Skip to content

filippo1649/chess_framework

Repository files navigation

Chess Framework — AI Competitive Intelligence System

What if you could map a company's competitive position like a chess game? Where piece placement = strategic strength, and board position = market reality.

Python Data Analysis Visualization


Origin

I was building a dataset of AI companies — scoring them on funding, model count, risk exposure, market dominance, unique advantages, and ecosystem strength. Seven companies. Six dimensions. Straightforward spreadsheet work.

But then I realized something: these dimensions are exactly what matter in chess. Material (pieces/resources), Mobility (how many moves available), King Safety (exposure to threat), Center Control (dominance of key squares), Passed Pawns (unique advantages), Piece Coordination (ecosystem integration).

What if I mapped each company as a chess position, where their piece placement showed their actual strategic state? Not as a metaphor, but as a real analytic framework that happens to look like chess.

The result: a system that translates corporate strategy into visual, readable, and — surprisingly — transferable across industries.


What I Built

Dataset

7 AI companies scored across 6 strategic dimensions:

Company Source
OpenAI Crunchbase, company disclosures, news
Anthropic LMSYS Chatbot Arena, enterprise signals
Google DeepMind Public announcements, GitHub model repos
Meta AI Investor reports, open-source activity
Microsoft AI Azure integration data, partnership announcements
Mistral EU regulatory filings, research papers
xAI Funding announcements, X/Twitter integration

Dimensions (Raw Scores 0–10)

Dimension Definition Source
Material Total funding raised Crunchbase, company disclosures
Mobility Model portfolio + API integrations Company websites, HuggingFace
King Safety Regulatory risk + runway EU AI Act compliance, burn rate analysis
Center Control LMSYS ranking + enterprise adoption Chatbot Arena leaderboard, case studies
Passed Pawns Unique models with no direct competitor Patent analysis, market differentiation
Piece Coordination Ecosystem strength (cloud ties, tools, partners) Integration announcements, partnership depth

Analysis Pipeline

Raw Data Collection
    ↓
Normalization (logarithmic for funding, min-max for others)
    ↓
Weighting System A: Chess Perspective
  (mobility & center control = 0.25 each — tactical focus)
    ↓
Weighting System B: Business Perspective  
  (material & coordination = 0.25/0.20 — structural focus)
    ↓
Position Generation
  (score → piece placement on chess board)
    ↓
Interactive Visualization
  (dual boards + radar charts + breakdown table)

Key Findings

1. Different Perspectives, Different Winners

Ranking Chess Perspective Business Perspective
#1 OpenAI (8.86) Google DeepMind (8.75)
#2 Google DeepMind (8.80) OpenAI (8.50)
#3 Anthropic (5.65) Microsoft AI (6.99)

Insight: OpenAI dominates tactically (models, mobility) but Google is stronger structurally (ecosystem, safety). Anthropic has the best center control (LMSYS #1) but lacks ecosystem scale. Microsoft dominates integration but lags on frontier models.

2. Convergence & Divergence

Companies where chess ≈ business (balanced position):

  • Google DeepMind (+0.05 delta) — equally strong tactically and structurally

Companies where chess ≠ business (exposed position):

  • Microsoft AI (-2.13 delta) — strong ecosystem, weak models
  • xAI (-1.51 delta) — funded but underdeveloped across dimensions

Methodology Notes

Why chess, not just a table?

Three reasons:

  1. Visualization clarity: A piece's position on the board immediately shows advancement. Advanced = high score. Back rank = underdeveloped. No need to read numbers.

  2. Transferability: This framework works for any competitive landscape where you can define 6 strategic dimensions and map them to pieces. Automotive (Material = scale, Mobility = model lineup), Pharma (Material = R&D budget, Passed Pawns = unique drugs), Fintech (Material = capital, King Safety = regulatory).

  3. Intuitive narrative: Non-technical stakeholders understand chess. CEOs, board members, marketing teams all immediately grasp "our king is exposed" (regulatory risk) or "we're losing the center" (market control).

Weighting logic:

Chess perspective weights mobility and center control heavily — in chess, a mobile piece in the center is more valuable than raw material quantity.

Business perspective weights material and ecosystem coordination — in business, capital and structural partnerships often matter more than feature count.

This isn't arbitrary. It's a deliberate choice to show how the same data yields different conclusions depending on what you prioritize.


Deliverables

Interactive HTML Report

Open chess_[Company_A]_vs_[Company_B].html in a browser to see:

  • Dual chess boards (chess vs. business perspective)
  • Real-time advantage bars
  • Captured pieces (missing strengths)
  • Dimension-by-dimension breakdown
  • Strategic interpretation

Generated Positions (Examples)

OpenAI vs. Anthropic (Chess Perspective)

  • OpenAI: Dominant mobility (pieces scattered across board), strong material, exposed king (safety risk)
  • Anthropic: Consolidated center control, weak ecosystem, underdeveloped material

Interpretation: OpenAI is "attacking" but has exposure; Anthropic is defending the center but not expanding.


How to Replicate on Another Industry

  1. Define your 6 dimensions — What matters in this industry? For automotive: scale, model lineup, charging network, autonomous capability, brand safety, supplier integration.

  2. Score 5-7 competitors — Use public data: financial reports, market research, news, patent filings.

  3. Normalize each dimension — Keep everything 0–10 scale for consistency.

  4. Map to chess pieces — Material→Rooks, Mobility→Knights, Center Control→Bishops, Coordination→Queen, Safety→King, Unique Advantage→Pawns.

  5. Generate positions — Run the Python script (provided in repo) to auto-generate board positions.

  6. Weigh and compare — Use dual perspectives (tactical vs. structural) to reveal strategy divergence.


Code & Tools

Language: Python 3.14 Libraries: python-chess (board generation), pandas (data normalization), numpy (weighting calculations) Visualization: Interactive HTML with CSS styling, no external charting library needed Data Format: CSV input, HTML output


Limitations

Current limitations:

  • Scoring is partially qualitative (enterprise signals, unique models require judgment calls)
  • Not real-time — based on snapshot data from June 2026
  • Limited to 6 dimensions (could expand to 8-10 but readability decreases)

Attribution & Methodology

Data collected from: Crunchbase (funding), company disclosures (model counts), LMSYS Chatbot Arena (rankings), news sources (partnerships), EU regulatory filings (risk assessment), GitHub (open-source activity).

Normalization: Logarithmic for heavily-skewed dimensions (funding), min-max for bounded dimensions (scores 0–10).

Interpretation: Based on public market signals, not insider information.


Chess Framework by Filippo Leonetti Luparini • June 2026

About

Competitive intelligence framework mapping companies across strategic dimensions using chess positions

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors