laskin
07-18-2010, 10:25 AM
Hello everybody. This is my first post in this forum. I hope I will be posting more in the future. I am really eager to learn more statistics than my current abilities.
My question is related to a study that I have planned but am not sure as to which statistical method best suits my purposes. The study involves people playing the game of Rock-Paper-Scissors against a computer. The computer's choices are predetermined (randomly) but the subjects think the computer is "playing" against them.
Then I record subjects' choices for each game (there will be about 100 consecutive games) which they give using the computer keyboard. The key thing is to try to extract interesting patterns from their choices, the way they depend on their past choices as well as the computer's past choices. In short, I want to find out how their current choice depends on their past history and the history of their opponent. E.g. if the computer chooses ROCK three times in a row, the subjects are more likely (than the expected 0.33 probability) to choose scissors on the 4th game, or if the subject plays a sequence like Scissors-Paper-Scissors, it is more likely (than the expected 0.33 probability) to choose Scissors again on the 4th game.
These are all speculations but I am just writing them to illustrate what kind of patterns I am trying to find (if they exist at all).
Basically, I want such a statistical method that it will use as input both subjects' N number of choices and computer's M number of choices and extract some interesting patterns in subjects' future choices. The numbers N and M are to be empirically found (that is, how many of the past games do subjects usually remember/take into account when trying to make a decision for their current choice).
I hope I could explain my idea in a non-vague manner.
Thank you in advance.
My question is related to a study that I have planned but am not sure as to which statistical method best suits my purposes. The study involves people playing the game of Rock-Paper-Scissors against a computer. The computer's choices are predetermined (randomly) but the subjects think the computer is "playing" against them.
Then I record subjects' choices for each game (there will be about 100 consecutive games) which they give using the computer keyboard. The key thing is to try to extract interesting patterns from their choices, the way they depend on their past choices as well as the computer's past choices. In short, I want to find out how their current choice depends on their past history and the history of their opponent. E.g. if the computer chooses ROCK three times in a row, the subjects are more likely (than the expected 0.33 probability) to choose scissors on the 4th game, or if the subject plays a sequence like Scissors-Paper-Scissors, it is more likely (than the expected 0.33 probability) to choose Scissors again on the 4th game.
These are all speculations but I am just writing them to illustrate what kind of patterns I am trying to find (if they exist at all).
Basically, I want such a statistical method that it will use as input both subjects' N number of choices and computer's M number of choices and extract some interesting patterns in subjects' future choices. The numbers N and M are to be empirically found (that is, how many of the past games do subjects usually remember/take into account when trying to make a decision for their current choice).
I hope I could explain my idea in a non-vague manner.
Thank you in advance.