I made a regression equation using shots on goal, foul differential (fouls suffered - fouls committed so a positive number is good), and corner kicks as independent variable and points as the dependent variable. I would have used shots on goal differential and corner kick differential, but MLS doesn't provide opponents' shots on target and opponents' corner kicks easily. What surprised me was that there was a slight negative correlation between corner kicks and points, although it was pretty close to 0. The equation was: Points = 0.4058*shots on goal + 0.07617*(fouls suffered - fouls committed) - 0.007080 * corner kicks - 17.6645 The coefficient for corner kicks was so low that the difference between the most and fewest corner kicks, 91, would expect to affect points by 0.644.
The result for the corner kicks is quite surprising. Maybe you could also add the number of goals scored as independent variable (or is it too much correlated to the points made?), or find the correlation between the goals and the corner kicks. I noticed (at least in Italian Serie A and Serie B) that there are very few teams who can kick a corner as it should be kicked, either because the manager developed some strange tactic that always ends with the losing of the ball, or because the ball goes too high, too low, or goes out even before it arrives in the box, and so on. I would be curious to know if some teams get a real advantage from having more corner kicks than others.
The regression equation for corner kicks (one independent variable) and goals scored (one dependent variable) is: Goals = 0.1827*Corner kicks + 18.0880 The correlation was 0.4237. Since corner kicks are positively correlated with goals scored but slightly negatively correlated with points, I would expect corner kicks to be positively correlated with goals allowed (meaning that taking more corner kicks goes with allowing more goals). That wasn't the case. The correlation was -0.0967. This makes me wonder if clubs with more corner kicks do a worse job at translating their goal differential into as many points as possible. It make me calculate the correlation between corner kicks and points without the other two original independent variables. This correlation was 0.3007. Therefore the regression line with three independent variables made the relationship between corner kicks and points look different than it actually is (or I made a mistake).