Attempt Strategy and Cut-offs

A typical selection applies two kinds of qualifying mark:

• a sectional cut-off you must reach in each test, and
• an overall cut-off on the total.

Both bind. A strong total does not rescue a section that fell below its own line, and clearing every section does not help if the total is short. The exact figures are set after the exam from the candidate pool and are not announced in advance.

The practical consequence: you cannot treat a weak section as something the others will cover. Each section needs enough safe attempts to clear its own line first; only then does maximising the total matter.

Once accuracy in a section falls much below the mid-eighties, each additional attempt adds very little net score and a lot of variance. The reason is the double loss on a wrong answer: you spend time you could have used on a question you knew, and you pay the penalty.

So the aim is not "attempt as many as possible". It is: attempt as many as you can while holding accuracy around 85 percent or above. The number of questions that allows depends on the paper's difficulty that day and on your preparation - it is not a fixed target you carry in.

For one-mark questions with a one-fourth penalty, if you make A attempts at accuracy p (a fraction between 0 and 1), your expected net score is roughly

net = A x ( p - 0.25 x (1 - p) )

The bracket is the net marks per attempt. At p = 0.85 it is 0.85 - 0.25 x 0.15 = 0.8125. At p = 0.70 it is 0.70 - 0.25 x 0.30 = 0.625.

Use it two ways: estimate the net a plan of attempts will yield, or fix a target net (say the likely sectional cut-off plus a margin) and back out how many accurate attempts you need. If that number is out of reach at your accuracy, the fix is accuracy, not more attempts.

One-mark questions, penalty one-fourth throughout.

Net score from a section. A candidate makes 68 attempts: 58 right, 10 wrong. Net = 58 minus 10 times 0.25 = 58 minus 2.5 = 55.5. Another, steadier: 45 right, 5 wrong gives 45 minus 1.25 = 43.75.

Expected value of one guess on a five-option question, by how many options you have ruled out. With n options left, the value is one chance in n of plus 1 and n minus 1 chances in n of minus 0.25:

• Rule out none (5 left): (1/5)(+1) + (4/5)(-0.25) = 0.
• Rule out 1 (4 left): (1/4)(+1) + (3/4)(-0.25) = +0.0625.
• Rule out 2 (3 left): (1/3)(+1) + (2/3)(-0.25) = about +0.167.
• Rule out 3 (2 left): (1/2)(+1) + (1/2)(-0.25) = +0.375.

So a guess is worth taking once two options are gone, and clearly so once three are.

Same attempts, different accuracy. 40 attempts at 85 percent accuracy nets 40 x (0.85 - 0.25 x 0.15) = 40 x 0.8125 = 32.5. The same 40 attempts at 70 percent net 40 x (0.70 - 0.25 x 0.30) = 40 x 0.625 = 25. Seven and a half marks, from accuracy alone.

Signs you have reached your ceiling in a section for the day:

• your last few attempts were two-option guesses, not solved questions;
• your running accuracy is sliding and each new attempt is harder than the last;
• everything left is a long set you have already judged not worth the time.

At that point, on a single overall clock, the marginal minute is better spent banking easy marks in another section. Under fixed sectional timing, you cannot move the time, so spend the remainder rechecking flagged answers you actually reasoned out - not blind-filling.

Stopping is a decision, not a failure. A section cleared at its cut-off with room to spare has done its job.

• Two cut-offs bind: each section's own qualifying mark and the overall total. A strong total will not save a failed section.
• Aim to attempt as many as you can while holding accuracy around 85 percent or higher; the number that allows varies with the paper.
• Net-score estimate: net = attempts x ( p - 0.25 (1 - p) ). The bracket is net marks per attempt: 0.8125 at 85 percent, 0.625 at 70 percent.
• Expected value of one guess: 0 with five options, +0.0625 after eliminating one, about +0.167 after two, +0.375 after three. Guess once two are gone.
• Same 40 attempts net 32.5 at 85 percent accuracy versus 25 at 70 percent.
• Stop a section when recent attempts are just two-option guesses; move the time to another section on a single clock, or recheck reasoned answers under sectional timing.

One-mark questions, one-fourth penalty. Give the answer and a one-line reason.

1. In reasoning you attempt 52: 47 right, 5 wrong. Net score?
2. Five options, and you can rule out two. Is a guess worth taking?
3. Your total clears last cycle's overall mark comfortably, but your English net is one mark below the English sectional cut-off. Do you go through?
4. A section run at 90 percent accuracy: net marks per attempt?
5. Same 40 attempts: candidate A at 80 percent accuracy, candidate B at 65 percent. Net score for each?
6. This section is separately timed with two minutes left; your last four answers were two-option guesses. Best use of the time?
7. Is "attempt exactly 60 in quantitative aptitude every time" a sound rule?

Answers. 1 47 minus 5 times 0.25 = 45.75. 2 Yes - three options left gives an expected value of about +0.17 per guess, clearly positive. 3 No - every section must independently clear its own cut-off; the total cannot compensate. 4 0.90 minus 0.25 times 0.10 = 0.875 marks per attempt. 5 A: 40 x (0.80 - 0.25 x 0.20) = 30. B: 40 x (0.65 - 0.25 x 0.35) = 22.5. 6 Recheck flagged answers you actually reasoned out - you are at your accuracy ceiling and the time cannot move to another section. 7 No - the attempt count should flex with the day's difficulty and your running accuracy against the cut-off, not be a fixed number.