This column is a practical guide to the 60 to 180 minutes inside an exam: how to stop losing points to the clock and to a one-row shift on the answer sheet. It's written for anyone sitting university entrance exams, TOEIC, Eiken, or professional certifications. It makes one claim: the points you lose on the day come from a gap in time design, not a gap in ability.

The reason you lose points on exam day is almost never "I didn't know it." It's that you ran out of time, or that your answer sheet was shifted by one row. You never reached a question you could have solved the night before. The answer you thought you'd marked has vanished by the time it's scored. Both wear the face of a gap in ability. The real cause is time that wasn't designed well enough.

The broader logistics of the day, such as how to spend the night before, your first ten minutes in the room, and recovering afterward, are covered in the Exam Day Checklist. This one is narrower. It deals only with the inside of those 60 to 180 minutes, from the start signal to the end. How to decide the order you solve in. Which questions to abandon. How to carve up your time. What to review first. How to prevent marking shifts with physics rather than willpower. And what to do with the time that remains.

One promise up front: I won't just list steps. Each step comes with one line on why it works. A step whose reason you don't understand is the first thing to evaporate under exam-day nerves. Once the reason has truly landed, your hand keeps moving even when your mind goes blank.

The design here isn't tied to one kind of exam. For university entrance exams, TOEIC, Eiken, and professional certifications alike, the logic of how you use your time on the day is the same. The format differs, but one thing never changes: not spilling the points you already have, within a limited time.

How do you allocate exam time? Pay your review time first

Work out a per-question cap with one formula, and subtract review time and a buffer before you solve anything. When the start signal sounds, don't pounce on question one. The first job isn't solving. It's dividing your time.

Per-question cap = (total time − review time − a buffer) ÷ number of questions

Here's an example. Say the exam is 80 minutes with 40 questions. Set aside 10 minutes of review and a 5-minute buffer first. That leaves 65 minutes. Divide by 40 questions and you get about 1.6 minutes per question. That's the cap: "I may not spend more than this on any one question."

The crucial part is to subtract the review time and buffer first.

→ Why it works: Review time you "do if there's any left" is never left. As you solve, time always gets eaten, and the final review disappears as "oh, zero minutes remaining." So you set it aside first. It's the same logic as why you'll never save money if you only save what's left after spending. Only what you subtract up front survives.

There are just three steps to run at the start.

  1. Grasp the total time and the number of questions. Flip through the booklet and check, at the start, how many minutes and how many questions you have.
  2. Subtract the review time and buffer. Lock in the remaining time as your solving time.
  3. Write the per-question cap in the margin of the booklet. Write "1.6 min/question." Don't try to hold it in your head.

With this formula in hand, you can see how many minutes you have left throughout the exam. When the remaining time is always visible, it builds a focus you can't pull yourself out of. Solve without the time visible, and somewhere along the way you will misjudge it.

What order should you solve in? Distrust "in sequence"

Don't follow the question numbers. Rebuild the order yourself, starting from the questions you're sure of. The booklet has numbers on it, but those aren't an instruction to solve in that order.

Four steps.

  1. In the first minute, flip through and gauge the difficulty. Don't start solving right away; look at the whole picture first. Get a rough sense of what's light and what's heavy.
  2. Start with the questions you're sure you can solve. Make your first question one you know you can do, not a hard one. Use it to build momentum.
  3. Clear the high-value sections while your head is fresh. Push them to the back and you'll be reaching for the big points with a tired brain.
  4. Raise the cognitive load from knowledge questions to thinking questions. Warm up on the light ones, then move into the ones that require thought.

→ Why it works: Chase your first success by certainty, not by difficulty, and your nerves drop physically. Solve one, and the feeling that "I'm moving" returns, and your hand stops freezing. A tired head also reliably fumbles the hard questions late in the exam. So put the high-value sections in the early half, when your head works best. You lose fewer points by pushing the low-value questions to the tired back half.

One caveat by exam type. In tests like the TOEIC or Eiken reading sections, where long passages tend to swallow your time, you can deliberately push the long passages to the back. Lock down the short, sure points first, then turn to the long passages with the time that's left. Getting swallowed by a passage, running out of time, and dropping even the short questions you could have solved is the single most common way to lose points on this kind of test.

How do you decide which questions to skip? Judge by break-even

Make the skip call by numbers, not by feel. The standard is the per-question cap from the previous section. Don't try to solve every question. Assume from the start that you'll abandon some. This isn't giving up. It's a genuine strategy for maximizing points in limited time.

  • The moment you exceed the per-question cap (that 1.6 minutes), mark it with a triangle and move on. The judge isn't the clock; it's the cap you set yourself.
  • If you don't see the path within 30 seconds to a minute, retreat. "I'm almost there" is the most dangerous trap. Even if you solve it after grinding, the time you spent doesn't come back.

