SKILL ALIGN · YEAR 12 · 2027
SACE Stage 2 General Mathematics 2027 — Practice Pack 0 Free Sample
Questions, worked solutions and marking guidance
10 additional independent questions introduce all 10 paid-course topic chapters, with worked support and separate solutions. This representative overview shows course breadth, not full-resource depth or complete syllabus mastery.
Version 1.3 · 10 questions · 36 practice marks · Booklet build: 20261114-SGM-R14
Independent free sample. These 10 questions and their worked examples are additional to the individual and School resources.
Independent Skill Align practice resource. Curriculum reference: SACE General Mathematics Stage 2 Subject Outline, reissued 2025; changes from 2025 onwards; Topics 1–5 (Matrices option) (checked 26 October 2026). This resource is not endorsed by SACE Board.
How to use this resource
- This independent overview introduces every paid-course topic chapter at representative depth. Read each chapter's prerequisites and worked support, then attempt its questions before opening Solutions.
- Use the material and performance conditions stated in each question. Record topics you have not studied separately; this sample has no course checks or mixed revision.
- Use the worked solutions below to award one mark per criterion. Accept equivalent mathematically correct working. Carry an earlier numerical error forward where the later method is correct; do not award a mark twice for the same step.
- Record your reasoning, independence and next step for each topic on the review sheet. Use a full resource for further practice, course checks and mixed revision.
Untimed SACE Stage 2 learning practice for compulsory Topics 1, 3, 4 and 5 plus the Topic 2 Matrices option. A school's alternative Open Topic is not included. The external examination covers only Topics 3–5; these whole-course workshops also teach the selected school-assessed topics. Use an approved non-CAS graphics calculator and show the model, working and interpretation. A spreadsheet may support a stated investigation; the official school-investigation software conditions are distinct from examination rules. No General Mathematics formula sheet is supplied. The examination permits one unfolded A4 handwritten sheet, both sides, and the approved calculator combinations. Unless a task states otherwise, interest is applied before an end-period deposit/payment, periods match, and calculations retain full precision until the requested final rounding. Rates, tax rules, fees and operational conditions are supplied fictional learning assumptions. Row-state transition models use R(next)=RT. Different checks are not equated progress measures or grade predictions. Free Individual Pack 0 gives one independent question for each of the ten substantive paid-course chapters. This representative introduction does not replace full workshop depth, complete syllabus study or examination preparation.
SACE approved calculator conditions
Suggested times guide study planning. Questions are grouped for learning; total marks and demand labels do not predict an external examination grade, ATAR or a validated diagnostic score.
Use the working space provided or extra paper labelled with the question reference. Keep your original attempt so you can compare its reasoning with the worked solution below each question.
LIN · Linear relationships and independent information
Outline 1.1: contextual, numerical, graphical and algebraic linear relationships; paired equations; unique, dependent and inconsistent information. Representative topic preview; full resources provide broader applications and depth.
Before you start: Stage 1 linear functions, units, substitution and rearrangement.
Suggested study time: 25 minutes, including review.
The final question in this chapter is your self-check. Attempt it independently before opening its solution.
LIN-preview 3 marks · Technology-active · Model, apply, interpret and justify
Three linked calibrations report b=2a+3, c=3b−4 and then a′=(c−2)/6. Compose one complete round trip. A technician says only an unfortunate starting value causes disagreement. Audit that claim and find the change after ten round trips. Replace the final constant 2 by k so the round trip returns every starting value exactly.
Worked solution — LIN-preview
- The first two rules give c=6a+5. Thus a′=(6a+5−2)/6=a+1/2, so every starting value gains 1/2 per trip.
- After ten trips the gain is 5; no choice of starting value removes the systematic offset.
- With return rule (c−k)/6 the composite is a+(5−k)/6, which is the identity exactly when k=5.
Marking guidance — one mark per criterion — LIN-preview
- Composes the rules and finds the universal offset 1/2
- Explains the starting-value claim and ten-trip drift 5
- Derives k=5 and verifies identity for every input
Acceptable alternatives — LIN-preview
- Credit repeated substitution or a labelled input/output table supporting the general affine result; ten trials alone do not prove the every-input identity.
