This dataset has information about the games completed by top 100 agents for rps competition.
It includes the following columns:
Player1: Player 1 username
Player2: Player 2 username
Result: Player 1 winner/loser/tie
Game: Total number of games completed by player 1 (from 1 to n)
P1FinalScore: Player 1 final score
P2FinalScore: Player 2 final score
Rounds: number of steps for that specific game
Movement1: Player 1 action
Movement2: Player 2 action
Reward1: Player 1 reward
Reward2: Player 2 reward
Player1CurrRank: Player 1 current ranking
Player2CurrRank: Player 2 current ranking
Player1CurrScore: Player 1 current score
Player2CurrScore: Player 2 current score
Player1TotSubmit: Player 1 total agents
Player2TotSubmit: Player 2 total agents
Player1lastAct: Player 1 activity
Player2lastAct: Player 2 activity
P1IniScore: Player 1 initial score
P1IniConfidence: Player 1 initial confidence
P1UpdateScore: Player 1 updated score
P1UpdateConfidence: Player 1 updated confidence
P1GameReward: Player 1 game reward
P2IniScore: Player 2 initial score
P2IniConfidence: Player 2 initial confidence
P2UpdateScore: Player 2 updated score
P2UpdateConfidence: Player 2 updated confidence
P2GameReward: Player 2 game reward
First action for each agent? Is there anything we could use to try to defeat it?