Now that we have the basics established, it’s time for a ton of data. Some of the most interesting correlations include April winning percentage compared with end-of-year winning percentage, which shows a correlation coefficient of .539. Surprisingly, April run differential actually implies a stronger correlation in terms of end-of-season winning percentage. It isn’t a huge difference, but the correlation of .584 is a bit stronger.
Typically run differential at the start of the season can have a small sample size problem. In this instance, however, a small sample size is not an issue because this data includes 570 different Aprils and over 14,000 games. This would imply that outscoring your opponents is a better indicator of end-of-season success than actual wins and losses. A very surprising discovery indeed.
There is a much weaker link when analyzing winning percentage through the end of April to winning percentage after April, coming out to only .376, which shows there is a correlation between April performance and after-April performance, but it is not very strong. This is much weaker than the .432 correlation of April run differential and after April winning percentage. Essentially, end-of-April win percentage means practically nothing for performance after the month of April and run differential in April is a much better indicator of performance after the month of April.