these things are often not driven by hand made or even understandable algorithms but by linear algebra and it is possible for a model to make that kind of connection even if there is a better indicator. idk what ubers pricing algorithm looks like but if it is a neural network which is used to determine what people are willing to pay for a ride then it could have a preference for higher prices at night and also its more likely for people to have low battery at night so it also has a weaker connection between low battery and high prices.