deutsch
englishBrainDump1 is a program that allows you to play Rock, Paper, Scissors online. At the beginning of each game, you can choose the number of rounds and a strategy.
The role of chance in Rock, Paper, Scissors
Rock, Paper, Scissors is a game in which you can never lose (or win) by making random decisions. It is not possible to beat a strategy that always chooses Rock, Paper or Scissors by random in the long run.
It already becomes clear that there is no certain way to win if one takes into consideration that a certain series of choices cannot be considered as generally good or generally bad. Is it dumb to choose Rock twice in a row? Three times? It could indeed be very dumb, yet, it could also be very clever. It depends on the specific game situation.
In Artificial Intelligence (AI), we call this the "psychogenic" aspect of stochastic mathematics. This aspect provides the game’s "thrill". That is why in the "real" Rock-Paper-Scissors scene, the game is generally only played until a player has won two rounds. In longer games, chance plays too big a role.
If humans want to win Rock, Paper, Scissors, they generally develop certain behaviour patterns by which they believe to increase the probability of a win. BrainDump1 tries to recognise patterns in such behaviours and to draw an advantage from this recognition. BrainDump "knows" more at the end of the game than in the beginning. (However, the current version forgets everything that it has learned after each game.)
At present, a few simple strategies and learning mechanisms are implemented, which can roughly be divided into three groups:
- Strategy 1 to 4: BrainDump1 acts according to certain patterns and does not react to the human player. (autonomous strategy)
- Strategy 6 to 7: BrainDump1 reacts to the human player; however, it does not learn. (primitive strategy)
- Strategy 8 to 11: BrainDump1 tries to calculate the probability of the next turn based on its present knowledge. i.e. it learns from the course of the game. (adaptive strategy)
BrainDump1 can be at its strongest only if the human player really attempts to win, since this leads her or him to reflect before each turn and to engage both with strategies and with the game's basic principles. If you just choose randomly, you will in fact not lose against BrainDump1 - but you will not win, either; at least not with the simpler strategies.