U.S. Patent No. 10,512,846: Emulating player behavior after player departure
Issued December 24, 2019 to Sony Interactive Entertainment LLC
Priority Date: March 7, 2017
U.S. Patent No. 10,512,846 (the ’846 Patent) relates to simulating a player’s gameplay after the player leaves a multiplayer game. The ’846 Patent details a method of tracking and storing player inputs for sequence and/or timing of in-game actions in order to simulate a player’s gameplay if the player exits a multiplayer game and “autopilot mode” is activated. This enables the multiplayer game to continue by simulating the absent player’s actions based on previous tracked and stored gameplay decisions. Simulated in-game actions are determined using the probability that the missing player would take a certain action based on their tracked behavior in similar situations, the probability of taking a certain action based on the previous actions that led up to the current point in the game, and the probability that a certain action would be optimal for the current situation.
Multiplayer video games involve multiple players playing using either a single computer system or multiple computer systems connected together. While a first player of the multiplayer game is playing the multiplayer game, the first player’s actions are tracked and stored. When the first player quits or stops responding before the multiplayer game has completed, the first player is replaced by a computer-simulated version of the first player that selects its actions based on probabilities calculated from the tracked actions of the first player. The multiplayer game is thus able to continue without interruption or disruption, and the first player’s playstyle is preserved.
1. A method for simulating game player behavior, the method comprising: receiving a plurality of action inputs from a first user device associated with a first player, the action inputs associated with a plurality of action sequences in one or more game titles played by the first player via the first user device; storing the plurality of action inputs in association with a respective one of the game titles played by the first player in memory; identifying that the first player has triggered a simulation mode during a game session of an identified game title; and launching the simulation mode in the game session of the identified game title in response to the identified trigger by: identifying a first player state associated with a control object previously controlled by the first player within the game session of the identified game title, wherein the identified first player state includes a current action sequence; identifying a condition encountered by the control object; identifying a plurality of possible actions to be taken by the control object in response to the condition and the first player state; selecting a first action of the plurality of possible actions based on the stored plurality of action inputs associated with the identified game title and based on the current action sequence indicated by the identified first player state; and commanding the control object to perform the selected first action within the game session in continuation of the current action sequence.