Common error: LIN-preview — Checking one convenient value and overlooking repeated offset accumulation.
Next step: LIN-preview — Compose the entire round trip before testing whether it is an inverse.
LP · Linear programming and feasible decisions
Outline 1.2: constraints and objective, feasible region, vertices, continuous/integer solutions, slack and sensitivity. Representative topic preview; full resources provide broader applications and depth.
Before you start: LIN, inequalities and non-negative variables; distinguish a feasible point from an optimum.
Suggested study time: 25 minutes, including review.
The final question in this chapter is your self-check. Attempt it independently before opening its solution.
LP-preview 4 marks · Technology-active · Model, apply, interpret and justify
Whole daily outputs x,y≥0 use resources x+2y and 2x+y. Under a daily-reset policy each is at most 5 and unused capacity expires. The supplied graph shows the continuous one-day boundary. Find the continuous maximum of x+y and the integer daily maximum; hence find the three-day maximum. Under a different policy unused resource capacity carries forward, with 5 new units of each resource available at each day's start. Test the schedule (2,1),(1,2),(2,2) by cumulative constraints and certify its three-day optimum. Explain why its final day is forbidden under daily reset.
x+2y=5: blue solid curve · 2x+y=5: orange dashed curve · Feasible boundary: purple dotted curve
| Output x (items) | Output y (items) |
|---|---|
| 0 | 2.5 |
| 3 | 1 |
| Output x (items) | Output y (items) |
|---|---|
| 1 | 3 |
| 2.5 | 0 |
| Output x (items) | Output y (items) |
|---|---|
| 0 | 0 |
| 2.5 | 0 |
| 1.6666666666666667 | 1.6666666666666667 |
| 0 | 2.5 |
| 0 | 0 |
Worked solution — LP-preview
- The intersection is (5/3,5/3), giving continuous bound x+y≤10/3 by adding the two constraints. Whole outputs therefore give at most 3 per reset day, achieved by (2,1) or (1,2), and at most 9 over three such days.
- With carry-forward, the schedule's cumulative resource uses are (4,5), then (9,9), then (15,15), within capacities (5,5),(10,10),(15,15). It produces 3+3+4=10 items.
- Across three days the two total constraints give 3(X+Y)≤30, so 10 is optimal under carry-forward. The last day's (2,2) uses (6,6), exceeding both daily-reset limits. This changes capacity timing, rather than allowing fractional items.
Marking guidance — one mark per criterion — LP-preview
- Finds continuous maximum 10/3 and whole daily maximum 3
- Justifies reset three-day maximum 9
- Checks every cumulative capacity and certifies carry-forward optimum 10
- Explains why the final day fails the daily-reset policy
Acceptable alternatives — LP-preview
- Credit complete integer enumeration or an equivalent objective-bound proof. A feasible ten-item schedule must satisfy each day's cumulative supply, not merely the final totals.
Common error: LP-preview — Applying a total three-day capacity to every individual day.
Next step: LP-preview — Write dated cumulative constraints only when unused resources really carry forward.
NET · Connectivity, matrix powers and weighted evidence
Outline 2.1: directed connectivity, powers and multistage walks, redundancy and weighted dominance; interpret limits. Representative topic preview; full resources provide broader applications and depth.
Before you start: Matrix rows/columns and ordered multiplication; a walk may revisit vertices.
Suggested study time: 25 minutes, including review.
The final question in this chapter is your self-check. Attempt it independently before opening its solution.
NET-preview 3 marks · Technology-active · Model, apply, interpret and justify
The supplied unit-weight directed network has vertices ordered A,B,C,D. A separate score matrix, not its adjacency matrix, is W=[[0,2,−2,0],[0,0,0,3],[0,0,0,3],[0,0,0,0]]. Calculate (W²)AD. Construct the network's binary adjacency matrix K and calculate (K²)AD. A report says the zero score proves no two-step A-to-D walk exists. Audit the report using both the actual routes and the meanings of the entries.
Worked solution — NET-preview
- The signed products give (W²)AD=2×3+(−2)×3=0.
- K=[[0,1,1,0],[0,0,0,1],[0,0,0,1],[0,0,0,0]], so (K²)AD=1×1+1×1=2. The routes are A→B→D and A→C→D.
