The main point of discussion on Tuesday, was multiple linear regressions. We discussed the importance of how variables are either useful (# of beers when calculating a BAC) and not useful (# of days per week a person attends church when calculating a BAC) in multiple linear regressions. Having a million different variable continually increasing the R^2 value unless you are looking at an adjusted R^2 which penalizes each additional variable.
R^2 is a correlation coefficient. This means that it is the proportion of the error that can be explained by the variables. It varries from -1 to 1 and the closer to these numbers R^2 is, the stronger the linear relationship between the variables.