▴ Aileron Games — Rulebook & Design Document · formerly HIGGS

GOD DOES NOT PLAY DICE

A Standard-Model card duel · 2 players · ~15–25 min

Einstein insisted that God does not play dice. Quantum mechanics disagreed — and won. In this game every collision of quarks and leptons rolls dice over real Feynman-vertex outcomes… unless you spend hidden variables to overrule probability and choose. The title is the mechanic: you are Einstein, clawing determinism back from the quantum, one collision at a time.

⚄ Play now →

Contents

1. Poking holes in the brief
2. The rules
3. The deck (60 fermions)
4. The collision table
5. Gauge-boson effects
6. Physics fidelity notes
7. Balance verification

1. Poking Holes in the Brief (and the fixes)

The original design brief was strong — fermion cards, boson effects, probabilistic Feynman collisions, a mahjong-style reference wall. Playtesting-by-simulation and a physics audit found these holes. Each one shaped the final rules:

Hole #1 — The force mapping was wrong

The brief assigned gluon→gravity and Z→strong force. In the Standard Model the gluon IS the strong force, the Z is the weak neutral current, and gravity isn't in the Standard Model at all — there is no graviton in the theory.

Fix: gluon = strong (confinement steal), photon = EM (repulsion), W = weak charged (flavor exchange), Z = weak neutral (measurement). Gravity's absence is now a joke on the forbidden-vertex wall.

Hole #2 — All-dice collisions remove agency (ironic, given the title)

If every outcome is a straight probability roll, the game is a slot machine. Nobody outplays anyone.

Fix: hidden variables — spend one before a roll to choose the outcome. Z-boson "measurements" and famous discoveries (J/ψ, Υ, Ω⁻) generate them. Einstein's argument became the resource economy.

Hole #3 — Whiff turns

With one action per turn, a collision that can simply fail means whole turns evaporate. Feels terrible.

Fix: legal collisions always produce something. The dice decide what, never whether. (Also better physics: the S-matrix always has an outcome.)

Hole #4 — Two goals fighting

"Collect points en route to a Higgs" splits the player's objective: is this a race or an engine game?

Fix: one currency. Hadrons score points. First to 20 wins. The Higgs is a +20 finisher from gluon fusion (30% roll — or guaranteed for 3 hidden variables: Einstein's gambit). Simulation: ~92% of games end at the target, so the Higgs stays a rare thunderclap rather than the only plan.

Hole #5 — Draw is strictly worse than play

One card per draw against one collision per turn means nobody draws until forced, then the game crawls.

Fix: Draw 2 (hand cap 8).

Hole #6 — The W effect had unstable polarity

"Give one of your cards to your opponent" is a bonus in any economy where shedding cards helps, and a penalty otherwise — same button, opposite meanings.

Fix: W = swap (you choose what to give, chance decides what you get) — which is also what the W boson actually does: charged-current flavor exchange.

Hole #7 — Undefined economy = undefined ending

The brief never said how hands refill, where discards come from, or when the game ends.

Fix: scored hadrons remove cards from the game, so the deck provably runs out. Deck-out gives each player one last turn, then high score wins. Annihilation debris feeds the discard (Salvage) pile.

Hole #8 — Hand-lock deadlock (found by simulation)

In 36% of 2,000 simulated games, both players hit the hand cap holding cards with no legal collision — and the rules offered no state-changing action. Infinite stall.

Fix: Jettison (beam-dump a card to the debris) as a pressure-valve action. Re-simulation: 0 stalls in 2,000 games.

Hole #9 — "All fermions" vs. "each card is a quark"

The brief said both. Leptons with no role are dead cards.

Fix: quarks and charged leptons. Lepton pairs are pure boson factories (e⁻e⁺ → γ/Z). And e + μ is a forbidden vertex — lepton-flavor conservation, taught by refusal.

Hole #10 — Coin toss + dice stacked the luck

Simulation showed the second seat winning 55%: the shared collider favors the reactive player.

Fix: the toss winner opens the beam and pockets a hidden variable. Re-simulation: 50.4% first-seat winrate.

Hole #11 — The top quark can't be a card

The top decays in 5×10⁻²⁵ s — before it can hadronize. A top card that behaved like other quarks would be wrong.

Fix: no top card. Its absence is a forbidden-wall tile, which is itself the lesson.

Hole #12 — Flavor-changing neutral currents (caught mid-build)

A first draft let neutral cross-flavor pairs (c + u̅) annihilate to a Z. The Standard Model forbids that — no FCNCs at tree level (the GIM mechanism).

Fix: neutral cross-flavor pairs always hadronize (100% meson); only charged pairs can radiate a W. The prohibition is on the wall.

2. The Rules

Setup

Shuffle the 60-card fermion deck. Each player draws 6.
Flip one card face-up into the shared collider (the beam re-seeds itself whenever it empties).
Coin toss: the winner opens the beam and takes 1 hidden variable.

Your turn — exactly one action

Collide — play one card from your hand onto any object in the collider. Look the pair up on the Wall of Diagrams. If it's a legal vertex, resolve one of its outcomes. Roll the dice, or spend 1 hidden variable to choose. Hidden variables come from Z-boson measurements, famous discoveries (J/ψ, Υ, Ω⁻), the coin toss — and the underdog measurement: every opposing score that opens a lead of 8+ grants the trailing player one.
Draw 2 — hand cap 8.
Salvage — take the top card of the debris pile.
Fuse — if 2 gluons sit in the collider: 30% Higgs (+20, ends the run) / 70% c c̅ pair production to your hand. Or spend 3 hidden variables for a guaranteed Higgs.
Jettison — dump one hand card into the debris.

