✦ 100 BILLION CARDSPlay now

Design & mathematics

How the simulation works

The game builds a population of a hundred billion possible identities through deterministic generation. A separate probability system governs which people appear in packs.

A fixed identity space

The model assigns a numerical identifier to each of 100,000,000,000 possible people. The ID is fed into seeded pseudo-random functions so that the same identity can be reconstructed on different visits. The game stores collected identity keys and age information rather than maintaining a database of all possible people.

Birth periods are represented by cumulative population anchor tables. A lookup associates an ID with a birth period and a month within that period. The game then uses region, era and cultural profile data to generate names and backgrounds. The population tables approximate the distribution of people across historical periods. A hundred billion is the size of the game’s simulated world.

Age and context

Encountering the same person at different ages allows the life stage, occupation and wish to differ. Life histories and social contexts are approximated from the game’s internal models; for example, naming conventions vary across cultural profiles and the occupations available to a child differ from those of an adult.

Known historical people use curated identity records and special overrides, while fictional characters are produced from general generation rules. The model excludes living people from its curated historical roster and includes only figures whose lives have ended. The game assigns attributes and wishes as part of its character design.

How overall scores are calculated

Each attribute or talent has a modeled rank. Smaller ranks indicate a stronger modeled position. A simple average of displayed scores would lose the distinction between an ordinary percentile and an unusually rare tail of the distribution.

In the game, the core combination uses a geometric mean of ranks: it sums their logarithms, divides by the number of included ranks, and exponentiates the result. The inputs include attribute ranks, qualifying talent ranks and, for ordinary people, a modeled lifespan rank. Historical figures have special scoring rules and exclude lifespan from that combined rank calculation.

Illustrative example: ranks of 100 and 10,000 combine geometrically to 1,000. The example shows the rank calculation before the score conversion. The resulting value is mapped through the game's score calibration and rarity thresholds.

The game uses calibration tables for ordinary scores below R and a logarithmic scale for high scores. Above the R threshold, each order of magnitude in the underlying ranking scale corresponds to ten score points. Tier thresholds are N = 0, R = 100, SR = 120, SSR = 140 and UR = 160. These thresholds determine a card’s rarity in the collection.

Pack probabilities

Sampling a person from a simulated population and choosing a rarity for a pack are distinct operations. The game uses rarity weights and pack filters to decide which kinds of cards can appear. Ordinary-pack weights depend on player level, while regional or name-themed packs use additional eligibility rules. Historical figures have curated weighting rules.

At level 1 the ordinary tier weights are 40,000 : 10,000 : 400 : 20 : 1 for N through UR. At level 100 they are 1,000 : 500 : 100 : 10 : 1. Divide each relative weight by the sum of all five weights to obtain its probability. Beyond level 100, high-tier probabilities gradually approach their long-term targets.

TierLevel 1 ordinary weightsLevel 100 ordinary weights
N40,0001,000
R10,000500
SR400100
SSR2010
UR11

Rare packs exclude N; gold packs use the SSR tier. Check the game’s probability display for the actual pack and level, as a particular pack may not follow these ordinary weights.

Codes, reproducibility and storage

A share code encodes a card’s numerical identity, age and wish selection with an integrity check. When a valid code is imported, the browser reconstructs the card using the same generation rules. Sharing creates a copy of the card in the recipient’s collection. Each player keeps their own save.

The browser stores the collection, inventory, level, settings and claim history in local storage. This keeps the game self-contained but makes local data dependent on that browser profile. The save and its integrity checks run locally in the browser.

Historical context

The game uses simplified demographic and cultural models to make a wide range of eras and places available for exploration. For documented historical lives, the historical archive offers biographies and links to academic and museum resources.

Read the player guide · Play the game