The NHL pre-season schedule is packed with games. 107 pre-season games are played over 18 days. That is an average of almost six games per night with each team playing an average of more than seven times. Is this really necessary? Players come to training camp in good shape. In the salary capped era, the vast majority of team rosters are set by the contract structure of their players. Most teams have at most one to two positions that may be available to be won in training camp. The majority of competition in training camp and pre-season games is to win roster spots in the AHL and not the NHL. The NHL season is enough of a marathon without tacking on an increasing number of pre-season games. These are games that exist for the most part to make the NHL money. Season ticket holding fans are forced to buy tickets to these games. Many of these games are played in neutral sites that have no other access to NHL hockey, so the game will remain a hot ticket despite the fact it is an exhibition game.
I have written a few sabermetrics and hockey about combining Corsi Number and zone starts. I listed the top 20 and worst 20 players by this metric. The main remaining flaw is that team effects are not taken into account. A player will get a good rating merely by playing on a good team and will get a poor rating by playing on a bad team. That effect is the next one to remove. This can be done by taking team Corsi Numbers which are also adjusted for zone starts and doing a team adjustment similar to the one Klein and Reif discuss in the Hockey Compendium. It is necessary to use zone start adjusted numbers consistently throughout the calculation to avoid double counting (or missing out) on zone start effects.
Every year, one or two promising restricted free agents have not signed new contracts when training camp opens up. This is a very poor method to run a hockey team. These young players are the future of the franchise and potential future stars who are being alienated because they have very little leverage in contract negotiations. Last year, the Los Angeles Kings did this with Patrick O’Sullivan. It is a logical consequence of the move that O’Sullivan has since been traded to the Edmonton Oilers (where he is a very promising young player) and all they have to show for it is Justin Williams, who struggled through injuries last year scoring only 14 points (O’Sullivan is younger and scored 44 points). The ill-will created by the tough contract negotiation usually ends badly and forces a trade down the line or else the player leaves as a restricted free agent as soon as he can.
In my look at sabermetrics and hockey, I am looking at the problem of combining zone starts and Corsi Number. Although a method has been proposed by objective NHL, I have made my own which I think better solves the problem.
Raw Corsi Numbers are used as a metric to rank puck possession. Because teams are more likely to allow shots if they are in the defensive zone or take shots if they are in the offensive zone, individual player’s Corsi Numbers are adjusted by 0.8 * (defensive zone starts - offensive zone starts), as it has been shown that an extra zone start on average yields 0.8 shots directed at the goal.
When the San Jose Sharks lost in the first round of the 2009 playoffs many of their fans called for wholesale changes. The team had done increasingly better in the regular season for the last few years but had no significant playoff success. They won the 2009 President’s Trophy and were expected to do better than a first round playoff loss.
The urge was to trade away key players. Joe Thornton and Patrick Marleau were among the candidates to be shipped out of town. That would have been a bad idea for the Sharks. They were not going to get better by subtracting parts. They were highly unlikely to add better pieces in their place if they were trading away their best players.
I have written in the past about travel problems for the western teams, but it looks like things will be even worse than usual for the Western Canadian teams this year (Calgary, Edmonton and Vancouver). The US Government has banned Canadian charter flights from flying from point to point in the United States. This ruling is an attempt to protect American jobs. Typically, an NHL team has a contract with a particular airline (most likely Air Canada if it is a Canadian team) that they use for charter flights throughout the season. They are flown to all of their games and other events. Since teams often have extended road trips with multiple road games in the US, they often require flights between US cities. In the past an exemption to this law had existed for Canadian sports teams and entertainers traveling in the US.
The NHL has several markets that are in poor shape financially. The most prominent example is that of the Phoenix Coyotes, who are in bankruptcy. The team will be auctioned off very soon, with no clear end in sight to all the litigation and potential financial issues. The Tampa Bay Lightning are probably the NHL’s next serious problem. Owners Oren Koules and Len Barrie have barely had enough money to keep the team afloat. With Barrie apparently dropping out due to his financial difficulties, it leaves behind Oren Koules to run an underfunded team with declining fan support.
Yesterday, I wrote about the sabermetrics and hockey problem of combining zone starts and Corsi Number. There is as method used by objective NHL that treats Corsi Numbers as a rate statistic and adjusts by a factor similar to the percentage of offensive - defensive zone starts divided by total zone starts. I do not think this is the best method to treat this data.
I take the raw Corsi Number as a starting point. Since each extra offensive zone start is worth on average 0.8 Corsi (or each extra defensive zone start costs on average 0.8 Corsi), I merely add on a factor of 0.8 * (defensive - offensive zone starts). I think this method is more straight forward and better accomplishes the adjustment we are looking for.
In my sabermetrics and hockey posts this summer, I have introduced the concepts of the Corsi Number and of zone starts. Corsi Numbers measure the number of shots directed at goal (including shots on goal, missed shots and blocked shots) and zone starts measure which region on the ice that a player is in when he is on the ice for a faceoff. It should be clear that if you start frequently in the offensive zone that you should see a benefit in your Corsi Number as a result. Similarly, if you take a lot of faceoffs in your own zone, it should hurt your Corsi Number. This is one thing that can be adjusted for. In fact, objective NHL has done this.
The medal round of the Women’s Canada Cup has been played in Vancouver, British Columbia this weekend. For preliminary round results look here.
In the semi-finals, USA met Finland and Canada met Sweden. Here are the results:
USA 4 Finland 0 This was a re-match of the only preliminary round game that USA lost. This time they had a better game. Meghan Duggan scored a hat trick for the US team, with Gigi Marvin assisting on each of the goals. Jocelyne Lamoureux added the other US goal. Jessie Vetter stopped 19 Finland shots for a shutout. Noora Ratty was in goal for Finland.
Canada 7 Sweden 2 The outcome of this game was never really in question. Canada took a 6-0 lead into the dressing room at the end of the second period. Emotions got a little out of control and a couple near brawls occurred. In the third period, Sweden managed to get on the scoreboard but it was too little too late. Hayley Wickenheiser and Sarah Vaillancourt each scored two goals for Canada. Kim St Pierre was the winning Canadian goalie. Isabelle Jordansson and Danileja Rundqvist scored the Swedish goals. Sara Grahn was in goal for the losing Swedish effort.
About The Puck Stops Here
Who am I? A diehard hockey fan.
Why am I blogging? I want to.
Why are you reading it? ???