The axis of judgment is break-even. Rather than spend five minutes on a 3-point question, spend those five minutes picking up three other questions for nine points. The same five minutes earns triple the points. Think of a skip call not as "giving up" but as "an investment decision," and your hand starts to move.

In steps:

  1. Force a retreat when you exceed the cap. Hold the line you drew yourself.
  2. Use two kinds of marks. Distinguish "no idea at all (×)" from "not confident, but I can see the path (△)."
  3. Return to the △ ones first in review. Don't chase the × ones; keep only the △ ones as your upside.

→ Why it works: A head worn down by a hard question goes on to drop the easy ones too. The true cost of grinding on one question isn't only that question's time. It includes the later points you drop with a worn-down head. So skipping isn't running away; it's maximizing scoring efficiency. And separating × from △ means that in review, "where do I go back to grow my score" is visible at a glance. Put the same mark on everything and that information disappears.

How do you prevent marking mistakes? Confirm with your finger, not with grit

"Being careful" won't stop a shift. Fix your marking method to one the night before, confirm the row with your finger whenever you skip, and mark in blocks of five. The biggest accident on a multiple-choice answer sheet is a one-row marking shift. Your ability could be perfect, but the moment you're off by one row, everything after it can go to zero. Grit won't stop a shift. Stop it with physics.

Four steps.

  1. Fix your marking method to one, the night before. Decide in advance whether you'll mark right after each question or mark everything at the end, so you don't waver on the day. Either is fine. Not wavering is what matters.
  2. If you skip a question, skip its answer row too, and confirm the number and row with your finger. Press your finger on the skipped question's number and the answer sheet's row number and confirm they line up, not with your eyes alone.
  3. At each section break, check the question-number-to-row-number match once. Place a checkpoint at the natural breaks to confirm you haven't shifted.
  4. Mark in blocks of five ("block marking") to localize chain shifts. Not one at a time, not all at the end, but five questions at a time. If a shift does happen, the damage stops inside that block of five.

→ Why it works: Most marking shifts happen the instant the method changes mid-exam. At the seam where you unconsciously mix mark-as-you-go and mark-at-the-end, you shift by one question. Fix the method to one and the source itself disappears. One skipped question becomes the biggest accident there is, shifting every row after it by one. Physically confirming the row with your finger the moment you skip is the only reliable prevention. Eye-only confirmation gets missed easily by a tense head. Touch it with your finger and the misses physically vanish. Block marking works because even if a shift occurs, the damage is sealed within five questions. Mark everything at the end, notice the shift then, and it's already too late.

The one point that overlaps with the separate column on day-of logistics is this marking-shift discussion. That's exactly why this column adds the deeper tools the other one doesn't have: block marking, and the break-even of skip calls. Overlapping points are precisely where you dig deepest, bringing a new tool.

What should you review first? Review everything and you fix nothing

The full-row marking check comes first. Then the △ questions, then calculation and transcription errors, then blanks. Here's where many people fail with the review time they set aside: trying to review everything, from the start, in order. Do that and you run short, the back half goes sloppy, and in the end you fully tighten nothing. Review has a clear priority order.

  1. Make the full-row marking check the top priority. This is accident prevention with the largest point value. If one shifted row erases several questions' worth of points, preventing it is the highest return.
  2. Return to the △, not-confident questions. This is where the upside is largest. Don't chase the × ones; go for the △ ones.
  3. Re-check calculation and transcription errors. Did you drop a question you'd actually solved to a careless slip?
  4. Confirm there are no blanks. Scan last for anything you forgot to fill in.

In time terms: if review is 10 minutes, 3 minutes for the marking check, 5 minutes for the △ ones, 2 minutes for the blank scan. Decide this split in advance too, not on the spot.

→ Why it works: Review time always runs short. So starting in order of greatest upside maximizes the points you can pick up in limited time. Review everything evenly and you never reach the highest-return marking check, and everything ends up half-done. A priority order means deciding first what to abandon. The marking check goes first because it's the most cost-effective review there is: one check protects several questions' worth of points.

What should you eat on exam day? Balance digestive load against reassurance

Leave an hour between the meal and the exam, and test sugar and caffeine before the real day. Beyond that, decide by the balance between digestive load (body) and reassurance (mind). On a different axis from time allocation, your physical condition ties directly to your score on the day. The pre-exam meal in particular is easy to dismiss, and it tells.

Honestly, I can only write generalities here, and I shouldn't write more. How sugar and caffeine hit you varies widely from person to person; there's nothing you can flatly declare as "eat this and your focus rises." So the only thing I can say is: test it on your own body in advance.

