A problem tells you there are 18 objects in all. Some are worth 3 points and the rest are worth 5 points. The total score is 72.
How many 5-point objects are there?
A learner may say, “I’ll guess.” That can sound mathematically weak.
It is weak only when the guess is random.
Controlled guess-and-check is a search strategy: choose a sensible trial, calculate its consequence, compare the result with the target, then use the size and direction of the error to choose the next trial.
The intelligence is not in guessing. It is in how each failed attempt reduces uncertainty.
The quick answer: a good guess creates information
For the 18-object problem, let x be the number of 5-point objects. Then 18 − x are worth 3 points.
Suppose we try x = 6.
Score = 6×5 + 12×3 = 30 + 36 = 66.
The target is 72. We are 6 points too low.
Changing one 3-point object into one 5-point object increases the total by 2 points.
We need 6 more points, so increase the number of 5-point objects by 3.
Try x = 9.
Score = 9×5 + 9×3 = 45 + 27 = 72.
Answer: 9 objects are worth 5 points.
The second trial was not another guess in the everyday sense. It was calculated from the first error.
Random guessing and mathematical search are different activities
Random guessing looks like this:
- try 4;
- try 12;
- try 7;
- try 15;
- hope one works.
Controlled guess-and-check looks different:
- define the possible range;
- choose a useful first trial;
- calculate the result accurately;
- compare it with the target;
- identify whether the trial is too high or too low;
- estimate how much one step changes the result;
- move deliberately;
- verify the final solution against every condition.
A wrong trial is useful if it tells you where not to search next.
Use a table so the search becomes visible
For a two-quantity problem, record each trial instead of carrying it mentally.
| 5-point objects | 3-point objects | Total score | Compared with 72 |
|---|---|---|---|
| 6 | 12 | 66 | 6 low |
| 9 | 9 | 72 | exact |
The table reduces working-memory load and exposes the relationship between the trial and the outcome.
The most useful observation is often the change per step
In the example above, replacing one 3-point object with one 5-point object changes the total by +2 while keeping the number of objects fixed.
That fixed change creates a fast correction rule.
If one trial step changes the outcome predictably, the error itself tells you how many steps to move.
This is why many “guess-and-check” problems quietly contain linear structure.
Worked example: two ticket prices
A school sells 40 tickets. Adult tickets cost $8 and child tickets cost $5. Total sales are $260. How many adult tickets were sold?
Start with a simple trial: 20 adult and 20 child tickets.
Revenue = 20×8 + 20×5 = 160 + 100 = $260.
The first trial happens to work, so 20 adult tickets were sold.
Now change the total revenue to $278.
Our 20-and-20 trial gives $260, which is $18 too low.
Changing one child ticket to one adult ticket adds $3 while preserving 40 tickets.
$18 ÷ $3 = 6.
Increase adult tickets from 20 to 26.
Then child tickets = 14.
Check: 26×8 + 14×5 = 208 + 70 = $278.
Notice how the method became almost algebraic once the per-step effect was recognised.
Worked example: when the first trial should be a boundary
A puzzle asks for a whole number between 1 and 50 that satisfies several conditions.
If increasing the number always increases the computed result, a middle trial such as 25 may be efficient.
If 25 gives a result that is too high, values above 25 can be eliminated. If the next trial 12 is too low, the solution lies between 12 and 25.
This begins to resemble interval narrowing or binary search.
The general lesson is useful beyond Primary Mathematics:
choose trials that remove large parts of the search space.
When guess-and-check is a strong choice
- The answer must be a whole number from a small or bounded range.
- There are several conditions but no obvious direct formula.
- A trial can be checked quickly.
- Changing the trial has a predictable effect on the result.
- The problem involves counts of two types with a fixed total.
- The learner can use a table to record outcomes systematically.
When another method is better
Do not use guess-and-check merely because a problem is difficult.
- If a direct unitary method exists, use it.
- If an invariant gives the relationship immediately, use the invariant.
- If there are many independent cases, systematic listing may be safer.
- If the relation is algebraically clear, an equation may be shorter.
- If the search space is huge and unstructured, trial-and-error becomes inefficient.
Good problem solving includes method selection.
A surprising bridge to algebra
The 40-ticket problem can be written directly as:
8x + 5(40 − x) = 278.
Simplify:
8x + 200 − 5x = 278.
3x = 78.
x = 26.
The controlled guess-and-check reasoning discovered the same structure informally: each replacement changes the total by 3.
This makes guess-and-check a useful bridge rather than an intellectual dead end.
Common misconception 1: the first guess should be lucky
The first trial does not need to be close.
It needs to be informative.
Repair: choose a midpoint, an easy boundary or a balanced case that makes the arithmetic transparent.
Common misconception 2: every wrong trial means start over
A wrong trial contains directional information.
Repair: write “too high” or “too low” and calculate the size of the error.
Common misconception 3: changing one quantity does not affect the other
If the total number of objects is fixed, increasing one type decreases the other.
Repair: preserve every stated constraint in each trial.
Common misconception 4: once the target is reached, checking is unnecessary
A trial may satisfy one condition but violate another.
Repair: verify the final answer against total count, value, units and any inequality or whole-number requirement.
A diagnostic ladder
- Can the learner state the possible answer range?
- Can the learner preserve all fixed conditions in a trial?
- Can the learner calculate the consequence of a trial accurately?
- Can the learner identify whether the result is high or low?
- Can the learner quantify the error?
- Can the learner identify the effect of changing the trial by one step?
- Can the learner choose the next trial from that information?
- Can the learner record attempts in a table?
- Can the learner recognise when another method would be more efficient?
- Can the learner verify the final answer independently?
How this fits Singapore Primary and PSLE Mathematics
The current MOE Primary Mathematics syllabus places mathematical problem solving at the centre of learning and organises content across Number and Algebra, Measurement and Geometry, and Statistics. “Guess-and-check” is best understood as a problem-solving heuristic rather than a standalone examinable content strand.
For the 2026 PSLE Mathematics examination, SEAB states that pupils are assessed not only on recall and straightforward procedures, but also on interpreting information, applying concepts in varied contexts, reasoning mathematically, analysing information and selecting appropriate strategies. Controlled guess-and-check belongs naturally inside that strategic repertoire when the problem structure supports it.
What parents should listen for
- “My first trial is inside the possible range.”
- “This result is 12 too high, so I should move downward.”
- “Every time I change one object type, the total changes by 2.”
- “I am recording the trials so I do not repeat them.”
- “This method works, but an equation would be faster now that I can see the relationship.”
What teachers and tutors should protect
- Do not praise lucky guessing as strategy.
- Require each trial to generate a reason for the next one.
- Use tables when several trials are needed.
- Ask what remains fixed when one trial variable changes.
- Compare the heuristic with model, unitary and algebraic solutions after the learner understands it.
- Include at least one problem where guess-and-check is possible but not the best method, so method selection itself is trained.
The deeper lesson: failed attempts can be evidence
Mathematics is often presented as if strong students move directly from question to correct method.
Real problem solving is sometimes messier.
A trial can fail and still improve the solution.
The difference between random trial-and-error and mathematical search is that the second attempt knows something the first attempt did not.
A controlled guess is not a surrender of reasoning. It is an experiment designed to make the next decision better.