Right now the network is looking at each value independently. This has worked, but not lead to a win.
Instead, let's make it compare each slot against the maximum value on the board, so it will consistently know what the largest value on the board is, with less guess work.
Right now the network is looking at each value independently. This has worked, but not lead to a win.
Instead, let's make it compare each slot against the maximum value on the board, so it will consistently know what the largest value on the board is, with less guess work.