trading_env step(action) returns previous observation

[henrydeclety]

Going through your code for TradingEnv, something seems odd to me, the function step(action) is supposed to return the next observation right?

In step(action), if it is using prices[self._current_tick] to calculate the reward given by action, the it should return the observation which contains the data point at self._current_tick right? In that case shouldn't get_observation be

def _get_observation(self):
        return self.signal_features[(self._current_tick-self.window_size+1):self._current_tick+1]

instead of

def _get_observation(self):
        return self.signal_features[(self._current_tick-self.window_size):self._current_tick]

?

Thank you for your help

[AminHP]

Hi @henrydeclety,

I think you are right. I had fixed the issue in my other repo but forgot to update it here. Thanks for noticing. I released a new version.