Test Prep
Negative Marking Explained: When to Guess on CEnT-S and TR-YÖS
There is a persistent myth that floats around international student forums: "If an exam has negative marking, never guess." It sounds cautious and reasonable. It is also wrong, and believing it will cost you points.
This article unpacks the actual mathematics behind the penalty systems used by two exams that matter enormously to internationally mobile students right now: CEnT-S, Italy's new English-taught admissions test run by CISIA, and TR-YÖS, the international student exam run by Turkey's ÖSYM. Once you see what the numbers actually say, you will replace vague anxiety about guessing with a clear, repeatable decision rule you can apply under time pressure in the exam room.
What CEnT-S Is and How Its Scoring Works
CEnT-S came into use in November 2025, replacing the earlier English TOLC-I, TOLC-E, and TOLC-F tests. It is CISIA's entrance examination for students applying to English-taught degree programmes in Italy, and it has quickly become a key hurdle for international applicants aiming at Italian universities.
The test contains 55 questions spread across five sections, with a total time allowance of 110 minutes. One important structural feature: once you move on from a section, you cannot go back to it. That changes your strategy in ways we will return to.
The raw scoring is straightforward:
- A correct answer: +1 point
- A blank: 0 points
- A wrong answer: -0.25 points
Across all 55 questions, your raw score therefore ranges from a theoretical floor of -13.75 (everything wrong) to a ceiling of 55 (everything right).
There is an additional layer worth knowing. After each registration period closes, CISIA applies a normalisation coefficient to each section. The coefficient is calculated from how easy each question turned out to be across all candidates who sat that version of the test. Because this depends on the performance of the entire cohort, you cannot know your normalised score in advance, and the final score can actually exceed 55. For full details on how normalisation is calculated, see our CEnT-S scoring guide.
What TR-YÖS Is and How Its Scoring Works
TR-YÖS is administered by ÖSYM, the same Turkish body that oversees domestic university entrance, for international students seeking places at Turkish universities. The exam contains 80 questions to be completed in 100 minutes, split across two separate tests. Crucially, the quarter-mark deduction for wrong answers is calculated independently for each of those two tests, not across the exam as a whole. Like CEnT-S, TR-YÖS uses five options per question.
For more on the structure and registration process, see our TR-YÖS overview.
The Myth of the Dangerous Guess
Both exams deduct 0.25 marks per wrong answer. Both use five-option questions. This is not a coincidence: the -0.25 penalty for a five-option question is specifically designed so that blind, random guessing has an expected value of exactly zero.
Here is the arithmetic. If you guess randomly on a five-option question, you have a 1-in-5 chance of being right and a 4-in-5 chance of being wrong. Your expected score is:
(1/5) × 1 + (4/5) × (-0.25) = 0.20 - 0.20 = 0.00
The penalty was engineered to neutralise random guessing, not to punish it. A blind guess on a five-option question is worth the same as leaving it blank: nothing. The exam designers are not trying to trick you into losing points by guessing; they are trying to ensure that guessing without any knowledge gives no advantage.
The danger is not in guessing. The danger is in the mistaken belief that guessing is dangerous, which leads students to leave questions blank that they could have answered profitably.
The Elimination Rule: When a Guess Becomes Worth Taking
The situation changes the moment you can eliminate even one option. Every option you rule out shifts the expected value of a guess from zero into positive territory.
The table below shows the expected value of guessing once you have eliminated some options. It is calculated as: p - (1-p)/4, where p is your probability of picking the right answer from the remaining options.
| Remaining options | Expected value of one guess | |---|---| | 5 (no elimination) | 0.00 | | 4 (one eliminated) | +0.06 | | 3 (two eliminated) | +0.17 | | 2 (three eliminated) | +0.38 | | 1 (four eliminated — you know the answer) | +1.00 |
Even eliminating a single option makes guessing mathematically correct. At two options remaining, a guess is worth more than a third of a mark on average. By the time you are down to two plausible answers and cannot decide, the expected cost of guessing wrong (-0.25) is far outweighed by the expected gain of guessing right (+0.75, weighted by probability).
The decision rule is therefore simple: if you can eliminate at least one option with any confidence, answer the question. Do not leave it blank.
The Cannot-Go-Back Constraint on CEnT-S
CEnT-S adds a layer of strategic complexity that TR-YÖS does not. Because sections cannot be revisited once you advance, you cannot adopt the strategy of flagging uncertain questions and returning to them with fresh eyes later in the exam. What you see in a section is your only chance.
This matters for the elimination rule in two ways.
First, it raises the importance of working deliberately through each section rather than rushing to the end and hoping to circle back. You will not circle back.
