Predicting the 2008-09 Season, Part 1
Two years ago the Montreal Canadians had the most efficient Power Play unit in the entire NHL. The following season Montreal saw PP specialist Sheldon Souray depart for Edmonton and yet they repeated as the best in the NHL in PP efficiency.
Players, fans and GMs often fall talk about the upcoming season by saying "well our club was pretty good at ____ and we should be better in our weak areas, therefore I expect us to move up the standing." That way of thinking about the upcoming season is usually very flawed. Why? Because some things tend to carry over from one season to another while other team attributes don't.
I was mulling this over the other day and it struck me that I had never actually looked at which team strengths and weaknesses are MOST likely and LEAST likely to carry over from one season to the next one. So that's what I will do in this post and the next post.
I pulled all the team level data between 2001 and 2008 from ESPN's NHL stats page and ran some basic correlations in order to figure what tends to carry over and what does not carry over. Below I discuss the the four areas that are MOST likely to be consistent from one season to the next.
1. Shots on Goal Against
The team's that were either strong or weak at preventing shots tend to repeat that the following season. It doesn't mean that there is zero change from season to season, but change is this area is difficult and few teams move significantly up or down the rankings in one season. I have to admit that I slightly surprised that SOGA finished so high.
Why is it that SOGA is rather consistent across seasons? I suppose it is because SOGA are primarily determined by the performance of team's defensemen and on most clubs turnover among the defensemen is usually a gradual process. Contrast that with say team SV% which is determined by the goaltender position--here we see much more variation from one season to the next because it is highly dependent on the starter. If your starting goalie gets hot or cold and/or gets injured it can have a huge effect on SV%, whereas losing one defensemen has less of an impact on SOGA.
2008 SOGA Leaders: DET, SJS, NYR
2008 SOGA Trailers: ATL, NYI, TBL
Thrashers Rank: 30th
2. Team Discipline (aka Times Short Handed)
Right up there with SOGA as the most consistent team statistic across season is Times Short Handed (also called Penalty Kill Opportunities). This suggest to me that team discipline--the ability to play within the rules--doesn't change all that much from one year to the next. Undisciplined teams tend to stay undisciplined and clean playing teams continue to stay out of the Penalty Box.
2008 xSH Leaders: COL, SJS, LAK
2008 xSH Trailers: ANA, CGY, PHI
Thrashers Rank: 18th
3. Shots-On-Goal-Differential.
You could argue this is a repeat of #1, but teams that get out shot tend to stay that way and vice versa. Notice that SOGA and SOG Differential both made the top four most consistent team attributes, but shots on goal for (SOGF) is less consistent from season to season and did not finish as high. So offensive pressure (measured in SOGF) is less consistent that defense (measured as SOGA).
The fact that SOG Differential tends to stay the same is not good news for Thrashers fans. On the other hand the club did bring in two players who may play top four minutes on the blueline in Ron Hainsey and Zach Bogosian. On the other hand Hainsey isn't famous for his defensive proess and Bogosian while talented is just 18 years old. I expect only modest improvement.
2008 SOG Differential Leaders: DET, SJS, NYR
2008 SOG DifferentialTrailers: ATL, EDM, LAK
Thrashers Rank: 30th
4. Goal Differential (Goals For minus Goals Against)
This is the stat I expected to rank very high and it did. Goal Differential is the single best predictor of the NHL standings season after season. If you look at all of NHL history Goal Differential is an EXTREMELY strong predictor of the standings. The correlation between Goal Differential and the NHL Standings is north of .90 and is the hockey equivalent of the law of gravity.
Again this is not good news for the Thrashers after losing the offensive production of Marian Hossa and Mark Recchi. The young kids are unlikely to replace that production next season so the key will need to come from the defense and Kari Lehtonen in a contract year. This team must make huge strides against the odds to stay out of the Tavares sweepstakes.
