Every MECEE-BL question is worth +1 for a correct answer and −0.25 for a wrong one; a blank scores zero. Most candidates take this as a warning — "do not guess" — and leave marks on the table. The penalty is not there to stop you guessing. It is there to make blind guessing worthless while leaving informed guessing profitable. The difference is arithmetic, and it is worth doing once.
The break-even
With four options and no idea, your chance of being right is 25%. Expected score from a blind guess:
0.25 × (+1) + 0.75 × (−0.25) = 0.25 − 0.1875 = +0.0625So even a completely blind guess has a slightly positive expectation — about a sixteenth of a mark. That is the design: −0.25 is exactly the point where a four-option blind guess is (barely) worth it, so the penalty punishes recklessness without making it fatal.
Now eliminate one option you are sure is wrong. Three remain:
(1/3) × (+1) + (2/3) × (−0.25) = 0.333 − 0.167 = +0.17Eliminate two:
(1/2) × (+1) + (1/2) × (−0.25) = 0.5 − 0.125 = +0.375Each option you can rule out roughly doubles the value of answering. A question you have narrowed to two options is worth answering every single time.
The rule that follows
- Can you eliminate at least one option with confidence? Answer.
- Genuinely no idea, all four look plausible? The expectation is still marginally positive, but the variance is high and your time is better spent elsewhere. Leave it, flag it, come back if there is time.
- Never answer on a hunch that one option "sounds right" if you cannot say why the others are wrong. That is where wrong answers cluster.
Why people still lose marks
The penalty hurts not because of guessing but because of confident errors — questions you did not think were guesses at all. In practice this is where most of the −0.25s come from: a misread stem ("which is NOT…"), a unit slip in Physics, a Chemistry exception you half remembered. The fix is not caution; it is reviewing every wrong answer until the pattern behind it is obvious.
Training it in mocks
Our mock player scores exactly as the real paper does, and the results page separates wrong from unanswered so you can see your two numbers over time:
- Accuracy on attempted questions — should climb toward 85–90% with practice.
- Attempt rate — how much of the paper you answered. Strong candidates attempt well over 90%.
If your accuracy is high but your attempt rate is low, you are leaving positive-expectation guesses blank: answer more. If your accuracy is low, you are answering on hunches: eliminate first. The analytics page tracks both across every paper you sit.
One more thing the arithmetic says: at 200 questions, twenty extra informed guesses at +0.375 each is 7–8 marks — several hundred ranks in the middle of the merit list. That is the most reliable score gain available in the last month before the exam, and it costs nothing to learn.