© 2013 Michael Parkatti Boston-Bruins-Preseason

Yes, the Preseason Matters

Every preseason we hear the same refrain from talking heads and fans alike: “the preseason doesn’t matter”.  With the Oilers holding down a fairly sexy 2013 preseason record of 5-1-1 for a points percentage (PTS%) of 78.6%, it’s worthwhile to question whether that actually means anything.  I’m happy to report that it does.

I compiled the preseason PTS% of every team’s preseason between 2005-06 and 2011-12, along with how those teams subsequently did in each regular season.  This resulted in 210 observations (30 teams * 7 seasons).  What I wanted to do was regress the regular season points percentage on the preseason points percentage — basically, can you use preseason PTS% to predict regular season PTS%? Here’s the output:


Our R-squared correlation is only 0.053, which doesn’t sound very high, but with this many observations it provides enough freedom to test minute influences.  Our preseason PTS% coefficient has a T-stat of 3.4, and a P-value of  0.0008.  What that means is that there’s only a 0.08% chance that a team’s preseason points percentage has no explanatory power in predicting a team’s regular season points percentage.  This obviously passes a 95% test of significance, so we can say that preseason performance does matter.

Here’s a more interesting way to show the relationship:


It kind of looks like a random cloud, but adding a linear trendline does show a positive relationship between the two variables.  Generally, the better you do in the preseason, the better you’ll do in the regular season.  It may be a slight influence, but it’s there and it’s statistically significant.

Here’s my favourite table:


Teams that have preseason points percentages below 42.5% have an average regular season points percentage of 53.3%.  They are also only playoff competitive 40% of the time, and non-competitive 60% of the time (I defined “playoff competitive” as having greater than 55% points percentage in the regular season, as this is generally the level where teams lock up a playoff spot or fall just a point short).

Teams that have a greater than 65% points percentage in the preseason go on to average a 58.3% regular season points percentage.  These teams are playoff competitive 74% of the time, and non-competitive only 26% of the time.

Teams between those two extremes average a 55.4% regular season points percentage, and have a 53% chance of being playoff competitive.

The Oilers playing well in the preseason does mean something. If they win tonight, they’ll have a points percentage above 80% in the preseason.  Historically, teams above that 80% mark in the preseason average a 59.7% points percentage in the regular season, and have an 87% chance of being playoff competitive.  So cheer for them tonight!  It means something :].


  1. Bruce McCurdy
    Posted September 27, 2013 at 1:45 pm | #

    Go, Data Points, Go!

  2. Cris
    Posted September 28, 2013 at 1:37 pm | #

    I have bookmarked this page, count on me to be back when the Oilers don’t make the playoffs again this year.

  3. Posted September 29, 2013 at 3:04 am | #

    I think you probably want to use a prediction interval rather than a confidence interval here. That would mean your interval for points would be a good bit wider here.

Post a Reply to schuckers

Your email is never published nor shared. Required fields are marked *


You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>