Wednesday 10 September 2014

Explanation & Validity of KIJHL Power Rankings

I know in the past there has been some misunderstanding of the Power Rankings, and let me give a little history.

Explanation

The idea behind calculating these stats is to present statistics in a clear way when teams face off. Similar to the first hockey video game I owned on my PC, that presented teams like this.

So I calculate offense, defense, goaltending and special teams.
  • Offense is the teams average goals forced per game (GFA). 
  • Defense is the teams average goals against per game (GAA). 
  • The special teams are calculated by adding the teams power play percentage with their penalty kill percentage. 
  • Goaltending is calculated by using the goalies, and by extension, the teams save percentage.
All these help indicate (a correlation not causation) of how a team is performing during the season (I have yet to figure out how to collect teams stats based on their past 10 games, which would help indicate where a team is presently at).

I vented frustration on twitter the other day about wanting the KIJHL to track shots on the team standings page, similar to how they track goals scored and goals against. Because to calculate the save percentage, I have to manually go into every teams individual statistics page and calculate the save percentage. This would be so much easier if it were on the standings page. Plus, we could then calculate offense efficiency by seeing the teams shooting percentage (goals scored divided by shots taken).

Validity

When I did the comparison post of Power Rankings vs Standings, I said it would be interesting to see how the Power Rankings predict for this coming season, to increase validity. But then I realized there is a whole history of stats on the KIJHL website. So I went back in time and calculated the Power Rankings for the past three seasons (2012-13, 2011-12, 2010-11). For simplicity I will list key stats, not all the stats.

  • 2010-11 Season
    • Osoyoos Coyotes - League Winner, 1st in League Standings, and 1st in Power Rankings
    • Castlegar Rebels - League Finalist, 3rd in League Standings, and 3rd in Power Rankings
    • Fernie Ghostriders - Conference Finalist, 2nd in League Standings, and 2nd in Power Rankings
    • Revelstoke Grizzlies- Conference Finalist, 4th in League Standings, and 4th in Power Rankings

  • 2011-12 Season
    • Beaver Valley - League Winner, 1st in League Standings, and 1st in Power Rankings
    • Kelowna Chiefs - League Finalist, 13th in League Standings, and 15th in Power Rankings (Hello there statistical anomaly and outlier)
    • Fernie Ghostriders - Conference Finalist, 5th in League Standings, and 5th in Power Rankings
    • Sicamous Eagles - Conference Finalist, 15th in League Standings, and 12th in Power Rankings (And there is another one, you are messing with my data!)
  • 2012-13 Season
    • Castlegar Rebels - League Winner, 1st in League, and 1st in Power Rankings
    • North Okanagan Knights - League Finalist, 3rd in League Standing, 5th in Power Rankings
    • Golden Rockets - Conference Finalist, 8th in League Standing, 9th in Power Rankings
    • Osoyoos Coyotes - Conference Finalist, 10th in League Standing, 10th in Power Rankings
Observation #1
So as we can see, the for those three seasons, not only was the top power ranked team the league winner, but there is a correlation that the top team in league standings also wins the league title. 
Observation #2
That was until this season past (2013-14) where the top power ranked team (Beaver Valley) was different from the top team in league standings (Kamloops), and the top power ranked team won the league title.
Observation #3
As we are learning with "advanced stats" at the NHL level, they may not predict the winner, and they are not the end all and be all, but they provide another level of depth of understanding a teams performance.

The first Power Rankings will arrive after two weeks of play. And I sure hope the KIJHL starts including shot totals on league standings!

No comments:

Post a Comment