Algorithmically generated perfect‑score city (bigscore1) shows minimal roads, no fire station and a downtown bonus.
*Jeff Quast’s Micropolis experiments expose how road‑free layouts, zero emergency services and negative unemployment boost scores, while population‑maximizing designs sacrifice points for sheer size.*
Jeff Quast, a software developer, spent months probing the open‑source MicropolisCore port of SimCity (1989). Using Python shims from gym‑city, he generated ten algorithmic cities: five that chase the elusive 900‑plus score and five that chase the 500k‑plus population milestone. The results expose a stark design dichotomy. Perfect‑score cities hover around 15k residents, run on a 1% tax, and strip away roads, rails, fire and police services. Megacity‑oriented designs swell to 700k inhabitants but surrender hundreds of points to crime, pollution and higher taxes. The experiment quantifies every factor the original manual left vague.
The five top‑scoring cities cluster around 12‑16k residents, a 1% tax rate and almost no infrastructure. bigscore1 hits 949.5 points with 16,560 people, 13 crime incidents and 36 pollution units. Roads and rails are sparse, often non‑contiguous, and fire or police departments are absent. By contrast, the five largest‑population cities exceed 670k residents but linger near 460‑525 points. bigpop1 reaches 709,405 citizens at a 2% tax rate, but its crime climbs to 140 and pollution to 174. The score gap stems from the seven penalty terms that punish crime, pollution, land value, taxes, traffic and fire, while rewarding negative unemployment. High‑pop designs accept those penalties to hit the Megapolis reward threshold.
Road tiles generate traffic and pollute residential zones, directly inflating the traffic penalty (max 576) and the pollution term. Rails emit no pollution and register no traffic, yet they satisfy the 30‑step delivery rule for commercial zones. The engine grants a downtown bonus of +64 that fades to -64 at map edges, so compact rail loops keep commercial growth without the cost of roads. Fire departments and police stations are among the largest budget items; funding them at 100% inflates annual costs, forcing a higher tax rate that shaves roughly 30 points off the score. Eliminating both services lets the perfect‑score cities run on a 1% tax, the lowest possible rate.
Unemployment is the only unbounded term. It is calculated as (resPop / (8 × jobs) - 1) × 255. When jobs outnumber residents, the result turns negative, adding up to about 128 points to the total. bigscore1 shows unemployment at -93, contributing a sizable boost. In the population‑focused megacities, unemployment is positive (e.g., bigpop2 at 175), offering no bonus. The formula explains why perfect‑score layouts deliberately over‑staff jobs: the negative unemployment offset outweighs any minor increase in other penalties.
The Super Nintendo edition is the only version that rewards megacity size with gifts that raise land value. MicropolisCore reproduces the original ruleset, and algorithmic runs reach 709,405 residents (bigpop1) with a score of 467.0. However, the engine caps practical population below the 911.2k reported by a SNES player, likely due to map size and performance limits. Even at 700k, the city endures high crime (140‑219) and pollution (171‑179), and must raise the tax rate to 2‑5% to cover costs, which further depresses the score. The data confirms that while the engine can host massive populations, it does so at a steep scoring penalty.
A 1% tax rate produces the highest scores; raising the rate to 2% drops the score by about 30 points.
Fire and police are large budget items; omitting them lowers annual costs, allowing the city to stay at a 1% tax rate and avoid score penalties.
The algorithm reaches 709,405 residents, but the city endures high crime, pollution and must raise taxes to 2‑5%, which significantly reduces the overall score.
Quast’s data proves that SimCity’s scoring engine rewards austerity over expansion. By exploiting rail loops, negative unemployment and the absence of emergency services, a city can edge close to the theoretical maximum of 1000 points. Players chasing the Megapolis badge must accept a lower score, higher taxes and a messier map. The findings lay bare the hidden calculus behind the classic simulation, turning what was once guesswork into a reproducible playbook.
Sources: Hacker News – A Perfect SimCity