Short answer: site coverage measures a qualifying top-down footprint relative to site area; gross floor area sums qualifying floor areas across levels.
This guide develops one missing reader question within the Zoning subject guide. It adds original worked decisions, changed-context tests and explicit limits. It should be read with The Gold Standard of Education | Learning That Transfers into Life and the How X Works.
The numerical cases are fictional teaching models. They do not grant planning permission, issue a weather or volcanic warning, certify a structure, or replace professional assessment. Current official sources and the responsible authority govern real decisions.
eduKateSG · Zoning
Find your next learning step
Build separate footprint and floor-area ledgers, then connect each to the correct site denominator and planning control.
ROUTE 1 · CHAPTERS 1–6
Define the two ledgers
Separate site area, footprint, GFA and non-unique building forms.
ROUTE 2 · CHAPTERS 7–11
Connect planning ratios
Relate GFA to plot ratio, storeys, retained area, mixed use and towers.
ROUTE 3 · CHAPTERS 12–17
Change classifications
Test vesting, projections, bounds, separate limits, open space and comparison.
ROUTE 4 · CHAPTERS 18–22
Communicate and continue
Document calculations, repair misconceptions and follow official guidance.
Chapter contents
Define the two ledgers · Chapters 1–6
1. Start with the short answer
2. Define the site-area denominator
3. Define the footprint boundary
4. Define gross floor area separately
Connect planning ratios · Chapters 7–11
7. Connect GFA to gross plot ratio
8. Do not read plot ratio as storey count
9. Keep existing and proposed quantities separate
Change classifications · Chapters 12–17
12. Changed context: road vesting changes net area
13. Changed context: a cantilever or projection
14. Use bounds when classifications are unresolved
Communicate and continue · Chapters 18–22
18. Communicate an illustrative calculation
19. Design a practical learner plan
20. Misconceptions that distort form
Site coverage measures the top-down ground footprint of qualifying building structures as a share of net site area under the applicable rule. Gross floor area adds qualifying floor areas across levels. One describes footprint; the other contributes to development intensity. Neither number determines the other by itself.
Write the quantity, unit, population or place boundary, and time window before calculating. Then mark each input as observed, stipulated, estimated or assumed. This short ledger prevents a familiar word from carrying an unstated denominator. It also lets another reader reproduce the result and identify which new observation would reduce the uncertainty that matters to the decision. Read the ledger aloud as a sentence: quantity, place, period, evidence state and calculation rule. If any element sounds vague, pause before doing arithmetic. A later reader should be able to replace one input without silently changing the others.
On a fictional 4,000-square-metre net site, a 40 per cent footprint is 1,600 square metres. Four equal qualifying floors of 1,200 square metres give 4,800 square metres of GFA. The footprint and total floor area answer different questions.
A useful transfer check is to change the number and size of upper floors while holding the footprint fixed. State what would make the answer change and what the example still does not establish. That final limit is part of the result, not an apology attached after it.
Gross site area, net site area and a sub-parcel can differ. A coverage percentage is meaningless until the applicable denominator is named.
Keep the mechanism beside the number. Ask what physical, administrative or behavioural relationship connects the input to the conclusion, and what could break that connection. A calculation is strongest when its conditions remain visible. If a changed case needs a different denominator, interval, density or responsibility, revise the model instead of forcing the new case through the old arithmetic. Draw one arrow from each input to the step it controls, then challenge the arrow with a counterexample. This makes the transfer task visible: learners must decide whether the relationship survives, needs qualification or no longer applies.
A gross parcel is 5,000 m², with 500 m² vested for a road reserve in the teaching model. Net site area is 4,500 m². A 1,800 m² footprint is 36 per cent of gross area but 40 per cent of net area.
A useful transfer check is to recalculate both percentages and explain which denominator the stipulated rule uses. State what would make the answer change and what the example still does not establish. That final limit is part of the result, not an apology attached after it.
A top-down footprint can include or exclude structures according to the applicable definition. A roof, projection, basement or landscape element should not be classified from intuition alone.
