WHITE WARP — Technology for Real Estate

Feasibility Engine
Methodology & Formula Reference

Complete disclosure of calculation methods, data sources, accuracy bounds,
and limitations — for professional review.
Engine Version: rev 00041+ Jurisdiction Coverage: Delhi, Gurugram, Noida, YEIDA, Chennai Issued: May 2026
Transparency Document — Public

White Warp is built on the belief that opacity is the biggest risk in real estate. This document explains exactly how every number in a White Warp feasibility report is calculated. A Chartered Accountant, RICS-certified Quantity Surveyor, or IIT-trained economist should be able to reproduce any output using the formulas below and the same input data. Where the engine makes assumptions, those assumptions are stated explicitly. Where the accuracy is limited, those limits are disclosed.

SECTION 01

Cost Estimation Model

1.1 Hard Construction Cost

Base rates are sourced from the Delhi Schedule of Rates (DSR) published annually by the Government of Delhi, PWD. We use the most recent edition (DSR 2024–25). DSR rates reflect government-grade construction. Private development costs are consistently higher due to higher-quality finishes and contractor margin structures.

Formula — Hard Construction Cost (per sqm)
Hard_Cost_per_sqm = DSR_base_rate × location_factor × finish_multiplier × OP_factor

Where: DSR_base_rate = PWD Delhi DSR 2024-25 (RCC framed structure, per sqm) location_factor = 1.00 (Delhi) | 1.08 (Gurugram) | 1.05 (Noida) | 0.92 (Chennai) finish_multiplier = 0.85 (govt-grade) | 1.00 (standard) | 1.40 (premium) | 1.60 (luxury) OP_factor = 1.20 (20% contractor overhead + profit; CPWD standard is 15%, private market is 18–25%; we use 20% as the conservative mid)
Why 20% O&P and not 15%? CPWD uses 15% because it awards contracts via competitive tender to the lowest bidder. Private developers in NCR pay 18–25% because they use negotiated contracts with preferred contractors and require faster delivery timelines. We use 20% as the midpoint. For luxury projects, 22–25% is more accurate.

1.2 Soft Costs (added on top of hard cost)

Formula — Soft Costs
RERA_fee = max(0.5% × project_cost, ₹50,000) [RERA Act 2016] approval_premiums = ₹50–₹150/sqft of built-up area [MCD/DDA/DTCP/GNIDA schedule] professional_fees = 2.5% × hard_cost [architect + MEP + PMC + structural] GST_on_services = 18% × professional_fees [GST Act, Sch. II, Clause 5(b)]

Total_Soft_Costs = RERA_fee + approval_premiums + professional_fees + GST_on_services

1.3 Interest During Construction (IDC)

Formula — IDC (Interest During Construction)
IDC = Hard_Cost × interest_rate × drawdown_years × average_utilisation Where: interest_rate = 14% per annum (MCLR + 3%, typical Indian developer financing) drawdown_years = construction_months / 12 average_utilisation = 0.50 (S-curve: 0% drawn at start, 100% at completion → avg 50%)

Example: Hard_Cost = ₹100L, 18-month construction IDC = 100L × 14% × 1.5yr × 0.50 = ₹10.5L

1.4 Total Cost (P50 and P85)

Formula — Total Project Cost
P50_Total_Cost = Hard_Cost + Total_Soft_Costs + IDC + Land_Cost P85_Total_Cost = (Hard_Cost + Total_Soft_Costs + IDC) × 1.275 + Land_Cost Note: P85 = "commitment-grade" estimate. The 1.275 multiplier is derived from: - MoSPI infrastructure overrun data: 20–37% average overrun on Indian projects - IIT Delhi construction cost studies: +20–25% for residential projects - Tayyab et al. 2023: >70% of Indian high-rise projects exceed initial budget We use 27.5% (P85) as the overrun budget for financing decisions. Rule of thumb: "Borrow against P85. Sell against P50."
SECTION 02

Revenue Projection Model

2.1 Sellable Area Calculation

Formula — Sellable Area
Built_Up_Area = Plot_Area × FAR × num_floors_adjustment Sellable_Area = Built_Up_Area × efficiency_ratio

Where efficiency_ratio: Standard residential (individual house) : 0.72–0.78 Group housing (corridor + lift core) : 0.55–0.65 Mixed use (ground commercial + upper res): 0.60–0.68

Sellable_Area_sqft = Sellable_Area_sqm × 10.764

2.2 Revenue Estimate

Formula — Revenue
Revenue_P50 = Sellable_Area_sqft × selling_rate_per_sqft

selling_rate_per_sqft is sourced from White Warp's locality rate table: - Primary source : PropEquity NCR transaction database (quarterly updated) - Secondary source: Anarock Research quarterly reports - Fallback : Magicbricks/99acres listed-price median × 0.88 (transaction discount) - Data vintage : shown on every report with quarter label (e.g., "Q4 FY25")

Revenue is further adjusted for absorption schedule: Phase 1 (0–6 months) : 30% of sellable area sold at launch premium +5% Phase 2 (6–12 months) : 40% of sellable area at base rate Phase 3 (12–18 months) : 30% of sellable area at −3% discount (unsold liquidation)
SECTION 03

Monte Carlo Simulation

3.1 Overview

The engine runs 10,000 simulations per report. Each simulation independently samples all uncertain variables, computes the full financial model, and records the outcome. The distribution of outcomes gives us the probability of profit, downside risk (P5), and upside (P95).

