def attempt(self):
        observation = self.discretise(self.environment.reset())
        done = False
        reward_sum = 0.0
        while not done:
            action = self.pick_best_action(observation)
            new_observation, reward, done, info = self.environment.step(action)
            if done:
                reward = 0.0
            new_observation = self.discretise(new_observation)
            self.update_knowledge(action, observation, new_observation, reward)
            observation = new_observation
            reward_sum += reward
        self.attempt_no += 1
        return reward_sum