To develop a model based upon historical data that may help in predicting the success of a game in advance thereby reducing certain level of uncertainty.
International Game Technology (IGT) is a multinational gaming company that produces slot machines and other gaming technology. The company is headquartered in London.
The company is engaged in operating and providing an integrated portfolio of technology products and services across various gaming markets, including lottery management services, online lotteries, electronic gaming machines, sports betting, interactive gaming and commercial services.
Gaming is one of the main economic sectors in the EU, with revenues of more than €80 billion a year and annual growth rate of 3%. Producing a game is a costly operation, however, the success of a game could not be evaluated until the game has been completely developed and made available to players. If the success of a game could be predicted at an early stage, it would allow game designers to make more informed decisions and modify game content to ensure success. The earlier the prediction could be made, the sooner game design could be modified resulting in major cost savings.
We want to develop a model based upon historical data that may help in predicting the success of a game in advance thereby reducing certain level of uncertainty. The historical data would include a set of successful and unsuccessful games described in terms of features, theme, bonus type, volatility level, and an image that captures the game’s theme.
The challenge has the following sample datasets available for download
The objective is to build an AI model to predict game’s success based on descriptive information such as genre, denomination, game features, number of reels, bonus type, pay table description (e.g., volatility), marketing image etc.
A predictor that can be used to predict games success. The predictor’s outputs should indicate the probability of success rather than successful or not. Accuracy level still unknown and would like to use the historical data and subject matter experts to validate our assumptions.