- Opposite signed score contributions cancel without deleting either connection. Non-negative binary adjacency counts walks; a signed score product has a different interpretation.
Marking guidance — one mark per criterion — NET-preview
- Calculates signed score 0 from both products
- Constructs the correct ordered adjacency evidence and finds two actual walks
- Rejects the no-walk inference by distinguishing signed scores from connectivity
Acceptable alternatives — NET-preview
- Accept listing both directed two-step routes as an independent check of K²; retain row-to-column order and distinguish the separate signed-score table.
Common error: NET-preview — Interpreting every zero matrix product as absence of a route.
Next step: NET-preview — Inspect what each original entry measures before interpreting a power.
TRANS · Transitions and the assumptions behind long-run models
Outline 2.2: row-state transitions in two and higher dimensions, repeated updates, steady state, changed probabilities and closed-system assumptions. Representative topic preview; full resources provide broader applications and depth.
Before you start: NET, row probabilities summing to one, expected counts and conservation.
Suggested study time: 25 minutes, including review.
The final question in this chapter is your self-check. Attempt it independently before opening its solution.
TRANS-preview 4 marks · Technology-active · Model, apply, interpret and justify
Conditional on being inspected, a closed two-state model uses T=[[.8,.2],[.1,.9]], ordered A,B. Each day half of A units and one-quarter of B units are inspected, with uninspected units staying. The same inspection and conditional movement rules apply each day. Derive the effective daily transition matrix, forecast the expected row from [60,40], and find a steady-state proportion row. Explain why applying T to everybody is a different model and why expected counts need not always be whole.
Worked solution — TRANS-preview
- Daily A→B probability is .5×.2=.1; B→A is .25×.1=.025. The effective matrix is M=[[.9,.1],[.025,.975]], each row summing to one.
- [60,40]M=[55,45]: expected flows are 6 from A to B and 1 from B to A.
- At a steady state .1a=.025b and a+b=1, giving [.2,.8]. Applying T to every unit assumes inspection of everybody. These are expected counts under the supplied continuing policy, not guaranteed realised movements; expectations can be fractional.
Marking guidance — one mark per criterion — TRANS-preview
- Derives both effective rows including non-inspection stays
- Finds expected row [55,45]
- Solves and normalises steady row [.2,.8]
- Distinguishes conditional inspection, daily movement and expected versus realised counts
Acceptable alternatives — TRANS-preview
- Credit a flow ledger or row-matrix calculation; a valid steady row must both sum to one and reproduce itself under M. No claim of an exact realised daily split is required.
Common error: TRANS-preview — Applying conditional-on-inspection probabilities to uninspected units.
Next step: TRANS-preview — Multiply each change probability by the appropriate state's inspection rate before completing the stay entry.
BIV · Bivariate investigation and model evidence
Outline 3.1: variables, scatter plots, association, r/r², regression, residuals, exponential models, interpolation/extrapolation and causality. Representative topic preview; full resources provide broader applications and depth.
Before you start: Stage 1 data, means, plotted pairs, units and linear/exponential functions.
Suggested study time: 25 minutes, including review.
The final question in this chapter is your self-check. Attempt it independently before opening its solution.
BIV-preview 4 marks · Technology-active · Model, apply, interpret and justify
The supplied graph gives observations (0,8),(1,9),(2,11),(3,12), candidate A: ŷ=8+x and candidate B: ŷ=6+3x. These are candidate lines, not claimed least-squares fits. Calculate actual-minus-predicted residuals and SSE for each. A fictional policy requests an extra check when a value is at least 11. Compare each line's four predicted decisions with the decisions from the observations. Does the smaller SSE guarantee fewer wrong threshold decisions here? Limit your conclusion to this evidence.
Observations: blue circles · A: y=8+x: orange dashed curve · B: y=6+3x: blue solid curve · Extra-check threshold y=11: orange dashed curve
| Explanatory value x | Response y |
|---|---|
| 0 | 8 |
| 1 | 9 |
| 2 | 11 |
| 3 | 12 |
| Explanatory value x | Response y |
|---|---|
| 0 | 8 |
| 3 | 11 |
| Explanatory value x | Response y |
|---|---|
| 0 | 6 |
| 3 | 15 |
| Explanatory value x | Response y |
|---|---|
| 0 | 11 |
| 3 | 11 |
Worked solution — BIV-preview
- A predicts [8,9,10,11], with residuals [0,0,1,1] and SSE=2. B predicts [6,9,12,15], with residuals [2,0,−1,−3] and SSE=14.
