Post by bertrand on Jul 25, 2016 11:39:23 GMT -8
Hello fellow General Game Players,
We are pleased to announce the upcoming International General Game Playing Competition for 2016. We hope that you used the past year to improve your player in new and meaningful ways. We look forward to witnessing strong player performance. The competition level rises every year; will you be up to the task?
This competition will feature a selection of brand new games in the spirit of General Game Playing. We are introducing a set of new games where we expect that the analysis of game rules will result in considerable performance improvements. There are already examples of such games in use, where the game can be factored in multiple subgames; examples include "Multiple Tic-Tac-Toe", "Joint Buttons and Lights", and "Best Buttons and Lights". We will expand the list of such games for the upcoming competition. We also plan to introduce games where the game's original formulation can be highly inefficient. A reformulation can model the exact same game, but with a considerable speed increase. This is the first set of games where the Game Manager will receive a syntactically different set of rules than the players, but the two set of rules being reformulations of each other, the game manager and the players will still be able to communicate as if they had the same set of rules!
You can expect new games illustrating these new points to be published on the Stanford repository gamemaster.stanford.edu in the upcoming weeks.
We would like to take this opportunity to describe the direction we would like the competition to take. Even though GGP players are expected to perform well on any possible game, it is understandable that they are tailored to address games which are currently played. As such, selecting some types of games in past competition events may introduce a bias. This bias becomes an issue when some technique yields lower returns than some other technique only because the current set of games favors only the latter. It is not true that a stagnant set of games leads to stagnation in research: chess has been at the root of many breakthroughs in game playing. But we believe that an appropriate choice of games can lead to even more fruitful results. The goal of General Game Playing is to perform automatic analysis of the problem at hand and efficient decisions during gameplay. Being a strong player is merely a consequence of it. We aim to pick a selection of games where advances in game analysis and advances in gameplaying both give worthy rewards.
But let's go back to the competition and talk about logistics. This year will feature a pre-selection phase and a main event. During the pre-selection phase, the players will play a series of games that are here to test their ability to communicate properly with the game manager as well as play a variety of games. We want to emphasize that this pre-selection phase is not meant to gauge the strength of the players, only their ability to give legal moves. Each pre-selection phase is expected to run for 2 hours. Players are free to participate in as many pre-selection phases as they desire. Registration will open on August 8th at games.stanford.edu.
Dates of the qualifiers:
Qualifier 1: August 15th 2016, 8 a.m PST
Qualifier 2: August 17th 2016, 2 p.m PST
Qualifier 3: August 20th 2016, 8 p.m PST
The main competition will be held during two days and will use the same format as last year. During Day 1, players will be assigned in pools where they will have to perform on a variety of games. At the end of the day, each player score will be summed and the highest ranked players will advance to day 2. Day 2 will consist of a single elimination bracket where the games are guaranteed to be two player zero-sum games. Each elimination match will consist of a series of different games testing various aspects of GGP.
Dates of the main competition:
Day 1: August 24th, 7 a.m PST
Day 2: August 25th, 7 a.m PST
As usual, we will hold a Carbon vs Silicon showmatch at the very end to see if humanity manages to avoid extinction for one more year!
Thank you very much, and happy gaming!
The Stanford GGP Community
We are pleased to announce the upcoming International General Game Playing Competition for 2016. We hope that you used the past year to improve your player in new and meaningful ways. We look forward to witnessing strong player performance. The competition level rises every year; will you be up to the task?
This competition will feature a selection of brand new games in the spirit of General Game Playing. We are introducing a set of new games where we expect that the analysis of game rules will result in considerable performance improvements. There are already examples of such games in use, where the game can be factored in multiple subgames; examples include "Multiple Tic-Tac-Toe", "Joint Buttons and Lights", and "Best Buttons and Lights". We will expand the list of such games for the upcoming competition. We also plan to introduce games where the game's original formulation can be highly inefficient. A reformulation can model the exact same game, but with a considerable speed increase. This is the first set of games where the Game Manager will receive a syntactically different set of rules than the players, but the two set of rules being reformulations of each other, the game manager and the players will still be able to communicate as if they had the same set of rules!
You can expect new games illustrating these new points to be published on the Stanford repository gamemaster.stanford.edu in the upcoming weeks.
We would like to take this opportunity to describe the direction we would like the competition to take. Even though GGP players are expected to perform well on any possible game, it is understandable that they are tailored to address games which are currently played. As such, selecting some types of games in past competition events may introduce a bias. This bias becomes an issue when some technique yields lower returns than some other technique only because the current set of games favors only the latter. It is not true that a stagnant set of games leads to stagnation in research: chess has been at the root of many breakthroughs in game playing. But we believe that an appropriate choice of games can lead to even more fruitful results. The goal of General Game Playing is to perform automatic analysis of the problem at hand and efficient decisions during gameplay. Being a strong player is merely a consequence of it. We aim to pick a selection of games where advances in game analysis and advances in gameplaying both give worthy rewards.
But let's go back to the competition and talk about logistics. This year will feature a pre-selection phase and a main event. During the pre-selection phase, the players will play a series of games that are here to test their ability to communicate properly with the game manager as well as play a variety of games. We want to emphasize that this pre-selection phase is not meant to gauge the strength of the players, only their ability to give legal moves. Each pre-selection phase is expected to run for 2 hours. Players are free to participate in as many pre-selection phases as they desire. Registration will open on August 8th at games.stanford.edu.
Dates of the qualifiers:
Qualifier 1: August 15th 2016, 8 a.m PST
Qualifier 2: August 17th 2016, 2 p.m PST
Qualifier 3: August 20th 2016, 8 p.m PST
The main competition will be held during two days and will use the same format as last year. During Day 1, players will be assigned in pools where they will have to perform on a variety of games. At the end of the day, each player score will be summed and the highest ranked players will advance to day 2. Day 2 will consist of a single elimination bracket where the games are guaranteed to be two player zero-sum games. Each elimination match will consist of a series of different games testing various aspects of GGP.
Dates of the main competition:
Day 1: August 24th, 7 a.m PST
Day 2: August 25th, 7 a.m PST
As usual, we will hold a Carbon vs Silicon showmatch at the very end to see if humanity manages to avoid extinction for one more year!
Thank you very much, and happy gaming!
The Stanford GGP Community
