I understand that these 4 players are not the primary focus on the Ranger's rebuild, the prospect and picks are. However, these players are the most comparable, as they are on the same playing field. I plan on doing an in-depth analysis closer to the draft of the Ranger's future assets, or as I have learned this semester in financial accounting, the Ranger's accounts receivables. One thing to note about this term, which is very apt for hockey is the uncertainty presented associated with it. Picks and prospects mean nothing when they don't create results. So for now, let's focus on those currently in the NHL. Now that we have seen the two play with their new teams for roughly a week and a half, I want to break down their play. I am writing this on March, 1st in reaction to not only Vlad Namestnikov and Ryan Spooner's contributions to the Ranger's scrappy win against the also-rebuilding Canucks, but also in response to, as of 11:30pm, J.T. Miller's two assists in Tampa's game against the Stars. All of my analytics are from Natural Stat Trick. Perhaps it is due to my sourness regarding his success on a rival team, but I want to focus on J.T. Miller. Please keep in mind that Miller is playing with elite offensive talents in Tampa Bay, and ALL of the sample sizes so far are incredibly small. However, I felt it necessary to provide a follow-up on their play. In just two games, Miller has produced offensively in ways he showed flashes of in New York. As I mentioned in my previous post, I believe Corsi to be a good indicator of team's long-term success. His Corsi, 2 games into his Lightning it up (pun intended), is a staggering 12-or-so percentage points ahead of when he was a New York Ranger. I would also like to highlight this statistic within the context of where he is taking the majority of his faceoffs, the defensive zone. In his first two games he has taken just 23% of his faceoffs in the offensive zone versus 55% during his time with the Rangers this season.
It's Thursday and instead of finishing that blog post, I think it would be better if I start a new one using more up-to-date information. First off, after reviewing last week's post, I realized that it was not accomplishing or working towards the reason why I started this. As the goal of this blog is to advance or provide a unique perspective on hockey analytics, I plan on doing what I initially said I would do and what the title of the blog suggests, find a relationship between statistics and success in hockey. So, to change the focus for this week's blog, I will do a simple regression(relationship analysis) between a team's PDO and a team's success during the playoffs. Before I do so, I want to provide clarification on both measures. PDO, called SPSV%(according to Google), is the team's shooting and save percentages added together. It is considered a measure of how lucky a team is, with those 1.000 or above considered luckier. Stupidly, I initially was going to regress goals versus success before I realized that obviously teams who play more hockey, those who go to the Stanley Cup Final, are much more likley to score more goals, due not only to their success to that point but also because they will be playing more games than most teams. Therefore, it's a pretty obvious analysis. Just a note for PDO, I decided to that it won't be affected by the number of games a team plays. Although you can argue that the longer they played the more it would regress to the mean, teams have already played an 82-game season and any changes can be accounted to their playing in the playoffs. Don't argue with that, you know everything changes in the big dance. For the term "success," I am going to first go on the basis of how many games the team won because let's be honest, that's all anyone actually cares about... Regarding the images below, I am noticing an interesting thing. If you look at the row "Regression," you will notice that under F and under P, we have 9.05 and .009. Both of these numbers indicate that in fact, there likely is a relationship between the PDO and #wins. That relationship is listed under regression equation, C2= -89.35+95.85C1. However, if you look at model summary, you'll notice that our r-sq =39.27%. This statistic represents the percentage of the data that is explained by the regression equation. This is a low percentage. Concerned with this, I looked at the graph. As you can see with the plotted points, there already seems to be a non-random relationship among the data, with teams having a higher PDO also having more wins. This makes sense because as every hockey fan knows, and as proved by the LA Kings in 2012 and the Predators last year, some of the success that teams have in the playoffs will come down to luck. That is illustrated in this data set. Conclusions: There is a relationship, and there's definitely another equation out there that better represents this relationship. If nothing else, this affirmed what every says about the playoffs, you need to play well and get a little bit of luck on your side. Please provide suggestions if you think there are any statistics that might be particularly interesting, otherwise I will do a regression with another equation for the statistic and conduct another regression; however, while the ladder will stay consistent week-to-week, I plan to change it as I become more familiar with advanced regression techniques. Thank you so much for taking the time to read my blog and until next week, Jonathan. P.S. I hope you're loving the Rangers' rebuild as much as I am 🏒!
It's Thursday and instead of finishing that blog post, I think it would be better if I start a new one using more up-to-date information. First off, after reviewing last week's post, I realized that it was not accomplishing or working towards the reason why I started this. As the goal of this blog is to advance or provide a unique perspective on hockey analytics, I plan on doing what I initially said I would do and what the title of the blog suggests, find a relationship between statistics and success in hockey. So, to change the focus for this week's blog, I will do a simple regression(relationship analysis) between a team's PDO and a team's success during the playoffs. Before I do so, I want to provide clarification on both measures. PDO, called SPSV%(according to Google), is the team's shooting and save percentages added together. It is considered a measure of how lucky a team is, with those 1.000 or above considered luckier. Stupidly, I initially was going to regress goals versus success before I realized that obviously teams who play more hockey, those who go to the Stanley Cup Final, are much more likley to score more goals, due not only to their success to that point but also because they will be playing more games than most teams. Therefore, it's a pretty obvious analysis. Just a note for PDO, I decided to that it won't be affected by the number of games a team plays. Although you can argue that the longer they played the more it would regress to the mean, teams have already played an 82-game season and any changes can be accounted to their playing in the playoffs. Don't argue with that, you know everything changes in the big dance. For the term "success," I am going to first go on the basis of how many games the team won because let's be honest, that's all anyone actually cares about... Regarding the images below, I am noticing an interesting thing. If you look at the row "Regression," you will notice that under F and under P, we have 9.05 and .009. Both of these numbers indicate that in fact, there likely is a relationship between the PDO and #wins. That relationship is listed under regression equation, C2= -89.35+95.85C1. However, if you look at model summary, you'll notice that our r-sq =39.27%. This statistic represents the percentage of the data that is explained by the regression equation. This is a low percentage. Concerned with this, I looked at the graph. As you can see with the plotted points, there already seems to be a non-random relationship among the data, with teams having a higher PDO also having more wins. This makes sense because as every hockey fan knows, and as proved by the LA Kings in 2012 and the Predators last year, some of the success that teams have in the playoffs will come down to luck. That is illustrated in this data set. Conclusions: There is a relationship, and there's definitely another equation out there that better represents this relationship. If nothing else, this affirmed what every says about the playoffs, you need to play well and get a little bit of luck on your side. Please provide suggestions if you think there are any statistics that might be particularly interesting, otherwise I will do a regression with another equation for the statistic and conduct another regression; however, while the ladder will stay consistent week-to-week, I plan to change it as I become more familiar with advanced regression techniques. Thank you so much for taking the time to read my blog and until next week, Jonathan. P.S. I hope you're loving the Rangers' rebuild as much as I am 🏒!
Comments
Post a Comment