The longest valid game we saw, after generating a few hundred games, was this one in the animation above:
A Return to Machine Learning
Kyle McDonald
63311

Did you consider having a human interact and play a game of chess with the trained network?

Also would it be possible to add training of the network to help filter out illegal moves?

Like what you read? Give Simon Dann a round of applause.

From a quick cheer to a standing ovation, clap to show how much you enjoyed this story.