Why going up beats going down on this ground
Homeowners short of space on a constrained plot naturally consider both directions, and on most of this side of the borough the comparison resolves decisively upward. Consider what a basement actually involves on low-lying ground near a river valley. The excavation is substantial and the spoil has to leave the site through the house on a terrace. The existing house has to be underpinned in sequence, which is skilled, slow and the phase where structural risk concentrates. Groundwater is likely to be encountered at a shallower depth than elsewhere, which turns waterproofing from a specification into an engineering problem: a fully tanked or drained-cavity system with a sump and pumps, and a pumped system needs power, maintenance and a failure plan. Flood risk bears directly on habitable accommodation below ground, which can require a site-specific flood risk assessment and can attract the sequential approach to siting vulnerable uses. And the finished space is the least pleasant in the house to be in, lit by rooflights or a lightwell, and the most expensive per square metre by a wide margin. Now consider the roof. The volume already exists and is currently storing boxes. On most addresses here it is permitted development, so the planning risk is close to zero and a certificate converts that into a document. There is no excavation, no underpinning, no groundwater and no pump. The resulting rooms have the best daylight in the house and, on a terrace, a view. The work is largely contained at the top of the building, so the family can usually remain in occupation. The honest caveats are that the roof volume is finite and cannot be increased, that the staircase and the fire strategy are real costs, and that some roofs simply do not have the head height. But where the roof does have the height, it is not a close comparison — and on ground where digging raises water questions, it is not close at all.