When Patrick Roy stepped down as Colorado Avalanche head coach, I predicted the Avs were in a tight place and would not get the man they would have picked as head coach had they started their search in April. They were left to pick coaches from a group of people who were unproven and had not already commit themselves fully for the upcoming season. That is exactly what happened. The Avs hired their new coach Jared Bednar a couple of days ago.
Bednar is not an NHL name. He was a hockey player but never made the NHL. He was never drafted and never signed an NHL contract. When he retired from his career in the ECHL he was named an assistant coach in South Carolina where he last played. After four years as an assistant he became their head coach and within two years won the Kelly Cup as ECHL champions - though it should be noted that he took over a top ECHL team and didn't build it as head coach.
He then left for the AHL. First he served as an assistant coach with the Abbotsford Heat and then as a head coach with the Peoria Rivermen. After being fired in Peoria, he joined the Columbus Blue Jackets organization as an assistant coach in Springfield and later as their head coach. When the team moved to Cleveland and became the Lake Erie Monsters, he remained head coach and was the coach of the Calder Cup winning team this year.
There isn't much hockey news in August. Evidence of this is the fact that we have all been waiting to see where a player with no NHL experience at all will sign. Jimmy Vesey is that player and he signed with the New York Rangers today.
Vesey was drafted by the Nashville Predators in 2012 but chose not to join them. He played at Harvard University for his four year college career. He won the Hobey Baker Trophy as NCAA MVP in his most recent season. When Vesey did not sign with Nashville he became an unrestricted free agent this summer, but not until August 15th. His rights were traded to Buffalo for a third round draft pick in June. He didn't sign in Buffalo either. Today we have learned that he will be a New York Ranger.
The question is how good is Vesey? As the MVP in the NCAA he is definitely a prospect. The last two winners before Vesey were Jack Eichel of Buffalo in 2015 and Johnny Gaudreau of Calgary in 2014. There is definite precedent for a Hobey Baker winner to be a star NHL player. However there are differences between those Hobey Baker winners and Vesey.
My summer sabermetrics series has turned to Tom Awad's goals versus threshold system. This is an attempt to score all players with a single number - the number of goals they were worth to their team above that of a replacement level player. I have already looked at the top 20 players in this system and want to look at worst players in this system today. These are position players who rank below a hypothetical replacement player yet nevertheless played in the NHL. I want to compare these players to the worst players by team and zone adjusted Corsi. Goals versus threshold will select players who did not score and posted poor +/- ratings as the worst players in the league. Corsi may pick players who score points but still have puck possession weakness. Any player who finishes badly in both systems is definitely one of the weakest players in the NHL.
Here are the worst 20 position players by goals versus threshold in 2015/16:
In my most recent sabermetrics and hockey posts I have been looking at Tom Awad's goals versus threshold system to rank players by the approximate number of goals they were worth to their team above the level of a replacement player. I argue that this system can be improved especially in its defensive rankings by incorporating Corsi rankings. Before getting to that, I want to fully establish what the current goals versus threshold system gives as results. I have already posted the top 20 position players in 2015/16. I think it is useful to also look at goaltenders and see how they rank relative to position players.
Let's start by listing the top 10 goalies in 2015/16 and discuss them and then let's look at where they rank among all players.
I have been writing about Tom Awad's goals versus threshold system to rank players by the number of goals they are worth to their team above a replacement level player. I listed the top 20 position players in 2015/16 to see the current state of the system. I argue that this system can be improved by improving the defensive parts of the system by using Corsi rankings. In order to show a problem I wrote about Brent Burns being second ranked in the league last season. I don't think there is any reasonable argument that Burns was the second best defenceman in the league - let alone the second best player. However Lakeside Mall disagreed and argues that Burns may actually be second best in the league. I think that claim is not particularly credible but I will offer another player who is clearly overrated by the goals versus threshold system.
Joe Pavelski was ranked fourth in the league. He posted 21.7 goals versus threshold. Pavelski tied for sixth in points in the league with 78 points. He is tied with Johnny Gaudreau and Blake Wheeler. His 38 goals are higher than any player in that tie. In fact only Patrick Kane and Jamie Benn had more goals and also more points than Pavelski. Pavelski was fifth in goals in the league with Alex Ovechkin and Vladimir Tarsenko also finishing ahead of him. Goals and points are important to the offensive portion of goals versus threshold. Pavelski ranks as the tenth best player in the offensive portion of GVT.
I have decided to look at Tom Awad's goals versus threshold method for my next sabermetrics posts. I want to try to improve the system by including Corsi results. I began by looking at the top 20 position players from 2015/16 by goals versus threshold. Not surprisingly Hart Trophy winner Patrick Kane leads the league with 28.2 GVT but Brent Burns i second at 23.5 GVT.
