I remember about ten years ago me has used a neural network to predict the resuе of lottery similar to this. Got of the shelf software  the only problem was to develop the automated teaching algo suitable for such kind of tasks, and I did it. The algo was unique (proud to say,  no body’s ever thought of that before), and I won some rounds, not always,  but won. But then I gave it up, еven though I didn’t lose in the end,  guess me should have kept playing.