PRO FORMA
METHOD NOTE

We publish the box.

Most simulations ask students to reverse-engineer a black box. Ours prints its own rules inside the game, and then proves, on every release, that no strategy owns the season. This page is the short version of how.

Open book, by construction

Every demand weight, price band, and scoring formula is printed inside the game itself, not summarized on a marketing page. The in-app Spec Book prints the engine's own live parameter values: the exact numbers running that season, not a paraphrase of them. The Analyst Brief and Intelligence screens go further and show the full factor-by-factor math behind a market-share number: which inputs moved it, by how much, and in which direction. The pro forma students consult before locking a decision is the actual engine running their whole year in advance. Nothing in the interface approximates the game. It is the game.

The demand model itself is published too: buyers respond to price against the cell average, hardest in the economy segment, gentler in premium, and a product whose competitive index beats the average earns price headroom before that penalty binds. A bare markup bleeds share; a justified premium holds. There is a best price inside every band, and finding it is the curriculum.

The fairness gate suite

Before any release ships, seven gates run over sixty seeded worlds. Plainly, they check: that the books balance to the cent, that five genuinely different strategies (a cost disruptor, a premium builder, a commercial specialist, a sustainability leader, a fast follower) each win a fair share of seasons and none dominates, that scripted degenerate plays all fail, that a single decision moves outcomes by a meaningful but bounded amount, that a badly-run company struggles without being guaranteed to die, that a team in last place after round three can still finish in the top half, and that replays are exact. This is the real harness output from today's release check, unedited:

Running 7 calibration gates over 60 seeds...[PASS] gate 1 ledger identity: 50 seeds, exact to the cent [PASS] gate 2 balance:disruptor: win rate 11.7% (need 10-35%) [PASS] gate 2 balance:premium_optimizer: win rate 25.0% (need 10-35%) [PASS] gate 2 balance:commercial_specialist: win rate 10.0% (need 10-35%) [PASS] gate 2 balance:esg_leader: win rate 25.0% (need 10-35%) [PASS] gate 2 balance:fast_follower: win rate 28.3% (need 10-35%) [PASS] gate 3 exploit:exploit_price_floor: top-2 rate 0.0% (need <20%) [PASS] gate 3 exploit:exploit_cash_hoard: top-2 rate 0.0% (need <20%) [PASS] gate 3 exploit:exploit_monopolist: top-2 rate 0.0% (need <20%) [PASS] gate 3 exploit:exploit_leverage: top-2 rate 0.0% (need <20%) [PASS] gate 3 exploit:exploit_stack: top-2 rate 0.0% (need <20%, ratchet only) [PASS] gate 4 sensitivity: +1 idx -> 6.66% rel share (need 2.5-8.5) [PASS] gate 5 solvency texture: naive emergency-line rate 10.0% (need 7-22%) [PASS] gate 6 comeback: last-at-round-3 in top-half of final scores 85.0% [PASS] gate 7 determinism: 3 seeds byte-identical (results + state) ALL GATES GREEN (13.1s)

The red team and the lever audit

Balance gates only catch the strategies we already thought to script. To catch the ones we didn't, an adversarial AI red team plays degenerate strategies against every release: price-floor dumping, cash hoarding, leverage abuse, and stacked residual exploits, among others. Any strategy that reaches the top two more than 20% of the time blocks the release until the underlying economics are repriced. A separate lever audit sweeps every player decision across its full range on each release, to prove each one is a genuine tradeoff rather than a dead knob nobody needs to touch or a no-brainer with one obviously correct setting. Both runs are real and merge-blocking. They are not marketing.

Written from the real decade

The world events are hand-authored from documented industry history, not generated. Before the Motors edition shipped its realism options, we ran a research pass across the 2015 to 2026 EV industry: demand curves, battery costs, policy shocks, the price war, the labor settlements, and compared every finding against the engine's numbers. Where reality disagreed with us, we changed the game; where our simplifications were deliberate, we wrote down why. The same discipline built the airline edition.

Determinism

Same seed, same world: replays come back byte-identical, results and state alike. That is what makes every classroom round exactly reproducible, grades defensible, and post-season debriefs honest. There is no hidden randomness to blame.

See it for yourself · the 15-minute demo