But let me write one core point. Heavy, hard-to-digest food like a pork cutlet is said to tend to work against you before an exam. Generally, it's said that blood flow and energy go to digestion, and your head tends to work poorly for a while after eating. That said, if the reassurance of eating what you love lowers your nerves, then choosing to eat by mood is perfectly valid too. Digestive load (body) versus reassurance (mind). Deciding by that balance is the essence. It's not a matter to settle by nutrition theory alone.

What I can say within a general range is about this much.

  • Your head is said to tend to feel foggy right after eating, so leave an hour between the meal and the exam.
  • Sugar and caffeine vary a lot by person, so try them once before the real day and learn your own reaction. Don't take them for the first time cold on exam day.

→ Why it works: Condition ties directly to your score. But the reassurance of "the same as always" often protects your day-of focus better than a perfect meal. Even if it's nutritionally optimal, putting something unfamiliar in on the day and leaving your stomach unsettled backfires. Settling yourself isn't about doing something special. It's about bringing your body to a state where you can put out your maximum. Think of it as condition management not for being comfortable, but for putting out everything you've got.

What do you do with the remaining time? Two branches: time left over vs. time short

If time is left over, start with the final marking check. If time is short, start by filling every blank. The endgame of an exam branches into two. Did time run over, or run short? Decide both moves in advance so you don't waver either way.

If time is left over, move like this.

  1. Do a final physical check of every marked row. The first thing you do with spare time is not a new question, but the final marking-shift check.
  2. Take another run at the △ questions. Return to where upside remains.
  3. Check written answers for typos. Pick up the leftovers in what you've written.

If time is short, move like this.

  1. Leave no blanks. For multiple choice, fill them by probability even when you don't know. A blank is a guaranteed zero, but filling it makes the expected value positive.
  2. For written questions that earn partial credit, write what you can, even partway. Even if you can't fully answer, you may pick up points for what you wrote.
  3. Put the last three minutes entirely into checking for missed marks. Use the final few minutes not for new questions, but for stamping out anything you forgot to fill in.

→ Why it works: A blank is a guaranteed zero, but filling it shifts the expected value positive. Even just guessing on a four-choice question lands one in a few. How you use the last few minutes swings one to several points. That handful of points decides pass-or-fail more often than you'd think. So in the endgame, don't move by feel. Trace the routine you decided in advance.

Here is one fixed routine for the last five minutes. When five minutes remain, don't touch new questions. Fill blanks by probability → confirm every marked row with your finger → check written answers for typos. Keep your hand moving in this order until the signal to stop.

Closing: knowledge is the job of the days before; the day is only about moving it without dropping a point

Here is the whole design again, one line each. Allocate by formula, paying your review time first. Build the order from certainty. Judge skip calls by break-even. Confirm marks physically, with your finger. Run review in order of greatest upside.

Building knowledge is the job of the days before. What you can do on the day is not to learn something new. It's to move the ability you already have onto the answer sheet without spilling a single point. Time allocation and mark management are nothing more than the foundation that supports that "not spilling." But several points swing right here.

passed.jp has "passing" in its name. I hope the version of you who shows up on exam day picks up even one of the designs written here, and gets to put out everything you've got. If you can put it all out, the result follows on its own.

FAQ

When and how should I decide my exam time allocation?

Right after the start signal, before you touch question one. Check the total time and the number of questions, subtract review time and a buffer first, and divide what's left by the number of questions. For an 80-minute, 40-question exam, subtracting 10 minutes of review and a 5-minute buffer leaves 65 minutes, or about 1.6 minutes per question. Write that number in the margin of the booklet.

What's the rule for deciding to skip a question?

If you exceed the per-question cap you set yourself, mark it with a △ and move on immediately. Retreat as well from any question where you can't see the path within 30 seconds to a minute. The axis is break-even: rather than spend five minutes on one 3-point question, spend the same five minutes picking up three other questions for nine points. A skip call is an investment decision, not giving up.

What's the most reliable way to prevent a one-row shift on the answer sheet?

When you skip a question, skip its answer row too, and press your finger on the question number and the row number to confirm they line up. Eye-only checks get missed by a tense head. On top of that, fix your marking method (mark-as-you-go or mark-at-the-end) to one the night before, and mark in blocks of five so any shift is sealed inside five questions.

I only have 10 minutes to review. What should I prioritize?

In this order: the full-row marking check, the △ questions, calculation and transcription errors, then blanks. With 10 minutes, split it as 3 minutes for the marking check, 5 minutes for the △ questions, and 2 minutes for the blank scan. Review everything evenly and you never reach the marking check, which has the highest return.

What should I do in the last five minutes of an exam?

Don't touch new questions. Fill blanks by probability, confirm every marked row with your finger, then check written answers for typos, and keep your hand moving in that order until the signal to stop. A blank is a guaranteed zero, but even guessing on a four-choice question lands one in a few, so the expected value moves positive.