Use an independent check that follows a different route from the main calculation. Reconstruct a total from components, compare a bound with a point estimate, or test the result against a changed example. Agreement does not prove every assumption, but disagreement exposes where a definition, unit or boundary has drifted. Record the mismatch rather than tuning an input until it disappears. A good check has its own written expectation before the answer is known. That protects the check from becoming a second performance of the same mistake and turns disagreement into evidence about where to inspect next.
The fictional plan has a 1,500 m² main block and a 120 m² qualifying above-ground structure. If both count, footprint is 1,620 m². Omitting the smaller structure understates coverage by 120 m².
A useful transfer check is to mark each plan component as included, excluded or unresolved before summing. State what would make the answer change and what the example still does not establish. That final limit is part of the result, not an apology attached after it.
GFA is not simply the footprint multiplied by the number of storeys. Floors can step back, contain voids or include spaces treated differently by the current handbook.
Match the precision and strength of the claim to the evidence. A tidy decimal cannot repair an approximate observation, and a broad planning statement cannot be proved by one convenient location. Report the result, the conditions under which it holds, and the next responsible action. Operational or safety decisions should use current official information and qualified assessment for the actual site. Separate the arithmetic conclusion from the decision conclusion in two sentences. The first states what follows from the supplied model; the second identifies the authority, observation or threshold still needed before anyone can act responsibly.
A building has qualifying floors of 1,500, 1,200, 900 and 600 m². Total GFA is 4,200 m², while the largest footprint remains 1,500 m² in the simplified case.
A useful transfer check is to compare the exact sum with the shortcut 1,500 times four and explain the 1,800 m² error. State what would make the answer change and what the example still does not establish. That final limit is part of the result, not an apology attached after it.
CHAPTER 5 OF 22 · Define the two ledgers
5. Worked example: same coverage, different GFA
Two schemes can occupy the same ground share and create very different floor totals. The vertical distribution matters.
Write the quantity, unit, population or place boundary, and time window before calculating. Then mark each input as observed, stipulated, estimated or assumed. This short ledger prevents a familiar word from carrying an unstated denominator. It also lets another reader reproduce the result and identify which new observation would reduce the uncertainty that matters to the decision. Read the ledger aloud as a sentence: quantity, place, period, evidence state and calculation rule. If any element sounds vague, pause before doing arithmetic. A later reader should be able to replace one input without silently changing the others.
On the same 4,000 m² net site, both schemes cover 1,600 m². Scheme A has two full 1,600 m² floors, for 3,200 m² GFA. Scheme B has five 1,200 m² floors above the same qualifying footprint arrangement, for 6,000 m² GFA under the stipulated simplified account.
A useful transfer check is to hold coverage constant and construct a third GFA total between the two. State what would make the answer change and what the example still does not establish. That final limit is part of the result, not an apology attached after it.
CHAPTER 6 OF 22 · Define the two ledgers
6. Worked example: same GFA, different coverage
The same total floor area can be arranged as a broad low form or a narrower taller form, subject to many other controls.
Keep the mechanism beside the number. Ask what physical, administrative or behavioural relationship connects the input to the conclusion, and what could break that connection. A calculation is strongest when its conditions remain visible. If a changed case needs a different denominator, interval, density or responsibility, revise the model instead of forcing the new case through the old arithmetic. Draw one arrow from each input to the step it controls, then challenge the arrow with a counterexample. This makes the transfer task visible: learners must decide whether the relationship survives, needs qualification or no longer applies.
Scheme C uses three 1,600 m² floors for 4,800 m² GFA and 40 per cent coverage. Scheme D uses six 800 m² floors for the same 4,800 m² GFA and 20 per cent coverage on the fictional site.
A useful transfer check is to state why equal GFA does not establish equal height, open space or approval. State what would make the answer change and what the example still does not establish. That final limit is part of the result, not an apology attached after it.
Gross plot ratio divides GFA by site or plot area under the governing definition. It is a ratio of total qualifying floor area to land area, not a percentage footprint.
