TEI Paper: Ready for Submission

✅ What’s Complete

Paper Structure

  • 19 pages of LaTeX content
  • 10 sections (Introduction → Conclusion)
  • 4 appendices (Figures, Tables, Code Map, Reproducibility)
  • 10 embedded figures (publication-quality PNG)
  • 3 embedded tables (LaTeX format)
  • Abstract (200 words, complete)
  • Placeholder references (5 foundational citations)

Data & Analysis

  • 19,000 games collected (38 configurations)
  • Statistical analysis complete (ANOVA, correlations, effect sizes)
  • Key finding: Skill increases with fleet size (contradicts intuition)
  • W18 @ 18p: Highest skill/luck ratio (8.10) - NOT a crapshoot
  • All figures generated from empirical data

Files Created

docs/
  tei-paper.tex           # Main LaTeX source (complete)
  tei-paper.pdf           # Compiled output (4.1 MB, 19 pages)
  tei-paper.bib           # BibTeX template (needs completion)
  SUBMISSION-CHECKLIST.md # Step-by-step submission guide
  READY-FOR-SUBMISSION.md # This file
  
tools/nn/
  figures/                # 10 PNG figures
  tables/                 # 3 LaTeX tables
  data/                   # 19K games + processed data

⏳ What Needs Completion (30-60 minutes)

1. Author Information (5 minutes)

File: docs/tei-paper.tex (lines 14-25)

Current:

\author{
  Author Name\textsuperscript{1,2} \and
  Author Name\textsuperscript{1} \and
  Author Name\textsuperscript{3}
}

Replace with:

\author{
  Your Name\textsuperscript{1} \and
  Collaborator Name\textsuperscript{2}
}

\textsuperscript{1}Your Institution, Department \\
\textsuperscript{2}Their Institution \\
\texttt{\{your.email, their.email\}@domain.edu}

2. Acknowledgments (10 minutes)

File: docs/tei-paper.tex (line 633)

Current: Template with TODOs

Example:

This work was supported by NSF Grant #XXXXXX. We thank the 
open-source community for contributions to the Warp 12 engine.
Computational resources were provided by [institution]. Special 
thanks to [beta testers] for feedback on game mechanics.

3. Citations (1-2 hours)

File: docs/tei-paper.bib

Status: Template with 5 foundational papers (MCTS, AlphaZero, Elo)

TODO:

  • Search Google Scholar for:
    • “domino games AI” (likely sparse)
    • “multi-player game AI”
    • “self-play reinforcement learning”
    • “skill rating systems”
  • Add 10-20 relevant citations
  • Add \cite{...} commands throughout text where appropriate

Key places needing citations:

  • Section 2 (Related Work) - each paragraph
  • Section 4 (Agent Architecture) - MCTS/ISMCTS references
  • Section 5 (TEI) - Elo/Glickman/TrueSkill
  • Section 8 (Discussion) - any strong claims

🎯 Submission Strategy

Why AIIDE?

  • Best fit for game AI + empirical evaluation
  • Values practical AI systems with real deployment
  • Appreciates reproducible research (GitHub, data)
  • Page limit (8-10) matches our content scope
  • Community values product-oriented AI design

Timeline:

  1. Complete author/acknowledgments/citations (1 day)
  2. Proofread and polish (1 day)
  3. Format to AIIDE template (0.5 day)
  4. Internal review (1 week)
  5. Submit before AIIDE deadline (typically June)

Alternative Venues

IEEE CoG (Conference on Games):

  • Pros: Broader games audience, IEEE prestige
  • Cons: Requires IEEE template (2-column)
  • Timeline: Similar to AIIDE

arXiv Preprint:

  • Pros: Immediate publication, no review delay
  • Cons: Not peer-reviewed (supplement, not replacement)
  • Strategy: Post to arXiv, then submit to conference

FDG (Foundations of Digital Games):

  • Pros: Design-focused audience
  • Cons: May need to emphasize UX/design aspects more

📋 Quick Start: Complete in 1 Hour

Step 1: Author Info (5 min)

# Edit docs/tei-paper.tex lines 14-25
# Replace "Author Name" with actual names
# Add affiliations and emails

Step 2: Acknowledgments (10 min)

# Edit docs/tei-paper.tex line 633
# Replace TODO template with actual acknowledgments

Step 3: Key Citations (30 min)

# Add these specific citations throughout text:

# Section 2.2: MCTS
Add: "Monte Carlo Tree Search \cite{browne2012mcts} ..."

# Section 4: ISMCTS
Add: "Information Set MCTS \cite{cowling2012ismcts} ..."

# Section 5: Elo
Add: "Standard Elo rating \cite{elo1978rating} ..."

# Section 7: AlphaZero comparison
Add: "Unlike AlphaZero \cite{silver2017alphazero} ..."

Step 4: Compile (5 min)

cd /Volumes/Code/Warp12/docs
pdflatex tei-paper.tex
bibtex tei-paper
pdflatex tei-paper.tex
pdflatex tei-paper.tex

Step 5: Proofread (10 min)

  • Read abstract out loud
  • Check figure captions
  • Verify all sections flow logically

🚀 Commands Reference

Compile Paper

cd /Volumes/Code/Warp12/docs

# Full compile (3 passes)
pdflatex tei-paper.tex
bibtex tei-paper
pdflatex tei-paper.tex
pdflatex tei-paper.tex

# Or use latexmk (automatic)
latexmk -pdf tei-paper.tex

# Clean auxiliary files
latexmk -c

Check Paper Stats

# Word count (approximate)
pdftotext tei-paper.pdf - | wc -w

# Find TODOs
grep -n "TODO" tei-paper.tex

# Check missing citations
grep -n "\\cite{}" tei-paper.tex

Export for Submission

# Create submission package
mkdir tei-paper-submission
cp tei-paper.tex tei-paper-submission/
cp tei-paper.bib tei-paper-submission/
cp -r ../tools/nn/figures tei-paper-submission/
cp -r ../tools/nn/tables tei-paper-submission/
zip -r tei-paper-submission.zip tei-paper-submission/

📊 Paper Highlights (for Abstract/Summary)

Novel Contributions:

  1. 19,000-game empirical study across 4 Warp factors × 38 configurations
  2. Counter-intuitive finding: Skill expression INCREASES with fleet size (r=0.84-0.89)
  3. W18 @ 18p is NOT a crapshoot: 1.74× skill vs 2p, highest skill/luck ratio (8.10)
  4. W12 empirically optimal for rating: highest sensitivity to player count (r=0.875)
  5. Dual-track TEI: Independent points/go-out ratings with fixed AI reference bands
  6. Ω self-play policy: Replaces heuristic Commander without separate lobby tier
  7. Multi-engine routing: Expectimax for 2p, ISMCTS for 3p+ (mode-dependent search)

Key Results:

  • ANOVA: Warp factor effect F=817.14, p<0.001, η²=0.290 (large effect)
  • Fleet size: r=0.875 (W12), r=0.893 (W15), r=0.840 (W18) - all p<0.001
  • Decision complexity constant (~2 moves/turn) despite 4× tile count variation
  • Hand entropy increases W9→W18: 2.09 → 2.66 bits

✉️ Next Steps

  1. Today: Fill in author info and acknowledgments
  2. This week: Complete citations and proofread
  3. Next week: Format for target venue
  4. Submit: Before conference deadline

Questions? Check SUBMISSION-CHECKLIST.md for detailed guidance.

Ready to submit? You’re 1 hour of editing away from a complete paper! 🎉

Warp 12 — a Double-Twelve domino variant for the NX Epoch.