Open Games Plurigrid Skill
World-to-world interactions via compositional game theory with GF(3) energy consistency
Trit: 0 (ERGODIC)
Color: #26D826 (Green)
Synergy: 0.59 (plurigrid ↔ jules-hedges)
Overview
This skill unifies:
- open-games: Play/CoPlay bidirectional optics
- ramanujan-expander: Spectral gap λ₂ ≤ 2√(d-1)
- glass-bead-game: Triangle inequality d(W₁,W₃) ≤ d(W₁,W₂) + d(W₂,W₃)
Core Insight
Play-Co-Play interactions are energy-consistent open games:
┌───────────────┐
X ──→│ │──→ Y
│ δ-Key Update │
R ←──│ │←── S
└───────────────┘
Play (+1): forward pass, propose action, derive delta
CoPlay (-1): backward pass, adjust world, verify proof
Bridge (0): equilibrium, combine deltas, coordinate
GF(3): (+1) + (-1) + (0) = 0 ✓ (energy conserved)
Discovered Ecosystem
Core Contributors (Git Log)
| Name | Contributions | Trit | Role |
|---|---|---|---|
| philipp-zahn | 216 | +1 | Maintainer |
| andrevidela | 76 | 0 | Contributor |
| jules-hedges | 40 | -1 | Founder |
| dxo | 15 | 0 | hevm integration |
Repositories
| Repo | Stars | Description |
|---|---|---|
| CyberCat-Institute/open-game-engine | 183 | Main implementation |
| jules-hedges/open-games | 13 | Original Haskell |
| plurigrid/open-games-hs | 0 | Plurigrid fork |
| LISA-ITMO/CGT4NN | 5 | Neural network games |
Key Bridge Person
clayrat (Alex Gryzlov) @ IMDEA Software
- Stars BOTH open-games AND Catlab.jl
- Repos: dialectica, coherence-spaces, comonads
- The cobordism between compositional game theory ↔ applied category theory
bgavran (Bruno Gavranović) - Categorical Deep Learning
- 1459 stars on Category_Theory_Machine_Learning list
- Lens_Resources (56 stars) - optics theory
- Compositional_Deep_Learning (151 stars)
World Hopping
Distance Metric
def world_distance(w1, w2, dr):
# Hamming distance on fingerprints (being)
being_diff = bin(fp1 ^ fp2).count('1') / 64.0
# Epoch difference (temporal)
epoch_diff = abs(w1.epoch - w2.epoch) / 100.0
# Domain crossing
domain_diff = 0.0 if w1.domain == w2.domain else 0.3
return sqrt(being_diff² + epoch_diff² + domain_diff²)
Triangle Inequality
For any hop W₁ → W₃ via W₂:
d(W₁, W₃) ≤ d(W₁, W₂) + d(W₂, W₃)
Elegance = direct_distance / path_distance (higher = better)
Spectral Analysis
Ramanujan Bound
For d-regular expander graph:
λ₂ ≤ 2√(d-1)
Observed: λ₁ = 0.053 after growing, bound = 2.14
Growing the Expander
Add edges that maximize spectral gap:
- Prefer cross-cluster connections
- Maintain GF(3) balance
- Respect domain boundaries
Files
open_games_expander.py- Graph discoveryplay_coplay_protocol.py- Delta update protocolplurigrid_world_hop.py- World hopping enginedelta_key_update.py- Original delta protocolgay_loot.py- Synergy-based loot tables
GF(3) Triads
temporal-coalgebra (-1) ⊗ open-games (0) ⊗ free-monad-gen (+1) = 0 ✓
sheaf-cohomology (-1) ⊗ open-games (0) ⊗ topos-generate (+1) = 0 ✓
ramanujan-expander (-1) ⊗ glass-bead-game (0) ⊗ world-hopping (+1) = 0 ✓
Commands
# Run world hopping
python3 plurigrid_world_hop.py
# Play-CoPlay protocol
python3 play_coplay_protocol.py
# Expander graph discovery
python3 open_games_expander.py
# Delta key update demo
python3 delta_key_update.py
Related Skills
open-games- Core game theoryramanujan-expander- Spectral boundsglass-bead-game- World hoppingdelta-key-update- Key rotation protocolgay-mcp- Color fingerprintsbisimulation-game- Observational equivalence
Para(Optic) atlas
Part of: para-mensch-commons.