Use an independent check that follows a different route from the main calculation. Reconstruct a total from components, compare a bound with a point estimate, or test the result against a changed example. Agreement does not prove every assumption, but disagreement exposes where a definition, unit or boundary has drifted. Record the mismatch rather than tuning an input until it disappears. A good check has its own written expectation before the answer is known. That protects the check from becoming a second performance of the same mistake and turns disagreement into evidence about where to inspect next.
With 4,800 m² GFA on a 4,000 m² site, the simplified GPR is 1.2. Coverage remains 40 per cent if footprint is 1,600 m². The two values coexist without being interchangeable.
A useful transfer check is to double the GFA while holding land area and footprint constant, then recompute GPR. State what would make the answer change and what the example still does not establish. That final limit is part of the result, not an apology attached after it.
CHAPTER 8 OF 22 · Connect planning ratios
8. Do not read plot ratio as storey count
A ratio does not specify identical full floor plates. Setbacks, height, envelope, use, circulation and exclusions affect form.
Match the precision and strength of the claim to the evidence. A tidy decimal cannot repair an approximate observation, and a broad planning statement cannot be proved by one convenient location. Report the result, the conditions under which it holds, and the next responsible action. Operational or safety decisions should use current official information and qualified assessment for the actual site. Separate the arithmetic conclusion from the decision conclusion in two sentences. The first states what follows from the supplied model; the second identifies the authority, observation or threshold still needed before anyone can act responsibly.
A GPR of 1.2 on 4,000 m² permits a simplified 4,800 m² GFA target, but that target could be distributed across several floor patterns. Dividing 1.2 by 0.4 suggests three equivalent full footprints only in a deliberately uniform model.
A useful transfer check is to state every extra assumption needed for the ratio-to-storeys shortcut. State what would make the answer change and what the example still does not establish. That final limit is part of the result, not an apology attached after it.
CHAPTER 9 OF 22 · Connect planning ratios
9. Keep existing and proposed quantities separate
A redevelopment check needs current retained areas, removed areas and new areas. Adding the proposed total to the existing total can double-count retained floors.
Write the quantity, unit, population or place boundary, and time window before calculating. Then mark each input as observed, stipulated, estimated or assumed. This short ledger prevents a familiar word from carrying an unstated denominator. It also lets another reader reproduce the result and identify which new observation would reduce the uncertainty that matters to the decision. Read the ledger aloud as a sentence: quantity, place, period, evidence state and calculation rule. If any element sounds vague, pause before doing arithmetic. A later reader should be able to replace one input without silently changing the others.
A site has 3,000 m² existing GFA. The proposal retains 1,000, removes 2,000 and adds 3,500. Final GFA is 1,000 + 3,500 = 4,500 m², not 6,500.
A useful transfer check is to reconcile final area from both retained-plus-new and existing-minus-removed-plus-new routes. State what would make the answer change and what the example still does not establish. That final limit is part of the result, not an apology attached after it.
CHAPTER 10 OF 22 · Connect planning ratios
10. Handle mixed-use apportionment carefully
Use shares of GFA when discussing floor-area mix and shares of site area when discussing land allocation. A 30 per cent commercial GFA share need not occupy 30 per cent of the ground.
Keep the mechanism beside the number. Ask what physical, administrative or behavioural relationship connects the input to the conclusion, and what could break that connection. A calculation is strongest when its conditions remain visible. If a changed case needs a different denominator, interval, density or responsibility, revise the model instead of forcing the new case through the old arithmetic. Draw one arrow from each input to the step it controls, then challenge the arrow with a counterexample. This makes the transfer task visible: learners must decide whether the relationship survives, needs qualification or no longer applies.
A 6,000 m² mixed-use development has 1,800 m² commercial GFA, or 30 per cent. If shops occupy 900 m² of a 1,500 m² footprint, they use 60 per cent of that ground-level footprint while remaining 30 per cent of total GFA.
A useful transfer check is to label every percentage as GFA share, footprint share or site share. State what would make the answer change and what the example still does not establish. That final limit is part of the result, not an apology attached after it.
