Wednesday, July 24, 2013

Why the Jets shouldn’t be worried about Mark Scheifele’s Future Projection…Yet

Mark Scheifele has come under attack by some Winnipeg Jet’s fans as being a bust. He was drafted 7th in the first round in 2011, and two years later and now at age 20, he has yet to be a contributing factor for the jets playing a total of 11 games for 1 point.

But Jet’s fans should not be calling him a bust yet, as there have been plenty of players aged 20, even 21, that were drafted when they were 18 and have gone on to have successful NHL careers. The frequency table below includes all centres drafted in the first round between 1990 and 2010 and have played at least 10 game a season. Not surprisingly, the total number of centres beginning their NHL career spikes at 19, with teams giving their prospect one more year to grow in junior/minor hockey, something that is not uncommon. At age 22, the total number of centremen entering the league drops substantially from 113 at age 21 to 25 centremen at 22.The reason most likely being that these players have failed to show that they will produce in the NHL and teams beginning to give up. So, at the age of 20, Mark Scheifele is still in the age frame where he will have an NHL career.


But what about points? Can the jets fan expect him to be a point per game player? While every player is different, we can look at all centremen drafted in the first round since 1990 again and see how well their careers panned out in terms of points per game.

Below are graphs produced for centres drafted at 18 but start their careers at different ages, up to 23, since the age 24 only has 2 observations. What is clear is that players that start their career at ages 18 and 19 are generally superstars, with some players averaging over 1.5 points per game in some seasons. This trend dissipates as players get older, however. Since Scheifele is 20, we will focus on that age group for now.








While players that begin their career at age 20 do not score more than 1.5 points per game, the overall average is almost on par with those starting at 18, 19, and even 21 (shown in the final graph). These players also tend to have more consistent careers than those who start playing at age 18 or 19, to the point where around the age of 30, they are playing just as well as those who started at 18 or 19.

So, in conclusion, jets fans should not be too worried yet, there is still good history that says he will be a good NHL centre in terms of points per games, he may not be the knight in shining armor they were hoping for, but there is no doubt he could end up being a top 10 centre in the NHL one day.

Monday, July 22, 2013

Why do Shutouts in the NHL follow a normal distribution?

Recently I graphed the amount of shutouts with the amount of shots per game in anticipation to see a downward sloping curve. The less shots per game, the more shutouts. The more shots per game, the less shutouts. It made sense.

However, that is not what the data showed. What was found is that the amount of shutouts spiked just above of the average shots per game of 27 from season 1980 to 2013. Why would shutouts spike above the average shots per game though?



One theory I believe makes sense is that teams with less shots per game have invested in good defense, not necessarily spending big money for a superstar goaltender, and thus giving up goals with the defense not giving up a lot of shots. Teams that average a lot of shots (over 30) per game, invest in a very good goaltender instead of a top notch defense. Thus, giving up a lot of shots where even a great goaltender will have trouble keeping teams to a shutout.

This leaves us with the middle. Teams that invest in a good  goaltender and a good defense. Not great, not poor, but a good balance of defense and goaltending. These are the teams that are able to shutout their opponents. The defense keeps teams to enough shots in order for their goaltender to have a reasonable chance of getting a shutout. Of course, it is possible for teams to have a great defense and a great goaltender in the data with there being no cap pre 2004 and teams not paying quality defensemen and goaltenders market value in their early years,. However in the cap system we see now, this is more difficult to get great goaltending AND a great overall defense. Teams need to find a good balance of defense and goaltending, as relying on one or the other to win games may not be a solid plan.

An example of this system is the St. Louis Blues in 2011-2012. They had a decent goaltender in Brian Elliot, with a very good defense as well. They averaged 25 shot against per game, but Elliot was still able to put up 9 shutouts.

If you have another theory about this, please feel free to comment on the story.

Thursday, July 18, 2013

Comparing Revenue and Win Percentage % in NFL Teams

It is often debated whether how well a team does actually reflects how well the team does in terms of income generated, or revenue. While this is a simple analysis using NFL revenue data for 2010 from and 2011 from the Forbes NFL valuation list, it can help us understand the relationship.


As the graph shows, there does appear to be a positive relationship between the two, although it a weak one. The better a team does, the more revenue the team reports following the season. In the future if demand is there, I will attempt to further this analysis for more than just NFL teams and for more years.

Wednesday, July 17, 2013

When does the average NHL Superstars maximize their Points per Game?

While every NHL player maxes out their talent at different stages of their career and life, it is nonetheless interesting to see when NHL superstars reach their max potential in terms of points per game. Here, I define superstar as averaging 1 point per game or greater in a single season.

Below, all superstars in any given season are graphed for the years 1980 to the past 2012/13 season. With no surprise, the amount of observations diminishes as points per game gets higher, and as the second graph shows, most those observation belong to… you guessed it, Wayne Gretzky (The second is not meant to necessarily compare Gretzky to Crosby in their performance, but to show how they fit in the graph as they are two of the most recognizable players ever).

