Targeting Systems

Straight-Line-Based

• Creating Distance-Minimization-Based Straght-Line Targeting Model

  • Use multiple SVR / SVM to determine straight line.

  • Note that each model must only have one-column input and output, representing each dimension.

  • Additionally, use the output as the same as input, so that the models will learn to create hard decision boundaries which can be extracted from the model parameters.

  • Also for SVM’s outputs, just use sign function on inputs.

  • Questionable implementation at best. However, I will not stop game designers from making their games look “smart”.

  • Minimal implementation takes a minimum of 30 minutes using DataPredict™.

Cluster-Based

• Creating Distance-Minimization-Based Clustering Targeting Model

  • Find the center of players based on number of clusters.

  • Best suited for precise targeting to multiple areas.

  • Minimal implementation takes a minimum of 30 minutes using DataPredict™.

• Creating Probability-Maximization-Based Clustering Targeting Model

  • Produces clusters that maximizes the likelihood of being hit.

  • Minimal implementation takes a minimum of 30 minutes using DataPredict™.

• Creating Probability-Based Targeting Model

  • Find the center of players by finding the area with high player density.

  • Can perform both precise and intentionally inaccurate, yet likely-to-hit targeting.

  • Only one cluster.

  • Minimal implementation takes a minimum of 30 minutes using DataPredict™.

• Creating Reward-Maximization-Based Targeting Model

  • If your map has terrains and structures, the model may learn to corner and “trap” players to certain locations for easy “kills” or “targets”.

  • Limited to one target at a time, but can take in multiple player locations.

  • Might be the most terrible idea out of this list. However, I will not stop game designers from making their games look “smart”.

    • The model will likely do a lot of exploration before it can hit a single player. Once that particular location is marked as “reward location”, the model will might overfocus on it.

  • Minimal implementation takes a minimum of 2 hours using DataPredict™.