3.2 Variable Distributions

Sampled Variables — Lognormal Distribution
Variable | Mean (μ) | Std Dev (σ) | Source ------------------|--------------------|----------------|--------------------------- Selling Rate | locality_rate_P50 | 12–18% | PropEquity NCR std dev, 10yr Construction Cost | P50_hard_cost | 10–15% | MoSPI project variance data Timeline | base_months | 15–20% | RERA Delhi completion data Interest Rate | 14% | 1.5% | RBI MCLR 10yr band Marketing Cost | 2% of revenue | 0.5–1% | Developer market practice
Why lognormal? Real estate returns are not normally distributed. The NAREIT index shows kurtosis = 10.31 — a fat left tail (more extreme losses than a normal curve predicts). Lognormal distribution captures this asymmetry. Using normal distributions systematically underestimates downside risk.

3.3 Correlation Structure (Cholesky Decomposition)

5×5 Correlation Matrix
Sell Rate Const Cost Timeline Int Rate Marketing Selling Rate : 1.00 −0.20 −0.30 −0.40 0.10 Const Cost : −0.20 1.00 0.45 0.25 0.00 Timeline : −0.30 0.45 1.00 0.20 0.00 Interest Rate : −0.40 0.25 0.20 1.00 0.05 Marketing Cost: 0.10 0.00 0.00 0.05 1.00

Rationale: When interest rates rise, selling rate falls and timeline extends (buyers exit, absorption slows). When construction extends, cost rises. These correlations are sourced from Hoesli (2006) Indian RE adaptation. Cholesky decomposition enforces these correlations across all simulations.

3.4 Black Swan / Jump Process

Poisson Jump Process
Jump_probability = λ × dt = 0.25 × (timeline_years) λ = 0.25 events/year (1 shock every 4 years — calibrated to: Demonetisation 2016, GST disruption 2017, IL&FS crisis 2018, COVID lockdown 2020)

If jump occurs: Revenue impact : −25% to −35% (uniform draw) Timeline impact: +3 to +5 months (uniform draw) Probability : ~18–22% of simulations experience a jump over 18-month project

3.5 Outputs

Monte Carlo Outputs
P5_profit = 5th percentile of profit distribution (severe downside) P50_profit = 50th percentile (median — most likely outcome) P95_profit = 95th percentile (strong upside) prob_profit = fraction of 10,000 simulations where profit_margin > 15% IRR_P50 = median IRR across simulations IRR_P5 = 5th percentile IRR (downside scenario)
SECTION 04

IRR Calculation

Formula — Internal Rate of Return (Newton-Raphson)
IRR is the rate r such that: Σ [CFt / (1+r)^t] = 0 for t = 0 to T Where: CF0 = −Land_Cost (outflow at t=0) CF1..n = −Construction_drawdowns (monthly, S-curve weighted) CFT = +Gross_Revenue (phased over absorption schedule)

Solver: Newton-Raphson with 1000-iteration cap and ±0.01% convergence threshold. Cross-check: MIRR computed using reinvestment rate = 12% (conservative NCR market rate).

Hurdle Rate: 18% per annum Rationale: NCR developers expect 18–22% IRR on standalone plot redevelopment (source: CREDAI NCR member conversations + Anarock developer survey 2024). Projects below 15% IRR are flagged as "HOLD". Projects below 12% are "RECONSIDER".
SECTION 05

FAR / FSI Calculation

Floor Area Ratio (FAR) — also called Floor Space Index (FSI) — determines the maximum built-up area permitted on a plot. Rules are jurisdiction-specific and plot-size-dependent.

Jurisdiction Governing Code FAR Range Key Notes
Delhi MPD 2041 + MCD Bye-laws 1.20 – 3.50 Plot-size tiered. Transit Oriented Zone up to 3.5. Lal Dora: no FAR applies.
Gurugram HSVP + DTCP + MCG 1.25 – 2.50 Sector-specific. DLF sectors differ from HSVP sectors. Road-width premium.
Noida GNIDA Building Regulations 1.50 – 2.00 1.80 for ≤500 sqm; 1.50 for >500 sqm (residential, Group A).
YEIDA YEIDA Regulations 2022 1.50 – 2.00 Yamuna Expressway corridor. Special zones along expressway may differ.
Chennai TNCDBR 2019 + CMDA 1.50 – 3.25 HRB zones by road width (2.0/2.5/3.25). CBA hard cap: 12m/4 floors.
Formula — Maximum Built-Up Area
Max_BUA = Plot_Area_sqm × FAR_lookup(jurisdiction, plot_size, road_width, land_use)

Setback deduction: Net_Plot_Area = Plot_Area − setback_area(front, rear, side) per jurisdiction rules Effective_BUA = Net_Plot_Area × FAR (conservative: setbacks reduce buildable footprint)

Height limit is separately calculated from road width: Permitted_Height = road_width × height_ratio per jurisdiction + permissible_floors cap
SECTION 06

Accuracy Envelope — Honest Disclosure

These are the current accuracy bounds of the engine as of May 2026. No back-testing on live completed projects has been done yet. These bounds are based on meta-analysis of construction cost studies, developer surveys, and academic literature on Indian real estate.

