12-06-2021, 11:05 AM
I decided to check out the game of the day, for once, now that the web player makes it even easier. That made me wonder about the value of the ludemic approach to historical/traditional games. The game was Bajr, and playing it game against AI, it feels completely broken. It is clear that some information is lacking in the description of the game. This could be information like whether it's a children's game, or whether Bajr-players all have some unspoken agreement about certain things one simply do not do when playing (like blocking the opponent's goal area).
In short, if you strip away the context in which this game is played – if you abstract out the mathematical entity from the cultural phenomenon – what you're left with is garbage. And the question is: For what kind of end is this garbage a valuable piece of data? If this was a rare effect of such an abstraction, there would obviously not be a problem, but my impression (which might be, and hopefully is, wrong) is that this is often what happens if you regard traditional games as mathematical entities apart from the complex way it is embedded in a history and a culture.
I'm not saying this to cast doubt on the value of this project, but to ask for a broader picture. The idea of racing to fill a portion of a board is definitely a real ludeme – it's a repeatable and replicating unit of "tabletop gaming" – so of course it's interesting to record it's occurrence in space and time and map its relation to other ludemes. So the (real, and more nuanced) question is: When Ludii eventually is going to reconstruct old games based on partial descriptions, how will it know if it has found a plausible reconstruction of a silly game that only works because it is played by drunk and blind-folded players, or an implausible reconstruction of a highly regarded game played in serious competition? It seems that kind of context is important in order to get the reconstruction "right".
In short, if you strip away the context in which this game is played – if you abstract out the mathematical entity from the cultural phenomenon – what you're left with is garbage. And the question is: For what kind of end is this garbage a valuable piece of data? If this was a rare effect of such an abstraction, there would obviously not be a problem, but my impression (which might be, and hopefully is, wrong) is that this is often what happens if you regard traditional games as mathematical entities apart from the complex way it is embedded in a history and a culture.
I'm not saying this to cast doubt on the value of this project, but to ask for a broader picture. The idea of racing to fill a portion of a board is definitely a real ludeme – it's a repeatable and replicating unit of "tabletop gaming" – so of course it's interesting to record it's occurrence in space and time and map its relation to other ludemes. So the (real, and more nuanced) question is: When Ludii eventually is going to reconstruct old games based on partial descriptions, how will it know if it has found a plausible reconstruction of a silly game that only works because it is played by drunk and blind-folded players, or an implausible reconstruction of a highly regarded game played in serious competition? It seems that kind of context is important in order to get the reconstruction "right".