Second, it means that time management within each section is more consequential than on a free-navigation exam. If you spend too long on a hard question early in a section and then run out of time at the end, you cannot recover those blanks. A reasonable approach is to move through questions at a consistent pace, making a quick elimination judgment on each one: can I rule out anything? If yes, answer and move on. If you genuinely see no path to eliminating any option and the question is consuming time, a blank is equivalent to a random guess — but that equivalence is rarer than most students assume.
How This Differs From the SAT
The contrast with the SAT is worth spelling out clearly, because many international students prepare for multiple exams simultaneously and it is easy to cross-contaminate strategies.
The SAT has no wrong-answer penalty at all. Every blank is a wasted opportunity, because a random guess on the SAT has a positive expected value: (1/4) × 1 = +0.25 per question (the SAT uses four options). On the SAT, you should always answer every question, even if you are completely guessing. Leaving anything blank is simply giving up points.
For more on SAT strategy, see our SAT guide.
Bringing SAT logic into a negatively marked exam leads to overconfidence in guessing. Bringing negatively marked exam caution into the SAT leads to costly blanks. Know which exam you are sitting, and apply its specific logic.
Practical Application: A Question-by-Question Protocol
Here is how to operationalise the elimination rule under real exam conditions.
Step one: Read the question and all five options. Do not read the question and then guess before checking the options. The options themselves often help you locate what the question is really testing.
Step two: Cross off anything obviously wrong. This can be fast. You do not need to prove an option is wrong beyond doubt; you need a reasonable basis for doubting it. That is enough to shift the expected value into positive territory.
Step three: Apply the rule. One option crossed off? Answer. Two crossed off? Definitely answer. Cannot cross off anything and genuinely have no feel for the question? Leave it blank. This situation is less common than students fear, especially after adequate preparation.
Step four on CEnT-S specifically: Do not agonise past your section time budget. An uncertain answer is worth more than a time debt that leaves questions unanswered. Set a rough per-question pace (110 minutes for 55 questions is exactly 2 minutes per question on average) and treat significant overruns as a signal to commit and move on, not a reason to think harder.
The Psychology of Negative Marking
There is a cognitive dimension here that exam strategy guides rarely address directly. The -0.25 penalty feels asymmetric even when it mathematically is not, because losses feel more salient than equivalent gains. This is well-established in decision research: the fear of seeing your score drop tends to loom larger than the rational expectation of a small positive return.
The result is that most students who have not thought through the mathematics over-skip, and over-skipping is a systematic source of lost marks. Understanding why the anxiety arises — and that it does not reflect the actual expected value calculation — is part of what lets you override it in the exam room.
Build the elimination habit in practice tests. Go through your practice papers and, for every blank you left, ask: could I have eliminated at least one option? If the honest answer is yes, you made the wrong call. Track this. Most students find they left a meaningful number of profitable guesses on the table.
Normalisation on CEnT-S: What It Does and Does Not Change
A question that comes up frequently: does the normalisation applied by CISIA after the exam change the guessing calculus?
The answer is no, for a straightforward reason. Normalisation is applied after the fact, uniformly across sections, based on aggregate difficulty. It does not make incorrect answers more or less costly relative to blanks within the raw scoring. The +1 / 0 / -0.25 structure governs how your behaviour in the exam translates into a raw score, and normalisation then scales that raw score. The relative advantage of an expected-value-positive guess over a blank is preserved through the scaling. Your strategy in the exam room should be based entirely on the raw scoring rules.
What normalisation does affect is the comparison of scores between different test sittings, which is its purpose: it makes scores from a harder sitting comparable to scores from an easier sitting. That is relevant when you are evaluating your result, but not when you are deciding whether to answer a particular question.
Putting It Together
Two exams, one underlying principle. The negative marking systems on CEnT-S and TR-YÖS are not traps. They are deliberately calibrated so that blind guessing is a zero-sum activity, and so that any real knowledge, even partial knowledge that lets you eliminate one option, has genuine positive expected value. The correct rule is not "never guess." The correct rule is "answer whenever you can eliminate at least one option, and leave a blank only when you genuinely cannot."
On CEnT-S, layer in the no-revisiting constraint: manage your time within each section, make your elimination judgment once per question, and commit. On TR-YÖS, remember that the penalty is calculated separately for each of the two component tests. On the SAT, forget all of the above and answer everything.
The students who do best on negatively marked exams are not the most cautious. They are the most calibrated: they have practised the elimination habit until it is fast and automatic, and they have made peace with the mathematics well before they walk into the room.
If you are preparing for CEnT-S, TR-YÖS, or any other international entrance exam and want to build a structured study plan around your specific timeline and target programmes, our advisors at AxoPrep are available for a free consultation. Talk to an Atlas advisor about your profile, your target institutions, and how to spend your preparation time most efficiently.
