The Knights play against the Storm in the Elimination Final this evening. The winner goes through to a Semi Final series of upto three games against the Blades (the other Semi Final is already set to be between the Flyers and Rangers). We have entered the serious end of the season. As such, now is a good opportunity to review the regular season just concluded, from both an individual and a team perspective.
My Season - GP 16 G 2 A 3 Pts 5 +4 10/4/1/1 PIMs 14
I played my second entire season in defence, didn't miss a game and had no serious injury events. More than half of the season we have played with only three on defence. I have generally been on the first pairing and had more time on ice than my partners. My game has become much more physical, primarily crease clearing in the early season, graduating to board battles in the corners as time passed. A perhaps necessary consequence of developing my physical game has been an accumulation of penalties. I have also developed a tendency to join the rush and have skated coast to coast on a number of occasions. Apart from games where my health was suffering I have finished games skating stronger than at the start.
My decision making has improved through the season, particularly when without the puck. Examples include taking my man, covering the rush, controlling the gap and keeping them on the outside, maintaining coverage in front of the net, setting up on the blue line and general weak side play and have all been reasonable to good.
I have not been so consistent with the puck on my stick, still having a tendency to find the opposition D on the blue when trying to clear up the boards from our zone or to ice the puck when there isn't enough pressure to warrant this. More head-up play is probably my best tactic to adopt to counter these failings this late in the season.
The increased physical performance has come about due a much more coherent pre-season training schedule than last year, training through the mid-season break and intelligent dietary choices (especially on game day). My improved technique and reading of the play has come about through a combination of accumulating experience and making the majority of team training sessions. Personal and team training combined have led to improved confidence, smarter play and a good foundation to rely upon as we enter the playoffs.
The Knights - 4th on the ladder G 54 (3rd) PIMs 108 (2nd)* SA 245 (1st) GA 37 (4th)
*if Stewy's ridiculous game suspension for momentarily removing his helmet prior the handshake at the end of a game is discounted (it triggered a major shakeup amongst officials with the responsible referee losing his position as head ref), we come in at between 4th and 6th in the penalty rankings.
It has been a good season for the Knights, especially considering that our roster was depleted of Bacon and then Matt before the xmas break. We have a total of only eleven on the roster (plus Tommy the goalie) as we enter the playoffs, but are a fairly tight unit and usually field at least nine or ten skaters. By comparison, the Storm generally fields thirteen skaters (occasionally twelve) while the Blades and Flyers generally only have eight or nine.
The primary stat of significance to me as a defender is that we have allowed less shots on goal than any other team. We will need to continue to do this.
All in all, we have a competitive team who can push any of the other playoff teams and, puck gods smiling, beat them. Certainly the Blades aren't invulnerable should we make our way past the Storm this afternoon. IF we can keep healthy through the playoffs, we'll give them a shake.
Perhaps equally important, if we can keep the bulk of this team together for next season (and there are many signs that most if not all intend to return) and recruit a couple more useful rookies in the interim, then we will truly be a force to be reckoned with!
Let's go Knights!
Fridge
Showing posts with label analytics. Show all posts
Showing posts with label analytics. Show all posts
Sunday, 31 January 2016
Tuesday, 7 July 2015
Fuzzy Wheeler Stats
The other night I worked out a bunch of stats for the top two teams in our Div II competition. Sadly, it appears I failed to save the results of my calculations so I won't be able to record them for posterity unless I go through the tedious counting and calculating all over again. No doubt it would be simpler to do the second time around, but I still can't be bothered. So I'll post my statistically inspired ruminations in their more fuzzy form.
I was able to derive comparative team statistics of Shooting Percentage (Goals/Shots on Goal), Save Percentage (Goals Against/Shots on Goal Against) and a raw team Corsi Percentage (SOG/(SOG+SOGA). I only worked up the numbers for the top two teams because we are drawing clear ahead of the other two. Although the sample size is only eight games (320 minutes), the regular season is only 12 games so it's a fair sample for that purpose, albeit subject to a large range of apparent variablility and wild regressions.
For interest's sake, I also looked at how many points the players on the teams generate as a measure of 'assisting goals' - this last due to the ready availability of the stats and the different styles of play of the two teams. For all the statistical weaknesses of my data, it is what I have so I made the most of it.
From what I recall, our shooting percentage dropped from 23% to 21% from game 7, Save Percentage went up a percentage point to 86%, Corsi remained about the same at 65%. About two thirds of our goals are 'assisted'. The Bumpers SP% was 21% (holding even), SV% was 91%, Corsi 57%. Less than a third of their goals were 'assisted'.
Until recently, the Big Question about hockey stats was whether they show something meaningful about the performance and style of the team/player analysed. That question has been answered fairly conclusively in the affirmative in the NHL world, and will no doubt permeate far and wide as a result. This small study is another affirmation of this answer, because what it shows does reflect the performance and stylistic difference between the two teams studied.