A podium-and-tower form makes the distinction visible. The podium may set coverage while upper levels add substantial GFA on a smaller plate.
Use an independent check that follows a different route from the main calculation. Reconstruct a total from components, compare a bound with a point estimate, or test the result against a changed example. Agreement does not prove every assumption, but disagreement exposes where a definition, unit or boundary has drifted. Record the mismatch rather than tuning an input until it disappears. A good check has its own written expectation before the answer is known. That protects the check from becoming a second performance of the same mistake and turns disagreement into evidence about where to inspect next.
A 1,600 m² podium and five 600 m² tower floors yield 4,600 m² total floor area if all count. Site coverage on 4,000 m² is 40 per cent, while simplified GPR is 1.15.
A useful transfer check is to remove one tower floor and show which metric changes and which stays fixed. State what would make the answer change and what the example still does not establish. That final limit is part of the result, not an apology attached after it.
CHAPTER 12 OF 22 · Change classifications
12. Changed context: road vesting changes net area
When the denominator changes, the same physical footprint produces a different coverage percentage and GPR. Preserve the applicable site definition and date.
Match the precision and strength of the claim to the evidence. A tidy decimal cannot repair an approximate observation, and a broad planning statement cannot be proved by one convenient location. Report the result, the conditions under which it holds, and the next responsible action. Operational or safety decisions should use current official information and qualified assessment for the actual site. Separate the arithmetic conclusion from the decision conclusion in two sentences. The first states what follows from the supplied model; the second identifies the authority, observation or threshold still needed before anyone can act responsibly.
A 1,600 m² footprint on 4,000 m² net area is 40 per cent. If an applicable 400 m² vesting changes net area to 3,600 m² with footprint unchanged, coverage becomes about 44.4 per cent.
A useful transfer check is to recompute GPR for 4,800 m² GFA under both denominators. State what would make the answer change and what the example still does not establish. That final limit is part of the result, not an apology attached after it.
CHAPTER 13 OF 22 · Change classifications
13. Changed context: a cantilever or projection
An upper-level projection can change the top-down envelope without adding ground contact. Whether it enters site coverage follows the actual rule, while its floor area may separately enter GFA.
Write the quantity, unit, population or place boundary, and time window before calculating. Then mark each input as observed, stipulated, estimated or assumed. This short ledger prevents a familiar word from carrying an unstated denominator. It also lets another reader reproduce the result and identify which new observation would reduce the uncertainty that matters to the decision. Read the ledger aloud as a sentence: quantity, place, period, evidence state and calculation rule. If any element sounds vague, pause before doing arithmetic. A later reader should be able to replace one input without silently changing the others.
In a fictional rule, a 150 m² projection counts toward the top-down footprint. Coverage rises from 1,500 to 1,650 m², while GFA also changes only if the projection’s floor area is included under the separate GFA definition.
A useful transfer check is to keep the two classification decisions separate instead of assuming one controls both. State what would make the answer change and what the example still does not establish. That final limit is part of the result, not an apology attached after it.
CHAPTER 14 OF 22 · Change classifications
14. Use bounds when classifications are unresolved
If two components await classification, show lower and upper totals rather than silently choosing the convenient interpretation.
Keep the mechanism beside the number. Ask what physical, administrative or behavioural relationship connects the input to the conclusion, and what could break that connection. A calculation is strongest when its conditions remain visible. If a changed case needs a different denominator, interval, density or responsibility, revise the model instead of forcing the new case through the old arithmetic. Draw one arrow from each input to the step it controls, then challenge the arrow with a counterexample. This makes the transfer task visible: learners must decide whether the relationship survives, needs qualification or no longer applies.
Base footprint is 1,450 m², with 80 and 120 m² unresolved structures. Coverage on 4,000 m² lies from 36.25 per cent to 41.25 per cent if all combinations are permitted.
A useful transfer check is to construct both endpoints and state which official clarification would narrow the range. State what would make the answer change and what the example still does not establish. That final limit is part of the result, not an apology attached after it.