The shared collider

Everything in the collider belongs to no one. A diquark you park there can be completed by your opponent for their points — watch what you leave on the table. Gluons produced by annihilation also park there, as fusion fuel for whoever moves first.

Winning

20 points of scored hadrons wins on the spot.
A Higgs is +20 and ends the run immediately (highest total wins).
Deck out: one final turn each, then highest total wins.

3. The Deck — 60 Fermions

Card counts are generation-weighted, like the universe: first-generation matter is everywhere, the heavy flavors are rare and precious. Every quark card carries a QCD color charge (assigned at shuffle — which quark is "red" is a gauge choice, so randomizing it is physically honest); antiquarks carry anti-colors.

Card	Count	Charge	Mass	Gen	Role
u / u̅	8 / 6	+⅔ / −⅔	2.2 MeV	I	bulk matter — gluons, pions, protons
d / d̅	8 / 6	−⅓ / +⅓	4.7 MeV	I	bulk matter
s / s̅	4 / 3	−⅓ / +⅓	95 MeV	II	kaons, hyperons, the Ω⁻ hunt
c / c̅	4 / 3	+⅔ / −⅔	1.27 GeV	II	D mesons, J/ψ
b / b̅	3 / 3	−⅓ / +⅓	4.18 GeV	III	B mesons, Υ — the premium flavors
e⁻ / e⁺	4 / 4	−1 / +1	511 keV	I	boson factory (annihilation)
μ⁻ / μ⁺	2 / 2	−1 / +1	106 MeV	II	boson factory

No top quark (decays before hadronizing), no neutrinos (nearly non-interacting — they'd be blank cards), no graviton (not in the Standard Model).

4. The Collision Table (the Wall of Diagrams)

Annihilation — same flavor, boson factory

Collision	Outcomes
u/d/s + own antiquark	g 55% · γ 25% · Z 20%
c + c̅	g 35% · γ 15% · Z 15% · J/ψ 35% (4 pts, +1 hv)
b + b̅	g 30% · γ 15% · Z 15% · Υ 40% (5 pts, +1 hv)
e⁻+e⁺, μ⁻+μ⁺	γ 55% · Z 45%

Hadronization — different flavors, mesons

Collision	Outcomes	Points
u + d̅ (charged pairs)	meson 75% · W^± 25%	π 2 · K 3 · D 4 · B 5 · Bᴄ 6
d + s̅ (neutral pairs)	meson 100% — no FCNC	same scale

Baryon assembly — the strong force always binds

Step	Result
quark + quark	diquark — parks in the shared collider, anyone may finish it
diquark + quark	baryon, scored: p/n 6 · Σ/Λ 7 · Ξ 8 · Ω⁻ 10 (+1 hv) · heavy exotics up to 12
3 antiquarks	antibaryon — same table

Gluon fusion — the endgame

Action	Outcomes
g + g (Fuse)	H 30% (+20, run ends) · c c̅ pair production 70%
Einstein's gambit	3 hidden variables: no roll, take the Higgs

Forbidden vertices (real conservation laws, taught by refusal)

e + μ — lepton-flavor conservation.
q + q̅′ → Z — no flavor-changing neutral currents (GIM mechanism).
q̅ + diquark — color can't close. Baryons need three matching quarks.
t — not in the deck. Decays in 5×10⁻²⁵ s, before hadronizing.
graviton — gravity is not in the Standard Model.

5. Gauge-Boson Effects

Boson	Force	Effect	Why it fits
g gluon	strong	take a random opponent card. The gluon parks in the collider as fusion fuel	confinement binds and drags
γ photon	electromagnetic	a random opponent card drops to the bottom of the deck	like-charge repulsion
W^±	weak, charged	swap a card you choose for a random opponent card	the W literally exchanges flavor & charge
Z	weak, neutral	+1 hidden variable, glimpse the opponent's hand	a neutral current is a measurement — information without exchange

6. Physics Fidelity Notes

What is real, and where the game compresses:

Real: all quark/lepton charges, masses and generations. Which vertices exist and which are forbidden (lepton flavor, FCNC/GIM, color closure, top non-hadronization, no graviton); coupling ordering (strong > EM > weak) reflected in the annihilation odds; W = flavor exchange; Feynman drawing conventions in the animations (solid fermion lines with arrows, reversed for antiparticles; wavy γ/W/Z; curled gluons; dashed Higgs).
Compressed: a "gluon" outcome stands in for the 2–3-gluon burst (single-gluon annihilation is color-forbidden), and "γ" for γγ, for the same reason; hadronization probabilities are gameplay numbers, not measured branching ratios; gluon-fusion Higgs production really proceeds through a top-quark loop, abstracted here to the fuse action.
Sources: the design brief pointed at Wikipedia's List of Feynman diagrams, Quark, and Mathematical formulation of the Standard Model; the build environment could not fetch the pages, so all values come from standard references (PDG values, textbook Standard Model) and are worth a proofread against them.

7. Balance Verification

2,000 AI-vs-AI games through the shipped engine: 0 stalls, 0 crashes · first seat wins ~50% · average game ~12 turns per player (longest ~45) · endings: ~92% reach the 20-point target, ~7% deck out, ~1% end in a Higgs detection. Live playtesting added the underdog measurement (a hidden variable to the trailer whenever a scoring play opens a lead of 8+) to soften snowballs. Tuning levers: the 20-point target (15 for a faster game), the fuse odds, and hidden-variable income.

⚄ Play the game →