- Observed extra-check decisions are [no,no,yes,yes]. A gives [no,no,no,yes], three correct; B gives [no,no,yes,yes], four correct.
- A has smaller squared error but one threshold error, while B has none on these four cases. Prediction-error size and a threshold decision loss differ. Four supplied observations do not establish either candidate's general forecasting performance.
Marking guidance — one mark per criterion — BIV-preview
- Finds both complete residual lists
- Calculates SSE values 2 and 14
- Reconstructs observed/predicted decisions and counts 3 versus 4 correct
- Explains the different criteria and limits generalisation
Acceptable alternatives — BIV-preview
- Accept a four-row decision table or explicit false-negative identification at x=2. Residuals must use actual minus predicted consistently; neither line is asserted to be an OLS fit.
Common error: BIV-preview — Treating the lowest numeric loss as a guarantee under every decision rule.
Next step: BIV-preview — Apply the actual threshold separately to observations and forecasts.
NORM · Sample workshop — observed shape and a proposed model
A limited selection from outline 3.2: interpret observed grouped data and compare an event with a supplied normal model.
Before you start: Histogram area, relative frequency and standard scores; use the supplied normal model conditionally.
Suggested study time: 25 minutes, including review.
The final question in this chapter is your self-check. Attempt it independently before opening its solution.
NORM-1 4 marks · Technology-active · Model, apply, interpret and justify
Use the supplied histogram of 30 observations. Find the count in the 3–5 class from its width and density, and its observed proportion of the full sample. A separate proposed normal model is N(4,2²). Calculate its probability for 3<X<5 and compare with the observed proportion. Explain why neither unequal bar heights nor one approximate interval comparison establishes or refutes the complete normal model; the within-class positions are unknown.
| Class | Frequency | Frequency density |
|---|---|---|
| [0, 2) | 4 | 2 |
| [2, 3) | 6 | 6 |
| [3, 5) | 10 | 5 |
| [5, 6) | 6 | 6 |
| [6, 8] | 4 | 2 |
Worked solution — NORM-1
- The 3–5 bar has width 2 and density 5, so contains 10 observations; its observed proportion is 10/30=1/3. The proposed normal boundaries are z=−.5 and .5, giving probability Φ(.5)−Φ(−.5)≈.3829249.
- This is about .049592 above the observed proportion, but a sample need not exactly reproduce a model probability. Heights must be read as density when widths differ. Inspect the whole shape and further evidence; grouped counts do not supply exact within-class positions or prove a continuous theoretical distribution.
Marking guidance — one mark per criterion — NORM-1
- Recovers count 10 from histogram area.
- Finds the observed proportion 1/3.
- Calculates the normal-model probability about .382925.
- Explains density, finite-sample variation and the limits of this one model comparison.
Acceptable alternatives — NORM-1
- Credit other valid algebraic, graphical, numerical, financial, statistical or algorithmic methods with the requested supporting evidence. Accept equivalent exact or appropriately rounded values and justified contextual conclusions. Follow a plausible earlier numerical error through a valid later method once without restoring the earlier lost mark; reject impossible probabilities, infeasible assignments, invalid time order or an unsupported model domain. Do not award the same reasoning twice.
Common error: NORM-1 — Comparing unequal-width histogram heights as if they were counts or treating a theoretical probability as a mandatory sample frequency.
Next step: NORM-1 — Recover counts from bar areas, then distinguish observed proportions from conditional model probabilities.
SAVE · Sample workshop — a tiered interest rule
A limited selection from outline 4.1: translate a supplied fictional tiered interest rule into a dated savings ledger.
Before you start: Percentage calculations and interest-before-deposit timing.
Suggested study time: 25 minutes, including review.
The final question in this chapter is your self-check. Attempt it independently before opening its solution.
