Published

Chess Multiverse Error Explorer

Sparsh Varshney
Sparsh Varshney
0

The definitive research platform for analyzing human decision errors, time pressure blunders, and opening complexities across elite chess.

Chess Multiverse Error Explorer

DOI Architecture DuckDB Chessboard.js Routing

A browser-native research platform for large-scale analysis of human chess errors, cognitive collapse under time pressure, opening complexity, and behavioral decision-making patterns.

The Chess Multiverse Error Explorer is built on top of the Chess Multiverse Error & Evaluation Dataset (CMEED v1.0) and utilizes DuckDB WASM to perform server-grade analytical queries directly inside the user's browser without requiring any backend infrastructure.

Unlike traditional chess databases that focus primarily on move quality, the Error Explorer focuses on why strong players fail, enabling reproducible research into human error generation across elite chess.

๐Ÿš€ Live Application

Launch the production application:

Live URL

https://www.chessmultiverse.org/p/chess-multiverse-error-explorer.html

๐ŸŽฏ Research Objectives

The platform is designed to support research into:

  • Human decision errors
  • Cognitive degradation under pressure
  • Time-management failures
  • Opening complexity analysis
  • Evaluation-loss modeling
  • Player vulnerability profiling
  • Tournament pressure studies
  • Behavioral chess analytics
  • Reproducible computational research

๐Ÿ— Architecture

The application follows a fully client-side analytical architecture.

Browser
โ”‚
โ”œโ”€โ”€ DuckDB WASM
โ”œโ”€โ”€ Web Worker Thread
โ”œโ”€โ”€ CMEED Parquet Dataset
โ”œโ”€โ”€ Chess.js Engine
โ”œโ”€โ”€ Chessboard.js Interface
โ”œโ”€โ”€ MathJax Renderer
โ””โ”€โ”€ Hash-Based State Routing

All computation occurs locally within the browser.

No server-side database is required.

๐Ÿ›  Technology Stack

LayerTechnology
Database EngineDuckDB WASM v1.29.0
Dataset FormatApache Parquet
Chess Rules Enginechess.js v0.10.3
Board Rendererchessboard.js v1.0.0
MathematicsMathJax v3
State ManagementURL Hash Routing
DeploymentStatic Hosting
Analytics LayerSQL Query Engine
Data ProcessingBrowser Web Workers

๐Ÿ“ Repository Structure

โ”œโ”€โ”€ Tests/
โ”‚   โ”œโ”€โ”€ test.js
โ”‚   โ””โ”€โ”€ test.html
โ”œโ”€โ”€ src/
โ”‚   โ”œโ”€โ”€ css/
โ”‚   โ”‚   โ””โ”€โ”€ style.css
โ”‚   โ””โ”€โ”€ js/
โ”‚       โ””โ”€โ”€ app.js
โ”œโ”€โ”€ index.html
โ”œโ”€โ”€ CODE_OF_CONDUCT.md
โ”œโ”€โ”€ CONTRIBUTING.md
โ”œโ”€โ”€ LICENSE
โ”œโ”€โ”€ README.md
โ”œโ”€โ”€ biblio.bib
โ””โ”€โ”€ paper.md

File Overview

File / DirectoryDescription
index.htmlMain application entry point
src/css/style.cssApplication styling and responsive layout
src/js/app.jsCore analytical engine, DuckDB integration, filtering, dashboards, visualizations, exports, and routing
Tests/test.jsAutomated verification suite covering analytical and platform modules
Tests/test.htmlBrowser-based test runner
README.mdProject documentation and usage guide
LICENSEMIT License
biblio.bibBibTeX bibliography for software and dataset citations
CONTRIBUTING.mdGuidelines for community contributions
CODE_OF_CONDUCT.mdCommunity standards and expected behavior
paper.mdProject paper and research software description

๐Ÿ“Š Dataset

The software is powered by:

CMEED v1.0

Chess Multiverse Error & Evaluation Dataset

Features

  • Nearly 1 million human errors
  • Engine evaluation changes
  • Clock metadata
  • Opening classifications
  • Event metadata
  • Player metadata
  • Behavioral metrics
  • Position snapshots

Dataset DOI

https://doi.org/10.5281/zenodo.20625716

๐Ÿงฉ Research Modules

1. Error Explorer

The central analytical environment of the platform.

