kth.sePublications KTH
Operational message
There are currently operational disruptions. Troubleshooting is in progress.
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Chess-GPT: A Transformer’s Approach to Chess
KTH, School of Electrical Engineering and Computer Science (EECS).
KTH, School of Electrical Engineering and Computer Science (EECS).
2025 (English)Independent thesis Basic level (degree of Bachelor), 10 credits / 15 HE creditsStudent thesis
Abstract [en]

This study investigates the capabilities of transformer-based models in chess move generation and gameplay when trained solely on human game notations. Three GPT-2 architectures (two trained on unfiltered games and one on high-Elo games (>1800)) were evaluated for legal move accuracy and playing strength. The models achieved a 99.5–99.65% legal move rate and demonstrated intermediate playing strength (1400–1500 Elo) against Stockfish levels 0–2, despite lacking hardcoded chess rules or search algorithms. The filtered model showed marginal improvement, suggesting dataset quality impacts performance. These results highlight the promise of pure pattern recognition in constrained domains while underscoring its limitations in achieving expert-level play without symbolic reasoning.

Abstract [sv]

Denna studie undersöker transformerbaserade modellers förmåga att generera schackdrag och spela schack när de tränas enbart på mänskliga schacknotationer. Tre GPT-2-arkitekturer (två tränade på ofiltrerade partier och en på partier med hög Elo (>1800)) utvärderades med avseende på giltiga drag och spelstyrka. Modellerna uppnådde en giltighetsgrad på 99,5–99,65% och visade en medelhög spelstyrka (1400–1500 Elo) mot Stockfish nivå 0–2, trots att de saknade hårdkodade schackregler eller sökalgoritmer. Den filtrerade modellen uppvisade marginella förbättringar, vilket tyder på att datamängdens kvalitet påverkar prestationen. Resultaten belyser både potentialen hos ren mönsterigenkänning i begränsade domäner och dess begränsningar i att uppnå expertnivå utan symbolisk resonemangsförmåga.

Place, publisher, year, edition, pages
2025. , p. 523-528
Series
TRITA-EECS-EX ; 2025:152
National Category
Electrical Engineering, Electronic Engineering, Information Engineering
Identifiers
URN: urn:nbn:se:kth:diva-376173OAI: oai:DiVA.org:kth-376173DiVA, id: diva2:2034543
Supervisors
Examiners
Projects
Kandidatexamensarbete i Elektroteknik 2025, EECS, KTHAvailable from: 2026-02-02 Created: 2026-02-02

Open Access in DiVA

fulltext(80627 kB)18 downloads
File information
File name FULLTEXT01.pdfFile size 80627 kBChecksum SHA-512
35ce0a386dafe4649eb99cbe0efdfed651a3c9044e3339612422234d17a7e8ec21d4fd4aa201500c3c7a8f57194994b78b3e0cfbd5319ecd49f18a5d8a7ff775
Type fulltextMimetype application/pdf

By organisation
School of Electrical Engineering and Computer Science (EECS)
Electrical Engineering, Electronic Engineering, Information Engineering

Search outside of DiVA

GoogleGoogle Scholar
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

urn-nbn

Altmetric score

urn-nbn
Total: 3862 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf