This website introduces the PHH file format, designed to concisely describe poker hand history while allowing ease in human usage and machine parser implementations. Our solution fills a long-standing gap in the existing literature on the computer poker field. The specification builds on top of the TOML specification, allowing software systems to leverage the rich ecosystem surrounding the TOML file format. By representing different game state information through fields, the hand can be described in a consistent, structured manner. In our dataset contribution, also hosted on Zenodo, we provide PHH files in a diverse number of variants to validate the format's flexibility and effectiveness.
The PHH format is suitable to be used in a variety of poker applications in use cases ranging from poker AI development, hand archival, analytical tools, and teaching materials. A promising future direction for research in the field of computational poker is in the development of automated systems that can digitalize televised poker hands into a formats, such as the PHH format, for further analysis and study. This technology would leverage computer vision and machine learning algorithms to recognize chip amounts, player actions, and dealt cards from video footages and/or onboard graphics. Such systems could greatly enhance the efficiency and accuracy of data collection for poker research, eliminating the manual entry of hand histories and reducing human error.