For example. After our most recent game against the Rockers it was interesting to hear what spectators and our opponents thought, as well as what our own opinions were. It seemed to me that we basically weathered the early storm, suppressed shots and shot attempts through the game, increasingly played our defensive game (ie. when we didn't have possession of the puck) up their end of the rink, passed the puck around a lot, combined to outfox the goalie and score our goals. I was told by both observers and opponents that they rarely had uncontested possession, couldn't get shots off, and we played an 'ice hockey style of game' with a lot of (passing) 'triangles' making it hard to get the puck off of us. In other words, we seem to play an aggressive as well as attractive team game and don't rely on any one player for options. Which is pretty cool for a Div II inline hockey team at the bottom of the world!
The Key stat, of course, is WINS. The Bumpers have 7 and we have 6. They have a total of 43 goals to our 40, and we have two more goals against than they do. Thus, we need to beat them by three or more goals next week if we want to take top spot.
It looks like being a great game.
I was able to derive comparative team statistics of Shooting Percentage (Goals/Shots on Goal), Save Percentage (Goals Against/Shots on Goal Against) and a raw team Corsi Percentage (SOG/(SOG+SOGA). I only worked up the numbers for the top two teams because we are drawing clear ahead of the other two. Although the sample size is only eight games (320 minutes), the regular season is only 12 games so it's a fair sample for that purpose, albeit subject to a large range of apparent variablility and wild regressions.
For interest's sake, I also looked at how many points the players on the teams generate as a measure of 'assisting goals' - this last due to the ready availability of the stats and the different styles of play of the two teams. For all the statistical weaknesses of my data, it is what I have so I made the most of it.
From what I recall, our shooting percentage dropped from 23% to 21% from game 7, Save Percentage went up a percentage point to 86%, Corsi remained about the same at 65%. About two thirds of our goals are 'assisted'. The Bumpers SP% was 21% (holding even), SV% was 91%, Corsi 57%. Less than a third of their goals were 'assisted'.
Until recently, the Big Question about hockey stats was whether they show something meaningful about the performance and style of the team/player analysed. That question has been answered fairly conclusively in the affirmative in the NHL world, and will no doubt permeate far and wide as a result. This small study is another affirmation of this answer, because what it shows does reflect the performance and stylistic difference between the two teams studied.
For example. After our most recent game against the Rockers it was interesting to hear what spectators and our opponents thought, as well as what our own opinions were. It seemed to me that we basically weathered the early storm, suppressed shots and shot attempts through the game, increasingly played our defensive game (ie. when we didn't have possession of the puck) up their end of the rink, passed the puck around a lot, combined to outfox the goalie and score our goals. I was told by both observers and opponents that they rarely had uncontested possession, couldn't get shots off, and we played an 'ice hockey style of game' with a lot of (passing) 'triangles' making it hard to get the puck off of us. In other words, we seem to play an aggressive as well as attractive team game and don't rely on any one player for options. Which is pretty cool for a Div II inline hockey team at the bottom of the world!
The Key stat, of course, is WINS. The Bumpers have 7 and we have 6. They have a total of 43 goals to our 40, and we have two more goals against than they do. Thus, we need to beat them by three or more goals next week if we want to take top spot.
It looks like being a great game.
Tuesday, 14 April 2015
A year of shooting shots
Close on a year ago (15/4/14, to be precise) I posted some results and analysis from my first 2,250 shots down at the primary school. I hadn't yet standardised my shooting drill, so the data included a large variety of shooting scenarios - 15' and 30', moving and stationary, set shots and rebounds. There's no way now of teasing out stats that are directly comparable to what I've been keeping records on since (ie. almost exclusively 'circuits' of 100 stationary shots from 15', divided equally into 'sets' of 25 shots each, being snap wrist shots, wristers, forward facing snaps and backhands).
In the last year I have routinised my shooting drills, and since July 13, 2014 have kept stats on 4000 shots from comparable drills. The bulk have been in this calendar year (in the last three months, actually). 4000 is a good number to work with, giving me 1000 each of the four types of shot that I've been using.
In between April and July I had a handful of sessions down the school, but they were sessions involving a wide variety of techniques (including slapshots!), I won't use those figures in this analysis. Except to say that I've probably put about 7000 shots on target at my shooting range since the project commenced.
So, without further ado, I'll now summarise the results of my shooting expeditions, and perhaps draw out a few trends if they are evident on the numbers...