SAVE-1 3 marks · Technology-active · Model, apply, interpret and justify
A fictional account starts at $800. At each year-end it credits 2% on the first $500 of the opening balance and 6% only on any excess, then receives a $100 deposit. Find its balances after years 1 and 2. For balances staying above $500, derive a single affine update. Compare with incorrectly applying 6% to the whole opening balance in both years, and quantify the final overstatement.
Worked solution — SAVE-1
- Year-1 interest is 10+.06(300)=28, so balance becomes 928. Year-2 interest is 10+.06(428)=35.68, giving final balance 1063.68 after the next deposit.
- Above 500, B(next)=B+10+.06(B−500)+100=1.06B+80. The wrong whole-balance rule gives 948 then 1104.88, overstating the final value by $41.20. The tier threshold is applied to each year's actual opening balance.
Marking guidance — one mark per criterion — SAVE-1
- Builds the two correctly timed tiered-interest ledger steps.
- Derives and states the valid above-threshold affine update.
- Calculates the wrong-rule balance and overstatement 41.20 with an explanation of the rate base.
Acceptable alternatives — SAVE-1
- Credit other valid algebraic, graphical, numerical, financial, statistical or algorithmic methods with the requested supporting evidence. Accept equivalent exact or appropriately rounded values and justified contextual conclusions. Follow a plausible earlier numerical error through a valid later method once without restoring the earlier lost mark; reject impossible probabilities, infeasible assignments, invalid time order or an unsupported model domain. Do not award the same reasoning twice.
Common error: SAVE-1 — Treating an excess-only advertised tier as a whole-account rate.
Next step: SAVE-1 — Split each actual opening balance at the threshold before applying interest and the dated deposit.
LOAN · Borrowing models and defensible comparisons
Outline 4.2: reducing-balance and interest-only loans, sinking funds, payments/balances/term, strategies and comparison rates. Representative topic preview; full resources provide broader applications and depth.
Before you start: SAVE, signed cash flows, interest versus principal and valid final payments.
Suggested study time: 25 minutes, including review.
The final question in this chapter is your self-check. Attempt it independently before opening its solution.
LOAN-preview 4 marks · Technology-active · Model, apply, interpret and justify
A supplied fictional contract starts with principal $1200 and unpaid interest $0. Each year new interest is 10% of principal only, added to a separate unpaid-interest balance. That balance never earns interest and is never capitalised. A $50 year-end payment first reduces unpaid interest; any remainder reduces principal. Show both balances after each of two years and find the total settlement then. Compare the single-balance update B(next)=1.1B−50 and explain any discrepancy. This supplied model is a learning assumption, not a real loan recommendation.
Worked solution — LOAN-preview
- Year 1 adds $120 interest, then pays $50, leaving principal 1200 and unpaid interest 70: total 1270.
- Year 2 again charges only $120 on principal. After payment, unpaid interest is 70+120−50=140, principal remains 1200 and settlement is 1340.
- The single-balance update gives 1270 then 1347, overstating settlement by $7. It charges 10% on the first year's $70 unpaid interest, contrary to the stated non-capitalising contract.
Marking guidance — one mark per criterion — LOAN-preview
- Records year-1 principal 1200 and unpaid interest 70
- Records year-2 principal 1200 and unpaid interest 140 using the correct interest base
- Finds correct settlement 1340 and single-balance result 1347
- Explains the $7 discrepancy as forbidden interest on unpaid interest
Acceptable alternatives — LOAN-preview
- Credit an equivalent two-account recurrence or cash-flow ledger. Do not compound the unpaid-interest account; principal stays unchanged because each payment is less than the interest due.
Common error: LOAN-preview — Combining balances that have different interest rules.
Next step: LOAN-preview — Keep principal and unpaid interest in separate dated columns before applying a payment.
CPA · Sample workshop — all controlling routes matter
A limited selection from outline 5.1: scan a complete activity-on-arrow network and compare stated one-activity reductions.
Before you start: Directed precedence, maximum at a join and complete route durations.
Suggested study time: 25 minutes, including review.
The final question in this chapter is your self-check. Attempt it independently before opening its solution.