Brent Burns was not the second best player in the NHL last year by any reasonable objective measure. So why does goals versus threshold rank him so highly? Largely this is because Burns is a high scoring defenceman who played on a good team who had shootout success. Offensively, defencemen are rated more easily than forwards. This is sensible because forwards tend to score more than defencemen. Burns finished 11th in scoring in the NHL last year with 75 points. He was second among defencemen in scoring behind Erik Karlsson. This makes him a highly valuable offensive player in this system.
Defensively Burns gets credit for playing on a good team in San Jose. A large part of the defensive goals versus threshold comes from the shots allowed by a team versus the league average. San Jose allowed the second fewest shots per game. This is a team success. As the playing time leader on the Sharks, Burns is given more credit for this defensive success than any other player on the team. This isn't an accurate assessment. Marc-Edouard Vlasic who was second in San Jose ice time is definitely a better defensive player and should get more credit for team successes. This is where a more individualized Corsi based analysis would be better.
It is August and hockey news is very slow. My posting has slowed down a bit and has been mostly about sabermetrics but something has happened that is significant. Patrick Roy stepped down as Colorado Avalanche coach. This move is surprising because coaches do not quit in August but at the same time it fits with what we have seen of Patrick Roy's personality. It is quite possible that he could be fired this season and instead of risking that situation, he decided it would be a better move to quit. However if he wanted to quit the team it would have made a lot more sense to quit in April after Colorado missed the playoffs.
Roy coached the Avs for three years. In his first season of 2013/14 he won the Jack Adams Trophy as NHL coach of the year. He has not made the playoffs since that season. Roy has shown that he is a poor coach in technical matters. He isn't good at strategy, matching lines or the kinds of things many coaches specialize in. His strength is in motivating players. He was very good at getting his team fired up for their games. However after a couple years of missed playoffs, he seems to have lost some of his "fire".
The problem with a coach quitting in August is that the best coaches have already been hired. Other teams looking for new coaches have had their new coaches for a couple of months. If any of those coaches would have been your first choice to replace Roy, they are gone now. There are other coaches that you may have selected who have already accepted assistant coaching jobs or jobs in the AHL or Europe. That is not to argue that no coaches are available as there are always unemployed potential coaches, but they are likely not your first or even second choice had you have been looking in the spring.
I am writing my next hockey sabermetrics posts on goals versus threshold. This is a system developed by Tom Awad to rate all players with one number which alleges to show how many goals they were worth to their team above a replacement level player. The first step to look at this system is to look at the top players in this system in 2015/16. We want to see how well those results mirror the rankings we would give to players more subjectively. For now I am only looking at position players and not introducing goaltenders yet.
Here are the top 20 position players in 2015/16 by goals versus threshold:
I have spent most of the summer writing sabermetrics posts which have mostly been about Corsi and its adjustments to make it an individualized stat. While it does a good job of measuring puck possession, there is no question that puck possession is not the only facet to playing hockey. For example looking at the top players by team and zone adjusted Corsi we see the top three in 2015/16 are Drew Doughty, Patrice Bergeron and Hampus Lindholm. Although they are good players, nobody would take things seriously if they were the Hart Trophy nominees. While there may have been a hope at one time of making a useful single number player ranking starting from Corsi, we see that this is not going to happen. Players can have good seasons as Patrick Kane, Jamie Benn and Sidney Crosby did without having dominant puck possession and players can be puck possession leaders and not have as strong seasons as thee players.
My hope is to find one number to reliably show how well players played relative to one another. Perhaps the most meaningful mainstream attempt at this so far is Tom Awad's goals versus threshold system which I have written about in the past. This system is not without problems. Most notably it is hard to measure defence and this system attempts to do so using +/- ratings. I think Corsi would be a much more valuable attempt. Thus I want to look at the results of the current system and its weaknesses and try to improve it. So in the next little while I will be writing about goals versus threshold. The attempt of this system is to measure how many goals a player is worth to his team above the amount that a replacement level player would provide. It might be a more useful way to make individualized ratings for different players.
A few days ago I posted the worst 20 players by team and zone adjusted Corsi ratings in 2015/16. The highest scoring player on the list is Bobby Ryan of the Ottawa Senators. Ryan scored 56 points last year. He also had the fourth worst team and zone adjusted Corsi in the NHL at -166.3.
Ryan has participated in the 2015 NHL All Star Game and in the 2010 Olympics for Team USA. He is a good enough player that he can be selected for top level teams but this does not happen all the time. He was not selected to the 2014 US Olympic Team or the 2016 US World Cup team. He has been criticized by those who selected the teams as having a "low compete level". That is quite clearly an old school scouting statement. Generally it is the kind of thing that is criticized using analytics but in this case the statistics seem to support that kind of idea.
Ryan has a poor Corsi after adjustments. Ryan is one of the worst players on the Ottawa Senators by raw Corsi. There are three other Senators who finished worse than Ryan's -152. The problem is Ryan put up this low puck possession number with a Sens leading 171 excess offensive zone starts. Ryan often started his shifts in the offensive zone and despite that his team was usually on the defensive.
About The Puck Stops Here
Who am I? A diehard hockey fan.
Why am I blogging? I want to.
Why are you reading it? ???