In the process we have built and evaluated hundreds of different ML models using different features, architectures and training regimes following state of the art as well as our own unpublished methods... Instead of making Gaming operators adjust..
Relevance: 21.251316