Capabilities

  • Severity filtering
  • Critical moment detection
  • Rating-based analysis
  • Player filtering
  • Event filtering
  • Opening filtering
  • Time-pressure analysis
  • Phase analysis
  • SQL-driven sorting
  • CSV export
  • Markdown export

Supported Error Types

  • Inaccuracy
  • Mistake
  • Blunder

Advanced Features

  • Critical position extraction
  • Evaluation-drop thresholds
  • Deep-link reproducibility
  • Interactive position inspection
  • Similar-error exploration

2. Opening Atlas

Opening-specific behavioral analysis engine.

Metrics

  • Error volume
  • Blunder percentage
  • Average evaluation loss
  • Pressure frequency
  • Opening Danger Index (ODI)

Research Questions

  • Which openings generate the most blunders?
  • Which openings become unstable under time pressure?
  • Which ECO families produce the highest evaluation loss?

3. Event Explorer

Tournament-level analytical environment.

Metrics

  • Total errors
  • Blunder rates
  • Average evaluation loss
  • Time-pressure frequency

Applications

  • Broadcast analysis
  • Tournament comparison
  • Event pressure studies
  • Competitive environment research

4. Player Profiles

Behavioral profiling system for individual players.

Metrics

  • Error frequency
  • Blunder counts
  • Average evaluation loss
  • Panic Index
  • Preferred openings
  • Vulnerability patterns

Applications

  • Elite player studies
  • Comparative behavioral analysis
  • Individual weakness identification

5. Research Dashboard

Interactive statistical aggregation module.

Included Visualizations

  • Error Phase Distribution (Opening, Middlegame, Endgame)
  • Severity Distribution (Inaccuracy, Mistake, Blunder)
  • Time Pressure Analysis
  • Color Bias Analysis
  • Move Heatmap

6. Methodology Framework

Included topics:

  • Expected Score Loss (ESL)
  • Opening Danger Index (ODI)
  • Panic Index
  • Metric derivations
  • Reproducibility tutorials
  • Validation workflows
  • Citation guidance

โšก DuckDB-Powered Analytics

The application dynamically generates SQL from user-selected filters.

Example

ABS(eval_before) <= 1.25
AND eval_change >= 1.50

Queries are executed directly inside DuckDB WASM.

No server communication is required.

๐Ÿ”— Reproducible Deep Linking

The platform maintains analytical state through URL serialization.

Researchers can:

  • Configure filters
  • Copy URLs
  • Share findings
  • Reproduce exact views

This supports transparent and reproducible computational research.

โ™Ÿ Interactive Position Analysis

Features include:

  • Interactive chessboard
  • Move replay
  • FEN export
  • Best-move comparison
  • Played-move comparison
  • Evaluation timeline
  • Expected Score Loss display
  • Lichess integration

๐Ÿงฎ Mathematical Framework

Expected Score Function

Models winning probability based on engine evaluation.

ES(e)=11+eโˆ’e/2.2ES(e)=\frac{1}{1+e^{-e/2.2}}

Where:

  • e = engine evaluation
  • 2.2 = calibration constant

Expected Score Loss (ESL)

Measures winning probability lost through a single human decision.

ESL=maxโก(0,ES(ebefore)โˆ’ES(eafter))ร—100ESL=\max\left(0, ES(e_{before}) - ES(e_{after})\right)\times100

Opening Danger Index (ODI)

Quantifies the behavioral risk associated with specific opening families.

ODI=(โˆ‘i=15wiNi)ร—100ODI=\left(\sum_{i=1}^{5} w_iN_i\right)\times100

Where:

  • (w_i) represents the normalized weight of component i
  • (N_i) represents the normalized score of component i

Components

VariableWeight
Error Volume0.28
Blunder Density0.22
Evaluation Magnitude0.24
Player Diversity0.14
Time Pressure Frequency0.12

Panic Index

For moves played under severe time pressure:

PI<60=1โˆฃM<60โˆฃโˆ‘xโˆˆM<60ฮ”exPI_{<60} = \frac{1}{|M_{<60}|} \sum_{x\in M_{<60}} \Delta e_x

๐Ÿ”ฌ Reproducibility Tutorials

Example Study 1

Hypothesis

Grandmaster blunder rates in hypermodern openings increase significantly when the clock drops below 30 seconds.

