Diagnose the Leak Before You Review Everything
Do not begin UPCAT Math or Science review by rereading every notebook. Take a controlled diagnostic, find the decisions costing you points, and repair them with fresh problems. ‘62% in Math’ is vague; ‘I solve equations but set up rate problems incorrectly’ tells you what to do tomorrow.
This guide gives you that system for UPCAT 2027, scheduled for August 1–2, 2026. It was checked on July 18, 2026 against the current UP Office of Admissions General Information Bulletin, UPCAT admissions homepage, and UP's central UPCAT page. These pages confirm Science and Mathematics as two of four subtests. The bulletin gives about five hours for the full administration and prohibits calculating devices.
Those pages do not publish a detailed Math or Science blueprint, fixed item counts, separate subtest timing, or topic weights. Every detailed topic list below is a prep framework, not an official forecast. Exact percentages or item counts on reviewer sites are third-party estimates.
For admissions and logistics, use the complete UPCAT guide. For all four subtests, see the free practice-test guide. The verbal half has its own Language and Reading review. This article owns Math/Science diagnosis, repair, and retesting.
Keep Official Facts and Prep Choices Separate
| Question | What UP currently confirms | What you decide in practice |
|---|---|---|
| Which subtests exist? | Four, including Science and Mathematics. | Keep separate Math and Science results. |
| When is UPCAT 2027? | August 1–2, 2026. | Count backward from your assigned date. |
| How long is each? | UP gives about five hours for the full administration, not separate public limits. | Build pace from your timed sets. |
| How many questions and what weights? | No fixed per-subtest counts or detailed topic weights are public. | Sample broad skill families, then follow your misses. |
| May you use a calculator? | The current bulletin prohibits calculating devices. | Practice written computation and estimation. |
This boundary keeps unsupported precision from steering your study. What matters is which school-level relationships you can retrieve, represent, and apply without help.
Run a Two-Layer Math and Science Diagnostic
- Untimed reasoning layer. Work without notes or a calculator. Write the setup or evidence first and mark confidence high, medium, or low.
- Recorded-time transfer layer. The next day, use a fresh comparable set. Record total time and stalls. Do not borrow an unofficial subtest countdown.
Review every miss, omission, and low-confidence correct answer; a lucky answer is not reliable yet. Track skill family, confidence, time, failure mode, corrective rule or evidence, and recheck date separately for Math and Science.
Use these broad prep-only sampling buckets:
- Math: number/proportional reasoning; algebraic representation; geometry/measurement; tables, graphs, statistics, probability; multi-step word problems.
- Science: observations/data; life systems; matter and chemical relationships; force, motion, energy, waves; Earth/environmental systems; experimental design.
These are diagnostic labels, not UP's blueprint. Use different names if your school notes make the weakness easier to retrieve.
Use Seven Failure Modes, Not ‘Careless’
Tag each miss with one primary cause.
| Failure mode | What it looks like | Repair |
|---|---|---|
| Concept | You cannot explain the relationship. | Retrieve it plainly, study one example, solve two contrasts. |
| Setup | You translate the situation into the wrong model. | Write known, unknown, and governing relationship first. |
| Representation/data | You misread an axis, table, diagram, scale, or trend. | State labels and units before interpreting. |
| Arithmetic | The model is right but computation fails. | Rework cleanly and estimate size or sign. |
| Unit | You combine incompatible units. | Put units beside numbers and cancel them line by line. |
| Evidence | A science answer is plausible but unsupported. | Cite the deciding datum; reject stronger claims. |
| Pacing | You stall, restart, or rush. | Practice a proceed, mark, or return checkpoint. |
Never log only ‘careless.’ Changing 0.4 kilometers to 40 meters is a unit error. Using the final price as the base for percent change is setup. Picking the tallest bar without checking an axis that starts at 95 is representation.
Three Math Mini-Diagnostics
These are illustrative practice examples, not official or recalled UPCAT items. Narrate the setup before solving.
1. Successive Percent Changes
A value of 800 rises 20%, then falls 20%: 800 × 1.20 × 0.80 = 768, not 800. The second percentage uses 960 as its base. ‘Equal changes cancel’ signals a concept error; subtracting 160 twice signals setup.