2008 Goal Differential Leaders: DET, MON, DAL
2008 Goal Differential Trailers: ATL, NYI, TBL
Thrashers Rank: 30th
In the next post I will examine the four LEAST consistent team attributes from season to season.
Players, fans and GMs often fall talk about the upcoming season by saying "well our club was pretty good at ____ and we should be better in our weak areas, therefore I expect us to move up the standing." That way of thinking about the upcoming season is usually very flawed. Why? Because some things tend to carry over from one season to another while other team attributes don't.
I was mulling this over the other day and it struck me that I had never actually looked at which team strengths and weaknesses are MOST likely and LEAST likely to carry over from one season to the next one. So that's what I will do in this post and the next post.
I pulled all the team level data between 2001 and 2008 from ESPN's NHL stats page and ran some basic correlations in order to figure what tends to carry over and what does not carry over. Below I discuss the the four areas that are MOST likely to be consistent from one season to the next.
1. Shots on Goal Against
The team's that were either strong or weak at preventing shots tend to repeat that the following season. It doesn't mean that there is zero change from season to season, but change is this area is difficult and few teams move significantly up or down the rankings in one season. I have to admit that I slightly surprised that SOGA finished so high.
Why is it that SOGA is rather consistent across seasons? I suppose it is because SOGA are primarily determined by the performance of team's defensemen and on most clubs turnover among the defensemen is usually a gradual process. Contrast that with say team SV% which is determined by the goaltender position--here we see much more variation from one season to the next because it is highly dependent on the starter. If your starting goalie gets hot or cold and/or gets injured it can have a huge effect on SV%, whereas losing one defensemen has less of an impact on SOGA.
2008 SOGA Leaders: DET, SJS, NYR
2008 SOGA Trailers: ATL, NYI, TBL
Thrashers Rank: 30th
2. Team Discipline (aka Times Short Handed)
Right up there with SOGA as the most consistent team statistic across season is Times Short Handed (also called Penalty Kill Opportunities). This suggest to me that team discipline--the ability to play within the rules--doesn't change all that much from one year to the next. Undisciplined teams tend to stay undisciplined and clean playing teams continue to stay out of the Penalty Box.
2008 xSH Leaders: COL, SJS, LAK
2008 xSH Trailers: ANA, CGY, PHI
Thrashers Rank: 18th
3. Shots-On-Goal-Differential.
You could argue this is a repeat of #1, but teams that get out shot tend to stay that way and vice versa. Notice that SOGA and SOG Differential both made the top four most consistent team attributes, but shots on goal for (SOGF) is less consistent from season to season and did not finish as high. So offensive pressure (measured in SOGF) is less consistent that defense (measured as SOGA).
The fact that SOG Differential tends to stay the same is not good news for Thrashers fans. On the other hand the club did bring in two players who may play top four minutes on the blueline in Ron Hainsey and Zach Bogosian. On the other hand Hainsey isn't famous for his defensive proess and Bogosian while talented is just 18 years old. I expect only modest improvement.
2008 SOG Differential Leaders: DET, SJS, NYR
2008 SOG DifferentialTrailers: ATL, EDM, LAK
Thrashers Rank: 30th
4. Goal Differential (Goals For minus Goals Against)
This is the stat I expected to rank very high and it did. Goal Differential is the single best predictor of the NHL standings season after season. If you look at all of NHL history Goal Differential is an EXTREMELY strong predictor of the standings. The correlation between Goal Differential and the NHL Standings is north of .90 and is the hockey equivalent of the law of gravity.
Again this is not good news for the Thrashers after losing the offensive production of Marian Hossa and Mark Recchi. The young kids are unlikely to replace that production next season so the key will need to come from the defense and Kari Lehtonen in a contract year. This team must make huge strides against the odds to stay out of the Tavares sweepstakes.
2008 Goal Differential Leaders: DET, MON, DAL
2008 Goal Differential Trailers: ATL, NYI, TBL
Thrashers Rank: 30th
In the next post I will examine the four LEAST consistent team attributes from season to season.