Example Study 2

Example Hypothesis

Magnus Carlsen exhibits measurable evaluation-loss vulnerability while defending difficult positions under severe time pressure.

Researchers can automatically generate validation subsets and export datasets for independent verification.

๐Ÿ“ Installation

Because the application depends on WebAssembly, Web Workers, and Browser Fetch APIs, it cannot be executed using the local file:// protocol.

A local web server is required.

1. Clone Repository

git clone https://github.com/sciencewithsaucee-sudo/Chess-Multiverse-Error-Evaluation-Dataset-CMEED-.git

cd Chess-Multiverse-Error-Evaluation-Dataset-CMEED-

2. Start a Local Server

python3 -m http.server 8000

Option B: Node.js

npx serve

Option C: Visual Studio Code

Install the Live Server extension and click Go Live.

3. Open Application

http://localhost:8000

๐Ÿ“ค Export Support

The application supports:

  • CSV Export
  • Markdown Brief Export
  • Shareable URLs
  • FEN Export

๐Ÿงช Automated Verification Framework

Current status:

16 Tests Passing
0 Failures
12 Core System Layers Verified

Validation Coverage

Module 1: Parquet Ingestion Layer

  • Dataset loading verification
  • Binary fetch validation
  • Schema integrity checks

Module 2: DuckDB Relational Projections

  • Runtime view construction
  • Clock bucket classification
  • Numeric evaluation casting

Module 3: Dynamic SQL Filter Compiler

  • Complex filter generation
  • Multi-variable query validation
  • Analytical protocol enforcement

Module 4: Expected Score Loss (ESL)

  • Probability transformation validation
  • Evaluation-loss calculations
  • Edge-case handling

Module 5: Opening Danger Index (ODI)

  • Composite risk-score calculations
  • Weight consistency verification
  • Numerical stability testing

Module 6: Panic Index Framework

  • Time-pressure aggregation logic
  • Evaluation variance calculations
  • Threshold validation
  • Structural position matching
  • Material isomorphism validation
  • False-positive prevention

Module 8: Reproducibility Layer

  • URL state serialization
  • State deserialization
  • Deep-link integrity checks

Module 9: Export Systems

  • CSV generation
  • UTF-8 validation
  • Record-alignment verification

Module 10: Dashboard Synchronization

  • UI aggregation validation
  • Database consistency checks
  • Reactive state verification

Module 11: Position Replay Engine

  • FEN reconstruction
  • Chess.js legality verification
  • Move replay validation

Module 12: Performance Benchmarks

  • DuckDB query latency testing
  • Asynchronous execution validation
  • Runtime stability checks

Benchmark Results

MetricResult
Total Tests16
Failures0
Verification Layers12
Benchmark Query Time88 ms

๐Ÿ“– Citation

If this software contributes to your research, please cite both the software and the underlying dataset.

Software Citation

Varshney, S. (2026).

Chess Multiverse Error Explorer (Version 1.0.0) [Computer software].

Zenodo.

https://doi.org/10.5281/zenodo.20681955

Dataset Citation

Varshney, S. (2026).

Chess Multiverse Error & Evaluation Dataset (CMEED v1.0) [Data set].

Zenodo.

https://doi.org/10.5281/zenodo.20625716

๐Ÿ“œ License

This project is released under the MIT License.

You are free to:

  • Use
  • Modify
  • Distribute
  • Commercialize

the software provided that the original license and copyright notice remain included.

See the LICENSE file for details.

๐Ÿ‘จโ€๐Ÿ”ฌ Author

Sparsh Varshney
Founder, Chess Multiverse Lab

ORCID: 0009-0004-7835-0673

Research Interests

  • Chess Analytics
  • Cognitive Performance Modeling
  • Human Error Research
  • Computational Behavioral Science
  • Open Research Infrastructure

๐ŸŒ Vision

The Chess Multiverse Error Explorer seeks to transform chess databases from repositories of moves into laboratories of human decision-making, enabling researchers to study how expertise, pressure, complexity, and cognition interact in one of the world's most demanding intellectual domains.