2. Build the Equation First
A tricycle fare has a fixed 15-peso charge plus 4 pesos per kilometer. A 47-peso trip gives 15 + 4k = 47, so k = 8 kilometers. Writing 4(k + 15) groups the fixed fee incorrectly. Repair by naming what repeats per kilometer and what occurs once.
3. Let Units Decide the Geometry
A garden is 3 meters by 250 centimeters. Since 250 centimeters = 2.5 meters, its area is 7.5 square meters. Reporting 750 square meters is a unit failure, not a geometry failure. Estimating the expected range makes the error visible.
Three Science Mini-Diagnostics
These are also illustrative, not official UPCAT questions. Target the reasoning move rather than another definition.
1. Separate Observation From Explanation
Compress a sealed syringe of gas at constant temperature and pressure rises. The observation is lower volume, higher pressure; a particle explanation is more frequent wall collisions in less space. ‘Particles became larger’ shows a concept gap. Reversing graph axes shows a representation error.
2. Protect the Experiment's Comparison
Two identical plants receive the same light, water, soil, and pot size; only one receives fertilizer. Fertilizer is manipulated, height is measured, and the unfertilized plant supplies comparison. Greater height supports an association under these conditions, not ‘fertilizer always helps every plant.’ That overreach is an evidence error.
3. Read Slope Before Reaching for a Formula
A distance-time graph rises from 0 meters at 0 seconds to 30 meters at 6 seconds. Its slope is 5 meters per second, the average speed. Dividing 6 by 30 reverses the relationship (setup); reporting 5 meters loses physical meaning (unit).
Turn One Miss Into a Repair Loop
For every miss, record the exact task, original reasoning, failure mode, deciding rule/evidence, attractive wrong path, repair task, and delayed recheck date.
After the percent miss, do not repeat the identical numbers. Explain why the bases differ, solve a fresh opposite-change problem, recheck two days later, then mix it among unrelated questions. For Science, identify variables in fresh experiments or interpret unfamiliar axes before describing a trend. Changed surface details reveal whether the reasoning transfers.
A 14-Day Math and Science Repair Plan
This is a prep recommendation, not an official weighting or timing plan.
| Day | Main job | Evidence of completion |
|---|---|---|
| 1 | Untimed Math/Science diagnostic | Tag every miss and lucky guess. |
| 2 | Fresh recorded-time diagnostic | Separate knowledge from clock failures. |
| 3 | Largest Math concept/setup family | Solve two fresh examples with setup shown. |
| 4 | Largest Science concept/evidence family | Cite the deciding observation. |
| 5 | Math units, estimation, arithmetic | Show units and expected range. |
| 6 | Science data and experimental design | Label axes, variables, controls, conclusions. |
| 7 | Mixed retrieval and log count | Rank failure modes by frequency. |
| 8 | Delayed recheck of Days 3–4 | Solve new examples without notes. |
| 9 | Second Math family | Improve without copying old solutions. |
| 10 | Second Science family | Connect evidence to claim. |
| 11 | Timed mixed set | Log stalls and rushed endings separately. |
| 12 | Pacing triage and short drills | Use a repeatable proceed/mark/return rule. |
| 13 | Final fresh mixed set | Compare failure counts with Days 1–2. |
| 14 | Light consolidation | Review recurring rules, units, and graphs. |
Progress means repeated error families shrink. If one stays unstable, rebuild its prerequisite instead of adding question volume.
Judge a Reviewer Without Believing Its Blueprint
A third-party reviewer can help without being authoritative. Check whether it labels original practice clearly, explains answers, avoids presenting fixed counts as UP facts, varies words/equations/tables/graphs/units, and lets you transfer a solution to a fresh example. Avoid leaked or purported live items.
Your Next 45 Minutes
Choose a small mixed Math/Science set. Record confidence and write the setup or deciding evidence before answering. Tag misses and lucky guesses, repair the most frequent failure with two fresh contrasts, and schedule a no-notes recheck in two days.
Buy on Amazon