Meeting a coverage value does not establish compliance with GFA, height, setbacks, use, access or other controls. A real proposal needs the applicable complete framework.
Use an independent check that follows a different route from the main calculation. Reconstruct a total from components, compare a bound with a point estimate, or test the result against a changed example. Agreement does not prove every assumption, but disagreement exposes where a definition, unit or boundary has drifted. Record the mismatch rather than tuning an input until it disappears. A good check has its own written expectation before the answer is known. That protects the check from becoming a second performance of the same mistake and turns disagreement into evidence about where to inspect next.
A fictional scheme has 35 per cent coverage below a 50 per cent classroom cap but a GPR of 2.0 above a stipulated 1.6 cap. Passing the first check does not cancel the second.
A useful transfer check is to build a two-column compliance matrix and avoid netting one surplus against another breach. State what would make the answer change and what the example still does not establish. That final limit is part of the result, not an apology attached after it.
CHAPTER 16 OF 22 · Change classifications
16. Do not confuse open space with uncovered site
Uncovered area is not automatically usable public, communal, landscaped or permeable space. Those qualities require their own definitions and evidence.
Match the precision and strength of the claim to the evidence. A tidy decimal cannot repair an approximate observation, and a broad planning statement cannot be proved by one convenient location. Report the result, the conditions under which it holds, and the next responsible action. Operational or safety decisions should use current official information and qualified assessment for the actual site. Separate the arithmetic conclusion from the decision conclusion in two sentences. The first states what follows from the supplied model; the second identifies the authority, observation or threshold still needed before anyone can act responsibly.
If coverage is 40 per cent, 60 per cent of the site is outside the counted footprint under the model. That 60 per cent can include driveways, service areas and other surfaces; it is not automatically 60 per cent park.
A useful transfer check is to classify the uncovered area by function before making a liveability claim. State what would make the answer change and what the example still does not establish. That final limit is part of the result, not an apology attached after it.
CHAPTER 17 OF 22 · Change classifications
17. Compare schemes with consistent rules
A comparison fails if one scheme uses net site area and another gross area, or one includes a structure the other omits. Freeze definitions before ranking.
Write the quantity, unit, population or place boundary, and time window before calculating. Then mark each input as observed, stipulated, estimated or assumed. This short ledger prevents a familiar word from carrying an unstated denominator. It also lets another reader reproduce the result and identify which new observation would reduce the uncertainty that matters to the decision. Read the ledger aloud as a sentence: quantity, place, period, evidence state and calculation rule. If any element sounds vague, pause before doing arithmetic. A later reader should be able to replace one input without silently changing the others.
Scheme A reports 45 per cent coverage on 4,000 m² net area; Scheme B reports 42 per cent on 4,500 m² gross area. Their footprints are 1,800 and 1,890 m², so the lower percentage has the larger footprint.
A useful transfer check is to restate both on a common authorised denominator before comparing. State what would make the answer change and what the example still does not establish. That final limit is part of the result, not an apology attached after it.
CHAPTER 18 OF 22 · Communicate and continue
18. Communicate an illustrative calculation
A clear note names site area, footprint components, GFA components, rule source and unresolved classifications. It avoids implying planning approval.
Keep the mechanism beside the number. Ask what physical, administrative or behavioural relationship connects the input to the conclusion, and what could break that connection. A calculation is strongest when its conditions remain visible. If a changed case needs a different denominator, interval, density or responsibility, revise the model instead of forcing the new case through the old arithmetic. Draw one arrow from each input to the step it controls, then challenge the arrow with a counterexample. This makes the transfer task visible: learners must decide whether the relationship survives, needs qualification or no longer applies.
A defensible classroom note says: ‘On the stipulated 4,000 m² net site, counted footprint is 1,600 m² (40%); qualifying floor areas total 4,800 m² (GPR 1.2).’ It then lists assumptions.
