To obtain the best line of fit, we need the line which minimizes the sum of the squared distances, Therefore, from visual inspection, we can conclude that the best line of fit will be one in which the distance between each individual distance and the line of best fit is minimized, for the first graph, only the upper points are close to the regression line leaving the lower values further away, however, the secubd graph gives a more uniform distance between all the data values and the regression line. Hence, it does a better job at capturing the trend in the data.
