prompt asks for a specific thing. Blurting is the first. A flashcard is the second.

Endres and colleagues ran that comparison within subjects: 54 university students studied two expository texts, then practiced retrieval on one with short-answer questions and on the other with a free-recall task. Short-answer practice produced better retention of the information those questions had targeted. Free recall produced better retention of the other material in the text, the parts no question had pointed at. The two formats also differed off to the side: short-answer practice improved metacognitive calibration, meaning students' sense of what they actually knew, while free recall raised self-efficacy and situational interest (Endres, Kranzdorf, Schneider & Renkl, 2020, Instructional Science, 48(6), 699–728).

So neither method wins. They keep different things, and the choice follows from the paper you are sitting.

Cards suit anything with a defined answer: mechanisms, reagents, dates, definitions, the forty reactions on an organic chemistry sheet. You will remember exactly what you carded, and you will have an accurate sense of which cards you have not learned, which is the quieter benefit and arguably the larger one.

Blurting suits anything where the exam asks you to produce structure: compare two mechanisms, explain why a synthesis takes one route rather than another, write six paragraphs on a prompt you have not seen. It also catches the material you never thought to card, which is precisely the material a card deck cannot warn you about.

The practical arrangement is not fifty-fifty. Card the items with defined answers, blurt the topic once a week with the cards face down, and treat anything that fails to appear in the blurt as a gap your deck was never going to find.

One caution about blurting specifically. A blurt is only retrieval if the notes stay shut. Blurting with the page open is copying with extra steps, and it feels almost identical while you are doing it.

The 14-day finals countdown

Two weeks is the window most students actually have, which is shorter than the optimal gaps above. The schedule below works within that constraint rather than pretending otherwise: it front-loads the first pass so that the longest possible gap sits inside the fortnight.

The gap doing the work is day 13 to day 7, six days on the first set of topics. Worth being straight about what that buys you. With the exam a week after the second blurt, you are in the shortest row of the table, where the measured gain from optimal spacing was 10%. Real spacing gains need a longer run-up: 59% when five weeks separate the review from the paper, 111% at ten weeks.

So the fortnight plan is damage control, not the method working at full strength. If you have five weeks, do the first blurt now and the second eleven days later, and this countdown becomes the last fourteen days of that schedule rather than the whole of it.

Days 10 to 8 are deliberately card-only. Blurting daily feels productive and destroys the gap you are trying to create.

Where Sovi.AI fits

The limit first, because the work this article describes is not the work a tool can take over.

No generated question set can stand in for the one you write yourself, because writing the questions is where you decide what the exam will ask. That decision is the first pass of the method above, and handing it over removes the part that does the work. Check your own account for what is available before planning around it, since the feature list moves. What this section needs is there: the quiz generator in AI Study takes the notes as a PDF and returns practice questions, alongside summaries and knowledge maps.

Use it after your own deck, not instead of it. A quiz you did not write tests the material you did not think to card, which is the same blind spot blurting addresses from the other side. Running both is not redundant.

Live Recording covers the upstream half, transcribing a lecture as it happens and converting it into structured notes, so that the raw pages this article starts from exist at all. Check the spoken-language menu in the recorder before you rely on it for a course taught in anything other than English. For condensing a term into something you can carry into the final week, Cheatsheet takes uploaded material and produces a review sheet, with a formula-sheet option for quantitative courses.

The prompts for getting a model to quiz you rather than explain at you are in the companion guide on using AI for studying. What none of it replaces is the blurt, because the blurt is the measurement.

The notes revision kit

• notes-revision-kit.pdf to print and pin somewhere visible.
• notes-revision-kit.docx to adjust the dates to your own exam.

Inside:

1. The 14-day countdown as a dated grid, with a column to tick each session off, a planned-time column and a column for the topics each blurt exposed.
2. A blurt sheet, blank except for the topic line and a timer box, with a narrow margin on the right for marking what you missed when you check against the notes.
3. A gap calculator, the four measured test-delay and gap pairs from the spacing study set out as a lookup table, with lines for your own exam date and the review dates that follow from it.

Fill the gap calculator before the countdown. If your exam is more than three weeks out, the countdown is the tail end of your schedule rather than the whole of it.

Frequently asked questions

1. Why not just review every day?

Because daily review removes the gap, and the gap is where most of the benefit lives. In the spacing data, the optimal gap roughly doubled recall over immediate restudy at a ten-week horizon, with study time held constant.

2. What if my exam is in three days?

Then spacing is not available to you and the question is different. Blurt, repair, blurt again, and accept that this is a performance for Friday rather than learning for the degree.

3. How many cards is too many?

When you stop getting through the deck daily, it is too big. Cards are for defined answers only; anything needing a paragraph belongs in the blurt.

4. Does the forgetting curve mean I should review within 24 hours?

Not on the evidence above. The 24-hour advice comes from schedules built on top of the curve, not from the spacing experiments, and those experiments found longer gaps better for anything you need to keep beyond a week.

5. Should I rewrite my notes neatly first?

No. Rewriting is copying, and it is the activity that most reliably feels like studying without being it. Knowing how to study notes is mostly knowing which activities only resemble studying; sort them by topic and leave them ugly.

6. How do I blurt a math or problem-based course?

Redo problems on blank paper rather than describing methods. Writing out the steps of a method you cannot execute is the same false signal in a different form.