public class genom2 {
   static class NN {
      double    temperature;
      int       depth;
      int[]     layer;
      double[] weight;
      NN(double t, int d, int[] l, double[] w) {
          temperature = t; depth = d;
          layer = l;       weight = w;

      }
      double sigmoid(double x) {
          return (1.0 / (1.0 + Math.pow(2.7182818, -x / temperature)));
      }
      boolean wantToSave(int pts, int old, int opp) {
          return evaluate(new double[] {pts, old, opp}) < 0.5;
      }

      double evaluate(double[] input) {
          int Weight = 0, Out = 0;
          double[] cell = new double[20];
          for (int i = 0; i < input.length; i++) cell[i] = input[i];
          for (int level = 1; level < depth; level++) {
             int In = Out;
             Out += layer[level - 1];
             for (int j = 0; j < layer[level]; j++) {
                double s = weight[Weight++]; // bias
                for (int i = 0; i < layer[level - 1]; i++)
                   s += weight[Weight++] * cell[In + i];
                   cell[Out + j] = sigmoid(s);
             }
         }
         return cell[Out];
      };
    }
/* Das Neuronale Netz wurde offline und einem Genetischen Algorithmus,
   gegen Roll17 trainiert, der die Weights evolviert */
static NN net5328=new NN(0.13, 3, /*layer*/ new int[] {3, 8, 1, }, new double[]   {
 /* in1.1 */ -373.1,   /*w3_0*/ 108.8,  /*w3_1*/-255.6,  /*w3_2*/   -0.2,
 /* in1.2 */    53.7,  /*w4_0*/   3.6,  /*w4_1*/   0.2,  /*w4_2*/   -0.6,
 /* [...] */
 /* in1.8 */     2.3, /*w10_0*/  5.6, /*w10_1*/  -0.1, /*w10_2*/   49.9,
 /* on2.1 */     0.1, /*w11_0*/ -0.1, /*w11_1*/   0.1, /*w11_2*/   -7.2,
                      /*w11_3*/424.1, /*w11_4*/ 108.8, /*w11_5*/-246.8,
                      /*w11_6*/ -0.1, /*w11_7*/  -0.0, });