Random Choice when doing prediction

[HeshamMahrous]

I am not sure why you you did the random prediction this way action = np.random.choice(self.action_space, p=prediction) also you are picking the random outcome as the chosen action during testing and training. I can understand why during training but why during testing as well ?

[pythonlessons]

You right, I shouldn't do this while testing

[HoaxParagon]

I don't think it's wrong if you look at the way that the choice method works. It selects one of the actions based on the probability set by the prediction. The reasoning for it in training makes sense but for testing I can't be sure. It does seem like it might randomly choose something that will cause it to lose.

[kitmir]

I realised this also and tried different things. Random choice works well with training but testing becomes better with a number. Best result I could achieve was to make the action when the probability/prediction is more than 67%.

[polonezu]

how can we conect this bot to demo exchange?