According to the data, the average superstar reaches their maximum output around the age of 26, which is actually higher than Wayne Gretzky’s at the age of 22 who seeing a steady decline afterwards.





Because of the change in goals scored per game over the last couple of decades, I broke up the data into 3 categories: 1980's, 1990's, and post 2000. What can be seen is that the definition of superstar might be changing over time. Players are not scoring as much now, thus an offensive superstar might average only 0.7 points per game now, compared to over 1.0 in the 1980's or 1990's.

What is also interesting to see is that over the decades, the data is beginning to flatten out. In the 80's the data is very top heavy in the early ages, around 24. In the 90's it is pushed to around 26/26, and lately while it is again at around 26/27, there is no peak like we see in the other 2 graphs, possibly showing that players are not only more even in today's NHL, but that they are performing at their best well into their 30's, a big difference compared to the 1980's.






While not all superstars are the same, it will be interesting to see how the next couple of years treat Crosby. If he stays on course with the league superstar average, next year will be his greatest. If he follows the same course as Gretzky, his best years could be behind him, and of course if he follows the trend of the past decade, he could have many good years ahead of him. This begs the question, is the greatest of a player defined by how well he does in a single or couple of seasons, but by how constant his points per game are above the league average? 


Monday, July 15, 2013

Is Home Advantage Important for Success in the NHL?

Having home advantage, specifically in the playoffs, is seen as an important step towards winning the Stanley Cup. Having your home fans behind your back can provide an extra boost for your team. But do teams with a good home advantage succeed more?

To begin, home advantage here is defined as the percentage of home wins divided by percentage of away wins. So, a team with a big home advantage should have a value greater than 1, meaning they win more games at home than away. Teams with poor home advantage will have a value less than 1.

The following tables provide the home advantage data for the top 5 and lower 5 teams. In the past season, one of the worst teams in the league with a record of 16-25-7 had the best home advantage in the league, winning 3 times more games at home than on the road. In fact, of the top 5 only 1 of teams made the playoffs, Los Angeles. The others did rather poorly, finishing in the bottom of the league in terms of wins. 

Looking at the teams with the worst home advantage, 3 of the lower 5 made the playoffs with Edmonton and Carolina missing out, 2 teams that arguably had the potential to make the playoffs.




Looking at the past 5 years, on average 1.2 teams from the top 5 made the playoffs, while 2.6 from the bottom five made the playoffs. It might seem that home advantage is not all that important looking at this. When all observations are plotted, there is a clear downward trend, but one noticeable trend, is that all Stanley Cup champions (Not just those that made the playoffs) lie at or above 1.0 in terms home advantage other than the Boston Bruins in 2010. They also, unsurprisingly, lie above 1.0 in terms of points per game. The black box in the following graph surrounds every Stanley Cup champion over the past 23 years.


I quickly ran a regression with points per game as the dependent variables and the home advantage variable as the dependent variable. As expected, there is a negative significant relationship. The bigger your home advantage, the lower your points per game. That is, a 1 unit increase in your home advantage variable lowers your points per game by 0.09. While not too substantial, it is still something that should be taken into consideration.

The lesson here, is that it is important to have a nice balance between home wins and away wins, while being able to win at home with your fan base behind you when needed. That is, having a slight home advantage is ideal, with those with too much of a home advantage failing to win on the road and making any substantial progress towards the Stanley Cup.





Sunday, July 14, 2013

Height, Weight, and BMI of NHL Players

I recently created a number of graphs to see whether the term bigger in the saying players are getting bigger, faster, and stronger to see whether this is true in the current NHL as there appears to be a large influx of smaller young and fast players post 2004 lockout (and in response to new rules that include further restrictions to hits to the head) as oppose to larger, hard hitting players.

 What the data and graphs below show is that the average NHL players does weigh less, while the average height of a player appears to be stabilizing.

 The following graphs use data from 1960 to 2012-2013 season. The average weight and weight is calculated, along with the standard deviation. The third degree polynomial trend line is shown using the black dotted lines.

 The average weight of NHL players saw an upward trend until the beginning of the past decade. After the 2004 lockout, the average weight of players sees a dramatic drop as teams move away from big, hard hitting players.



 The average height on the other hand has kind of stabilized post 2004 lockout. As well, the spread of players is larger. There is now a wider range of short and tall players than in the past. Teams are drafting shorter, faster players but not at the expense of taller (most likely defensive) players.


While the average human being is now taller (and bigger) than those in the past, it is of larger magnitude, with the average human measuring x in xxxx and x in xxxx, compared to around 71 inches (x feet x inches) and over 73 inches (x feet x inches) in 2012.




 Combing the height and weight of players to get the average BMI, it is shown that it has significantly dropped in recent years. Falling from 26.9 in 1996 to 25.4 in the past year.





 This provides some evidence on how the NHL has changed in past decade, and possibly where it is going in the future. The new rules put in place post lockout has changed the type of players teams are looking for, now going to less heavy (and most likely hard hitting) players for smaller (faster) players.