For this deep dive I wanted to look at how well the pre-season does at predicting the rest of the regular season. The pre-season gives everyone a first look at how teams are going to play in the coming year and evaluate the rookies selected only a short few days before. This is the first time that we get to see these teams out on the ice and to really get an understanding of what the new season has in store. But the question remains, do pre-season games actually matter? No, of course not. You don’t get points and you don’t win cups from pre-season games. Do they have some predictive quality? Let's find out! (Spoilers: No)


From the index I took the pre-season and regular season data from season 78. There is a lot of information there but I was only interested in the points percentage for each team. Gathering that information we get a table that looks like this!



Cool. Now what? Well, we need to do some simple stats and see if there is any correlation. To do that I looked at the Pearson Correlation Coefficient and ran a simple t-test. Is this the in-depth statistical analysis that would make my old stats professors proud? No not really but it gives a quick and dirty look into the relationship between the pre-season and the regular season which is what we are after! So what did we find?


The Pearson Correlation Coefficient gave us a value of 0.44. So, we can see that there is some positive correlation but it is not very strong. This makes sense though, good teams will still be good but because any team is only going to play seven other teams their opposition has a larger impact than the team itself. If you are an average team but play games against the seven best teams in the league you probably are going to have a pre-season that reflects poorly and undervalues your actual talent. Likewise, if you are an average team but are only playing against weak opposition you are going to look stronger than you actually are. In the regular season every team is playing every other team so we get a more accurate representation of their true talents.


What about the t-test? The t-test allows us to see if there are meaningful differences between the data or if they can be explained by random factors. Running a t-test we get a value of 0.32 this is well above the standard 0.05 we would look for if there was a meaningful difference between the two. From this we can say that there is no meaningful connection between the pre-season and regular season records for SMJHL teams.


So pre-season games are meaningless? Yes and no. It gives teams a chance to see how individual players are doing and evaluate the team for the coming year. This evaluation though has to take into account what teams they are actually playing. Losing close games to good teams puts you in a better spot than cleaning the floor with weaker ones. Teams that want to be competitive should use the pre-season as a chance to try new strategies and evaluate player pairings but GMs should not put too much weight on the results of the pre-season.


(553 words)

If Myers has million fans, then I’m one of them.

If Myers has one fan, then I’m THAT ONE.

If Myers has no fans, that means I’m dead.

@a_canuck Approved. Your TPE will be applied shortly.