Output Typical Accuracy Error Bound Bias Direction Source
Total Construction Cost ±15–25% P50 understated ~10–15% Optimistic MoSPI 2023, IIT Delhi
Sellable Area ±5–10% (clean plots) ±15–20% (edge cases) Slight optimism FAR rule precision
Total Revenue ±15–25% Over 18–24 month cycle Optimistic PropEquity rate variance
Estimated Profit ±35–55% Mean biased high ~10–20% Optimistic Compound of above
IRR ±400–700 bps Optimism bias: +100–250 bps Optimistic Hoesli 2006, Cornell 2022
Proceed / Don't verdict ~65–75% correct Worse on edge-case plots Optimistic bias Internal meta-analysis
What this means in practice: If the engine says "Total Cost: ₹2.0 Cr", the real cost is more likely to land between ₹1.6 Cr and ₹2.5 Cr — and is marginally more likely to be above ₹2.0 Cr than below it. The P85 figure (e.g. ₹2.5 Cr) is the number you should use for loan sizing and contingency planning. The Proceed / Don't verdict is a directional signal, not a guarantee. Always commission a full QS estimate before signing a purchase agreement.

6.1 Back-Testing Roadmap

White Warp is actively pursuing back-testing through:

Even if those numbers show meaningful error, we will publish them. Honest calibration data, published transparently, builds more trust than polished claims with no evidence.

SECTION 07

Data Sources & Vintage

Data Type Source Update Frequency Vintage Shown on Report
Construction base rates PWD Delhi DSR 2024–25 Annual (April) Yes — "DSR 2024-25"
Locality selling rates PropEquity + Anarock NCR Quarterly Yes — "Q4 FY25" etc.
FAR / zoning rules MPD 2041, HSVP, GNIDA, TNCDBR 2019 Tracked manually; updated on gazette notification Yes — jurisdiction code
Interest rates RBI MCLR (published monthly) Monthly Engine uses last available
Stamp duty / registration State revenue department schedules Per Finance Act amendment Jurisdiction-labeled
Climate / flood data IMD, CGWB, NDMA, SoilGrids, Overpass API Annual / real-time Date of fetch shown
Pre-flight GIS checks DDA GIS (Lal Dora), ASI buffer (100m/300m), Delhi Ridge zones Static + updated on DDA notification Layer version shown
SECTION 08

Limitations — What the Engine Cannot Do

⚠The engine cannot replace a licensed architect's structural or design assessment. FAR and area figures are indicative; actual permissible construction requires an architect's detailed drawings submitted to the relevant authority.
⚠Title risk is not assessed. The engine does not verify ownership, encumbrances, litigation, or chain of title. Always conduct independent title search via a registered advocate.
⚠Soil testing is not done. Foundation cost estimates assume standard Delhi NCR alluvial soil with bearing capacity 15–18 T/m². Expansive or soft soils will increase foundation costs significantly. Commission a BIS IS 1892 soil test before structural design.
⚠Lal Dora, Ridge, ASI buffer, PM-UDAY, and unauthorised colony flags are auto-detected via GIS coordinates, but boundaries are approximate (±50m accuracy). Always verify with the DDA or MCD for plots near boundary zones.
⚠Revenue rates are historical transaction prices. Future rates depend on macro conditions (interest rates, sentiment, policy) that the engine cannot predict. The Monte Carlo simulation models this uncertainty but does not eliminate it.
⚠The Proceed verdict is a pre-feasibility signal, not a professional opinion under the definition of the RICS Valuation Standards (Red Book). It should not be used as the sole basis for a financial commitment without engaging a qualified valuer or CA.
SECTION 09

Disclaimer

This report is generated by an automated engine for pre-feasibility screening purposes only. It does not constitute a valuation, investment advice, or professional certification. The inputs, assumptions, and outputs in this report are the sole responsibility of the requesting user. White Warp Technology Pvt. Ltd. and its founders, directors, and employees accept no liability for decisions made on the basis of this report without independent professional verification.

All formula descriptions above are accurate as of the engine version stated on the cover. Formula parameters may be updated as back-testing data improves model calibration. The updated methodology will be published at whitewarp.in/methodology.

WHITE WARP — Technology for Real Estate  ·  whitewarp.in Methodology v2.0 — May 2026  ·  Engine rev 00041+