From many possible futures to one you can defend.
A data-driven process of about eight weeks. We generate optimized futures for the site, inspect each one in depth and side by side, and resolve the decision by stress-testing every path against conditions no one controls, until one earns the recommendation.
Before the site is committed
The Sprint is built for the moment before a site's future is fixed: when the change is large enough that its effects need forecasting, the options can't yet be compared, and the trade-offs are still hidden.
Scenarios as instruments
Plans are usually treated as end products: one image of the future, fixed early and defended afterwards. The Sprint treats scenarios as working instruments. Each is a temporary version of the future, built from a different position, so that comparing and testing them reveals the structure of the decision: what every future shares, and the few choices that truly divide them.
That structure is the real product. It stays useful as conditions change, which a single image of the future cannot.
How it Works
We establish what must be decided, what is fixed, and chart the tensions that make the choice hard. We agree on how every future will be measured. You receive the Frame report delivered before anything else. You approve it before scenarios are built.
Three scenarios, each from a different end of the site's tensions, generated from data and optimized within their own logic. You receive each scenario's design variants, the trade-offs between them, and a 3D model of the selected version.
Each scenario is analyzed in depth: its form, how it works as a place, and whether it can be delivered. You receive a profile of each, with every figure sourced.
The three scenarios on equal terms, in the same measures, at the same visual quality. You receive a side-by-side view of what each gains and gives up.
What every scenario shares, the few choices that divide them, and where each path breaks under conditions you don't control. You receive the branching diagram: what to commit now, what to hold open, and for how long.
A recommended strategy, with the conditions it depends on and the signal that would change it. You receive the recommendation, illustrated with a regulating plan and renders of the recommended future.
A glimpse into a sprint
A fictional site, followed through all six steps.
Step 5 is live: change a condition and watch the recommendation respond.
An illustrative example of the Stress-test step. A fictional waterfront site with arbitrary figures.
Stress-test Example
Most of what this site needs is shared by every scenario. Only one choice diverges.
Move the housing sale price and see which path holds, and when the answer changes.
The horizontal axis is cumulative commitment, not calendar time. Branch width shows the share of the program travelling each path. The shaded field is the deferral window: every path stays reachable until the choice closes.
Commit now
- Street grid and utilities;
- Waterfront promenade;
- School plot reservation;
- Metro stop.
The fork
Phase 2 blocks: housing-led, employment-led or green-first?
- Holds while
- Signal to watch
- Decide by
- Phase 2 land allocation, estimated Q3 2028
Before we start
-
Political and stakeholder judgement stay outside the analysis and re-enter in the recommendation, labeled as judgement.
-
Sometimes the shared core is small or the choice stays wide. That is a finding, and we report it as one.
-
Detailed design of the recommended future, and any detailed statutory document production, are separate steps.
-
Our founder’s work Dynamic Futures, tested computational scenario generation on an Urban Renewal site at 8th and Market Street, Philadelphia, as an entry to the 2025 Edmund N. Bacon Urban Design Awards. The method builds on what is showed and on what it didn’t.
-
About eight weeks from approved framing to final delivery.
-
Your challenge, the site, the decision you’re facing, existing plans and data, and one contact person.
-
Scoped per site after a first conversation.
-
A presentation-style report, as a PDF and an interactive version with one chapter per step. It closes with a session where we walk through the recommendation together.
-
The strategy is built to be revisited. When a signal fires, we update it. When you commit, we can develop the recommended future further.
