How Many Questions Should You Attempt in VARC, DILR & Quant for 99+ Percentile?
A guide to calculating your own optimal CAT attempt count instead of chasing a borrowed number, built around the VALUE method. Includes an illustrative expected-value table and a section-by-section breakdown of how VARC, DILR, and QA attempt strategy differ.
"How many questions should I attempt for 99 percentile?" is one of the most asked CAT prep questions, and it has no single correct number.
Any specific number you've seen quoted, whether from a forum, a senior, or a coaching brochure, was true for that person, in that section, in that year, at their accuracy level. Applied to your own attempt with a different accuracy rate, the same number can be actively harmful advice, pushing you into questions where your expected outcome is negative.
This guide replaces the search for a magic number with a method for calculating your own optimal attempt count, section by section. We call it the VALUE method, because that's what attempt strategy actually comes down to: the expected value of each additional question, not the count.
- There's no fixed attempt number for 99+ percentile — it depends on your accuracy, not just your count.
- VALUE: Verify your accuracy, Apply expected-value math, Lock a ceiling, Use buffer time, Evaluate and adjust.
- Attempting a question below your demonstrated accuracy threshold can lower your net score, even as your attempt count rises.
- TITA and MCQ questions need different attempt logic, since only one of them typically carries a negative-marking penalty.
Why "how many should I attempt" has no single answer
Your net score isn't a function of attempt count alone. It's a function of correct answers, wrong answers, and the marking scheme's penalty for each, which means two candidates who attempt the exact same number of questions can end up with very different scores, purely based on accuracy.
Chasing an attempt-count number heard from a senior or a forum without knowing their accuracy rate. A senior who attempted 24 QA questions at 75% accuracy and you attempting 24 at 45% accuracy are running completely different strategies that happen to share a number.
The number that actually matters isn't how many questions you attempt — it's how many questions you attempt where the odds are in your favor. That's what the rest of this guide calculates.
Who should read this guide
This guide is for you if any of the following sounds familiar:
- You've set an attempt-count target copied from someone else's result without knowing their accuracy.
- You've attempted more questions in a mock and ended up with a lower score than a previous, lower-attempt attempt.
- You're not sure whether to guess on a question you're moderately unsure about.
- You treat VARC, DILR, and QA with the same attempt-count target, despite them working differently.
If none of that sounds familiar, skip ahead to the section-by-section breakdown and apply it directly.
The VALUE method for attempt strategy
The fix replaces a borrowed number with your own calculation, based on your actual accuracy and the exam's actual marking scheme. We call it the VALUE method, because every attempt decision comes down to expected value, not volume.
V — Verify your accuracy
Before calculating anything, check your actual accuracy rate per section from your last several mocks, not a guessed or hoped-for number. Break it down further by confidence level if you can: questions you felt certain about, versus questions you attempted despite real uncertainty.
Use at least three to five recent mocks to calculate this, not one. A single mock's accuracy can swing widely based on that day's paper and your form; an average across several gives a far more honest number to plan around.
A — Apply the expected-value math
Once you know your accuracy, the marking scheme tells you whether attempting at that accuracy is worth it. Under a commonly cited CAT MCQ marking scheme of plus 3 for correct and minus 1 for wrong, here's what expected value looks like at different accuracy levels:
| Accuracy | Expected value per MCQ attempted | Verdict |
|---|---|---|
| 20% | (0.20 × 3) − (0.80 × 1) = −0.2 | Below breakeven — skip |
| 25% | (0.25 × 3) − (0.75 × 1) = 0 | Breakeven |
| 40% | (0.40 × 3) − (0.60 × 1) = 0.6 | Worth attempting |
| 60% | (0.60 × 3) − (0.40 × 1) = 1.4 | Strongly worth attempting |
| 80% | (0.80 × 3) − (0.20 × 1) = 2.0 | Always attempt |
Illustrative math based on a commonly cited +3/−1 MCQ marking scheme. Always verify the current year's exact scheme in the official CAT notification, and recalculate the breakeven point if it differs.
Pure random guessing among four options sits right at this scheme's breakeven, roughly 25%. Eliminating even one obviously wrong option before guessing pushes your odds meaningfully above breakeven — which is why partial elimination is worth the few seconds it costs, even under time pressure.
TITA (type-in-the-answer) questions typically carry no penalty for a wrong answer under CAT's usual marking scheme, which changes this math entirely: if you have any reasonable approach, the expected value of attempting is positive or at worst zero, since there's no downside beyond the time spent.
