Payback period: what it means and why it matters

Median over the buildings in the group. Medians DO NOT COMBINE — never average these across groups; re-derive from the histograms instead. Payback ignoring the time value of money. Baseline: meritorder.

Solar + battery, merit-order sized: Same engine as sc2, but the target is what the building CONSUMES today x 1.5 (MO_SIZING_UPLIFT) rather than full latent demand.

In Osun, sizing sc2 to latent demand while crediting it only with displacing current spend produced a 6x-oversized system and a negative IRR that said nothing about solar. sc3 is the fix, not an alternative opinion.

Counterfactual: solar displaces the most expensive kWh first — generator before grid. savings = min(solar, genset_kWh) x genset_cost + max(0, solar - genset_kWh) x tariff. Approach B; the only baseline that is coherent for Bands B-E, where grid energy is rationed rather than purchasable.

Median over the buildings in the group. Medians DO NOT COMBINE — never average these across groups; re-derive from the histograms instead.

Median over the buildings in the group. Medians DO NOT COMBINE — never average these across groups; re-derive from the histograms instead. Payback ignoring the time value of money. Baseline: mixed.

Solar + battery: PV sized to cover FULL demand including round-trip losses: kwp = min(demand x (daylight + (1-daylight)/0.90) x 1.10 / net_yield, pv_max_kwp); battery = night demand / (0.80 DoD x sqrt(0.90)), capped at 2.0x PV kWp

Sized to LATENT demand — the energy the building would use if it could. Against a cash baseline that only reflects today's spend this looks oversized; that is what sc3 exists to fix.

Counterfactual: grid when available, generator otherwise, weighted by the NTL-derived grid availability. The realistic case, and what the headline numbers use. Always falls between grid and genset by construction.

Median over the buildings in the group. Medians DO NOT COMBINE — never average these across groups; re-derive from the histograms instead.