Logistic Regression vs predicting probability by splitting data into bin

Question

So I am exploring using a logistic regression model to predict the probability of a shot resulting in a goal. I have two predictors but for simplicity lets assume I have one