L — Lock a ceiling, not a floor
Once you know your breakeven accuracy, set a maximum attempt count for each section, the point past which remaining questions would need pure guessing to attempt. This is a ceiling to stop at, not a floor to reach no matter what.
If you finish reviewing every question above your accuracy threshold with time still remaining, that's a signal to double-check high-confidence answers, not to lower your standards and start guessing on questions you'd otherwise skip.
U — Use buffer time on high-confidence answers
Leftover time is more valuable spent verifying answers you're already confident about than rushing into new, low-confidence questions. A careless slip on a question you understood well costs the same negative marking as a wrong guess on one you didn't.
E — Evaluate and adjust
Your accuracy threshold isn't fixed forever. Recheck it after every mock, since it shifts as your preparation improves. A ceiling calculated three months out will likely be too conservative closer to CAT, once your accuracy on previously borderline question types has genuinely improved.
Pull up your last mock and split your attempts into "confident" and "unsure." Compare the accuracy of each group. If the "unsure" group's accuracy sits below your marking scheme's breakeven point, that's exactly where your next ceiling adjustment should happen.
How VARC, DILR, and QA differ
The VALUE method applies across all three sections, but what it looks like in practice differs, because the question formats aren't the same.
| Section | What attempt strategy mainly depends on |
|---|---|
| VARC | Reading speed and RC-level accuracy — attempt count is mostly a function of how many passages you complete in time |
| DILR | Set selection, not per-question count — cracking a set's logic often unlocks most or all of its questions together |
| QA | A mix of MCQ (needs the accuracy-threshold math) and TITA (usually worth attempting whenever you have an approach) |
An aspirant sees 22 QA questions with 34 minutes on the clock, roughly 15 of them MCQs, the rest TITA. Their verified accuracy from recent mocks: 70% on questions they immediately recognize the method for, 30% on questions requiring a guess between two plausible approaches.
V — Verify: 70% and 30% are their real, mock-derived numbers, not hopeful estimates.
A — Apply the math: At 70%, expected value per MCQ is strongly positive; at 30%, it's still slightly positive under a +3/−1 scheme, but close enough to breakeven that time cost starts to matter more than the marks.
L — Lock a ceiling: All "recognized method" questions get attempted first. "Requires a guess" MCQs are attempted only if time remains after the confident pass. All TITA questions with any workable approach are attempted regardless, since there's no penalty downside.
U — Use buffer time: With four minutes left and every confident question answered, the remaining time goes to re-checking two calculation-heavy answers rather than opening a new "requires a guess" MCQ.
E — Evaluate: After the mock, they check whether the "requires a guess" MCQs they did attempt paid off, and adjust next time's ceiling accordingly.
How we built this guide
The VALUE method applies standard expected-value reasoning to CAT's typical marking scheme to explain why attempt count alone is the wrong metric. The specific marking-scheme numbers used for illustration (+3 correct, −1 wrong on MCQs, no penalty on TITA) reflect a commonly cited CAT scheme; always confirm the current year's exact scheme from the official notification, since the underlying math changes if the scheme does.
Attempt strategy only matters if the underlying accuracy is real; our CAT error log guide covers how to actually raise it. Once you're consistently attempting near your ceiling, the number that matters shifts to percentile; our score vs percentile guide covers how that conversion actually works.
The CAT exam hub collects every section-wise and strategy guide in one place, and the CAT score predictor shows how a higher-accuracy attempt strategy moves your projected percentile.
Key takeaways
- There's no universal attempt number for 99+ percentile — it depends on your own accuracy, not just a count.
- Use the VALUE method: Verify your accuracy, Apply expected-value math, Lock a ceiling, Use buffer time on sure answers, and Evaluate and adjust after every mock.
- MCQs need an accuracy-threshold check; TITA questions are usually worth attempting whenever you have any approach.
- Attempt strategy differs by section, since VARC, DILR, and QA don't share the same question format or logic.
- Your ceiling isn't fixed — recalculate it as your accuracy genuinely improves.
Stop guessing your attempt ceiling
Bring your recent mock data to a free session. We'll calculate your real accuracy-by-confidence breakdown, section by section.
Get Your Free CAT 2026 Strategy Session →Questions aspirants ask about attempt strategy
Build your CAT 2026 study plan
Personalised daily plan that adapts to your section-wise mock scores.