Low-Risk Systems

Retention Systems

The worse case scenario is that the model either predicts too early or too late for appropriate intervention. The former leads to flooding a lot of events that could lead to player getting engaged, while the latter scenario is and equivalent of you not having a model to intervene and the player will leave anyways. This is only applicable if your intervention contributes positively to the players’ progress without any difficulty such as resource multiplier, rare item spawn, rare event spawn and so on.

• Creating Low-Risk Time-To-Leave Prediction Model

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

Targeting Systems

The worse case scenario is that the model will just miss the players, which is a desirable property for games since they are more focused in making the AIs fun and playable instead of accurate and difficult.

• 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™.

Load Shedding Systems

The worse case scenario is that the model will place the asset that are far away from the players, but this isn’t a problem given that most games literally place assets at random places for players to collect.

• Creating Probability-Based Clustering Placement Model

  • Identifies areas of high player densities for spawning assets.

  • Reduces the servers’ and clients’ computational resources by spawning items in places that are highly likely to be interact by the players.

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