| CPC A63F 13/358 (2014.09) [A63F 13/2145 (2014.09); A63F 13/42 (2014.09); A63F 13/44 (2014.09); A63F 13/69 (2014.09); G06N 20/00 (2019.01)] | 16 Claims |
|
1. A nontransitory computer-readable medium comprising program instructions for causing a computer to perform a method comprising:
storing a plurality of state values of a computer game corresponding to a plurality of player characters that are controlled in a computer game space by a plurality of player terminals over a communication network;
accepting a player input by an input device at a first timing for controlling a player character among the plurality of player characters within the computer game;
determining an estimation result corresponding to a second timing, which is later than the first timing, based on an estimation model,
wherein the estimation model produces the estimation result for the second timing based on per-timing information regarding a plurality of player inputs for the plurality of player characters that are accepted by the plurality of player terminals for the first timing,
wherein the per-timing information comprises matrix data that has rows and columns corresponding to an x-axis and a y-axis in a world coordinate system of the computer game space and in which a value of a component at a q-th row and an r-th column indicates a state of coordinates (Xq, Vr) in the world coordinate system,
wherein the matrix data indicates a first plurality of positions of the plurality of player characters in the computer game space and input content of the plurality of player inputs for the plurality of player characters, and
wherein the estimation model is a machine learning model that is generated using a learn function and a plurality of game logs that are obtained by a server for a plurality of computer games;
determining whether accepted input content of the player input accepted by the input device for the second timing coincides with predicted input content of the player input that is indicated by the estimation result for the second timing;
transmitting, in response to a determination result indicating a non-coincidence of the accepted input content and the predicted input content for the second timing, first correction information indicating the accepted input content of the player input;
acquiring second correction information transmitted by another player terminal;
compensating for the estimation result to produce a compensated estimation result based on the first correction information and the second correction information;
updating the plurality of state values of the computer game based on the compensated estimation result to produce a plurality of updated state values; and
automatically generating, based on the plurality of updated state values, a game screen that indicates the plurality of player characters are located in a second plurality of positions within the computer game space,
wherein the game screen is displayed on a display device.
|