CPA-1 3 marks · Technology-active · Model, apply, interpret and justify
Use the supplied network, with all activities required and sufficient resources for the stated parallelism. Complete earliest and latest event times and list every critical route. Compare shortening D by one hour with shortening common activity A by one hour; all other times stay fixed. Explain why finding one critical route alone is insufficient for the shortening decision.
Worked solution — CPA-1
- Earliest event times 1–6 are [0,2,3,4,5,7]; latest times are the same. The three critical routes are A–B–G, A–C–D–G and A–E–F–G, each 7 hours.
- Shortening D leaves the other two routes at 7, so the project does not improve. Shortening A reduces all three to 6, so finish becomes 6. A decision based on one listed critical route can miss another route still controlling the join.
Marking guidance — one mark per criterion — CPA-1
- Completes both event-time tables.
- Identifies all three seven-hour critical routes.
- Recalculates both shortening cases and explains why the common activity changes every controlling route.
Acceptable alternatives — CPA-1
- Credit other valid algebraic, graphical, numerical, financial, statistical or algorithmic methods with the requested supporting evidence. Accept equivalent exact or appropriately rounded values and justified contextual conclusions. Follow a plausible earlier numerical error through a valid later method once without restoring the earlier lost mark; reject impossible probabilities, infeasible assignments, invalid time order or an unsupported model domain. Do not award the same reasoning twice.
Common error: CPA-1 — Assuming a reduction on any one critical route must shorten the whole project.
Next step: CPA-1 — List every controlling route and recompute their maximum after the proposed change.
ASSIGN · Assignment models and optimality evidence
Outline 5.2: one-to-one allocation, Hungarian reductions/zero covers/adjustments, minimum/maximum/rectangular cases and limitations. Representative topic preview; full resources provide broader applications and depth.
Before you start: Matrices and feasible one-to-one choices; use original values for the reported objective.
Suggested study time: 25 minutes, including review.
The final question in this chapter is your self-check. Attempt it independently before opening its solution.
ASSIGN-preview 4 marks · Technology-active · Model, apply, interpret and justify
Four workers A,B,C,R cover jobs 1,2,3 and one standby place one-to-one. Real-job cost rows are A:[5,9,unqualified], B:[unqualified,5,9], C:[9,unqualified,5], R:[1,1,1]. Every standby cost is explicitly 0. First minimise planned roster cost without absence protection, giving an allocation and lower bound. Then require the standby worker to replace ANY ONE absent assigned worker directly, with no other reassignment; the standby worker is available and at most one assigned worker is absent. Find the cheapest qualifying roster and its extra planned cost. Explain why merely adding a zero standby column does not enforce this guarantee. Optimise the stated planned costs, not unspecified post-absence costs.
Worked solution — ASSIGN-preview
- At least two of A,B,C must be on real jobs, each costing at least 5; if R is active its cost is 1. Thus every roster costs at least 11, attained for example by R1/B2/C3 with A on standby. Its reserve A cannot replace C on job 3.
- The standby worker must be qualified for all three occupied jobs. Only R has all three qualifications, so R must stand by. A1/B2/C3 costs 15 and attains the row-minimum bound 5+5+5. The only other feasible full A/B/C allocation is A2/B3/C1, costing 27.
- The guaranteed roster is therefore A1/B2/C3/R-standby, with extra planned cost 15−11=4. The zero standby column expresses a place and its price, but says nothing about the standby worker's ability to replace every occupied job. This is one direct replacement, not a second complete backup allocation.
Marking guidance — one mark per criterion — ASSIGN-preview
- Certifies nominal optimum 11 with a feasible roster and lower bound
- Deduces that only universally qualified R can provide the direct-replacement guarantee
- Certifies qualifying optimum 15 and extra planned cost 4
- Explains why a zero standby column alone omits replacement qualifications
Acceptable alternatives — ASSIGN-preview
- Accept any nominal roster attaining 11. Credit a complete feasible-roster enumeration or the qualification/row-minimum certificate. Do not assign an invented cost to an unqualified pairing or charge unspecified replacement costs.
Common error: ASSIGN-preview — Treating an available reserve as qualified to replace every assigned worker.
Next step: ASSIGN-preview — Test each occupied job against the reserve's qualifications before calling a roster protected.