A useful transfer check is to rewrite a statement that calls plot ratio ‘forty per cent site coverage’. State what would make the answer change and what the example still does not establish. That final limit is part of the result, not an apology attached after it.
Give a simple site plan and section with labelled areas. Ask for footprint, coverage, GFA and GPR, then change one floor plate or denominator.
Use an independent check that follows a different route from the main calculation. Reconstruct a total from components, compare a bound with a point estimate, or test the result against a changed example. Agreement does not prove every assumption, but disagreement exposes where a definition, unit or boundary has drifted. Record the mismatch rather than tuning an input until it disappears. A good check has its own written expectation before the answer is known. That protects the check from becoming a second performance of the same mistake and turns disagreement into evidence about where to inspect next.
The learner first sums four floors, then identifies the maximum top-down qualifying envelope. A transfer card removes one floor and adds a projection, testing whether the two ledgers remain separate.
A useful transfer check is to require two independent totals and a sentence on what the plan cannot establish. State what would make the answer change and what the example still does not establish. That final limit is part of the result, not an apology attached after it.
Site coverage is not GFA; GPR is not storey count; uncovered area is not automatically public space; a map value is not permission; and identical ratios do not prove identical design.
Match the precision and strength of the claim to the evidence. A tidy decimal cannot repair an approximate observation, and a broad planning statement cannot be proved by one convenient location. Report the result, the conditions under which it holds, and the next responsible action. Operational or safety decisions should use current official information and qualified assessment for the actual site. Separate the arithmetic conclusion from the decision conclusion in two sentences. The first states what follows from the supplied model; the second identifies the authority, observation or threshold still needed before anyone can act responsibly.
Ask two learners to create different four-level floor patterns with the same total GFA. If both are possible under the teaching constraints, the ratio has not determined a unique form.
A useful transfer check is to produce a counterexample to each false equivalence. State what would make the answer change and what the example still does not establish. That final limit is part of the result, not an apology attached after it.
CHAPTER 21 OF 22 · Communicate and continue
21. Follow the existing zoning branches
Use unequal floor-plate limits for level-specific capacity, development-capacity transfers for sender-receiver rules, boundary-angle uncertainty for geometric clearance and activity time windows for use evidence. This guide owns footprint versus total floor area.
Write the quantity, unit, population or place boundary, and time window before calculating. Then mark each input as observed, stipulated, estimated or assumed. This short ledger prevents a familiar word from carrying an unstated denominator. It also lets another reader reproduce the result and identify which new observation would reduce the uncertainty that matters to the decision. Read the ledger aloud as a sentence: quantity, place, period, evidence state and calculation rule. If any element sounds vague, pause before doing arithmetic. A later reader should be able to replace one input without silently changing the others.
The branches share a planning setting but have different quantities and tests. Choose the one matching the actual reader job.
A useful transfer check is to name the target quantity before selecting a sibling method. State what would make the answer change and what the example still does not establish. That final limit is part of the result, not an apology attached after it.
CHAPTER 22 OF 22 · Communicate and continue
22. Sources, limits and a final transfer
URA’s current Gross Floor Area handbook defines GFA and gross plot ratio for development control. URA’s site coverage guidance for flats and condominiums describes the top-down site-coverage computation and net site area for that development type. Sources checked on 26 October 2026.
Keep the mechanism beside the number. Ask what physical, administrative or behavioural relationship connects the input to the conclusion, and what could break that connection. A calculation is strongest when its conditions remain visible. If a changed case needs a different denominator, interval, density or responsibility, revise the model instead of forcing the new case through the old arithmetic. Draw one arrow from each input to the step it controls, then challenge the arrow with a counterexample. This makes the transfer task visible: learners must decide whether the relationship survives, needs qualification or no longer applies.
The official pages apply within their stated development-control scope and may change. The fictional shapes and classifications here do not interpret a real property or establish approval.
A useful transfer check is to solve a new stepped-building case, cite the current applicable source and list every unresolved component before advising on a real proposal. State what would make the answer change and what the example still does not establish. That final limit is part of the result, not an apology attached after it.