also if it seems to keep failing just click green flag again how many attempt does it take you to do a full round? (it took me 9 attempts)
how it works : when the project is inititialized, it is given a random weight in all of the 4 nuerons in the hidden layer, these starting weights will basically as you can say, "plant the seed of the whole run", what these weights do is that based on the sensor input, it will multiiply by the weights and output a turn value based on all the sensors and weights, it will be scored based on the travel distance, if the travel score is better than its highscore, the current weight mutation will be copyed to the template weights. and each time it fails, it will mutate its weights to find a sweet spot for them, based on the weights, the sensor input will be modifyed thus giving 1 output neruon which returns a value between -1 to 1. thanks for listeniing to my terrible explaination neural network layout : 4 inputs, 1 hidden layer, 4 hidden layer neurons, 1 analogue output neuron,