13/7/14 4/9/7/6 - 26
19/7 11/9/9/5 - 34
20/7 4/8/9/7 - 28
13/9 5/9/9/5 - 28
20/9 4/3/5/1 - 13
3/10 9/4/4/3, 6/4/4/3 - 20, 17
5/10 6/4/9/7, 9/8/7/1 - 26, 25
21/11 5/4/8/3 - 20
24/1/15 6/8/5/3 - 22
25/1 5/8/8/4 - 25
18/2 5/6/5/3 - 19
3/3 8/8/7/2 - 25
12/3 7/8/8/5 - 28
13/3 5/4/9/3, 6/9/7/1 - 21, 26
17/3 8/2/9/1, 8/10/6/2 - 20, 26
19/3 1/6/7/4, 6/7/9/7 - 18, 29
21/3 7/3/10/5, 7/10/9/4, 7/11/9/4 - 25, 30, 31
22/3 6/7/8/5, 7/10/9/6, 5/5/11/2 - 26, 32, 23
24/3 11/11/5/3, 10/10/9/1, 10/14/7/2 - 30, 30, 33
27/3 4/6/9/6, 7/12/5/5, 7/7/7/5 - 32, 29, 26
29/3 6/6/9/3, 5/5/15/2, 6/9/3/5 - 24,27,23
11/4 7/10/12/6, 8/8/6/3 - 35, 25
13/4 6/6/8/5, 9/9/6/3 - 25, 27
Highest % in 100 shots - 35% 11/4/15
Lowest % in 100 shots - 13% 20/9/14
Highest Forehand Snap - 11/25 19/7/14, 24/3/15
Highest Wristshot - 14/25 24/3/15
Highest Forward Snap - 15/25 29/3/15
Highest Backhand - 7/25 several occasions
Overall shot % - 25.725%
Forehand Snap - 26.3%
Wristshot - 29.8%
Forward Snap - 30.6%
Backhand - 14.8%
First Thousand - 23.7%
Second Thousand - 23%
Third Thousand - 28.9%
Fourth Thousand - 27.3%
Make of that what you will.
In the last year I have routinised my shooting drills, and since July 13, 2014 have kept stats on 4000 shots from comparable drills. The bulk have been in this calendar year (in the last three months, actually). 4000 is a good number to work with, giving me 1000 each of the four types of shot that I've been using.
In between April and July I had a handful of sessions down the school, but they were sessions involving a wide variety of techniques (including slapshots!), I won't use those figures in this analysis. Except to say that I've probably put about 7000 shots on target at my shooting range since the project commenced.
So, without further ado, I'll now summarise the results of my shooting expeditions, and perhaps draw out a few trends if they are evident on the numbers...
13/7/14 4/9/7/6 - 26
19/7 11/9/9/5 - 34
20/7 4/8/9/7 - 28
13/9 5/9/9/5 - 28
20/9 4/3/5/1 - 13
3/10 9/4/4/3, 6/4/4/3 - 20, 17
5/10 6/4/9/7, 9/8/7/1 - 26, 25
21/11 5/4/8/3 - 20
24/1/15 6/8/5/3 - 22
25/1 5/8/8/4 - 25
18/2 5/6/5/3 - 19
3/3 8/8/7/2 - 25
12/3 7/8/8/5 - 28
13/3 5/4/9/3, 6/9/7/1 - 21, 26
17/3 8/2/9/1, 8/10/6/2 - 20, 26
19/3 1/6/7/4, 6/7/9/7 - 18, 29
21/3 7/3/10/5, 7/10/9/4, 7/11/9/4 - 25, 30, 31
22/3 6/7/8/5, 7/10/9/6, 5/5/11/2 - 26, 32, 23
24/3 11/11/5/3, 10/10/9/1, 10/14/7/2 - 30, 30, 33
27/3 4/6/9/6, 7/12/5/5, 7/7/7/5 - 32, 29, 26
29/3 6/6/9/3, 5/5/15/2, 6/9/3/5 - 24,27,23
11/4 7/10/12/6, 8/8/6/3 - 35, 25
13/4 6/6/8/5, 9/9/6/3 - 25, 27
Highest % in 100 shots - 35% 11/4/15
Lowest % in 100 shots - 13% 20/9/14
Highest Forehand Snap - 11/25 19/7/14, 24/3/15
Highest Wristshot - 14/25 24/3/15
Highest Forward Snap - 15/25 29/3/15
Highest Backhand - 7/25 several occasions
Overall shot % - 25.725%
Forehand Snap - 26.3%
Wristshot - 29.8%
Forward Snap - 30.6%
Backhand - 14.8%
First Thousand - 23.7%
Second Thousand - 23%
Third Thousand - 28.9%
Fourth Thousand - 27.3%
Make of that what you will.
Tuesday, 15 April 2014
2,250 shots and counting...
In a period of 24 days I had 10 sessions involving shooting. It was all on inlines at the local schoolyard, and is the only concentrated shooting practice I've had. I counted 2,250 individual shots during that time, of which 502 hit the the 'goal' (a netball goalpost about 2" wide) on the shot and another 35 on the rebound (in the 400 dynamic plays). A gross shooting percentage of 24.0%.
The only stats I kept on my first 100 shots on 21 March were distance (15') and on target percentage (22%). So, I guess I've improved!
In sessions 2-6 I shot at 35%, 19%, 19%, 31%, and 14%. Overall, shot at 21.7%. All of these sessions include 50 shots at 30' (total of 250 shots at that distance. My accuracy at 30'was 17.2%. Session 6 totaled only 150 shots and was a really low motivation day.
In Session 7 I introduced the 'new routine' of 100 each of static and dynamic at 15', concentrating on shooting mechanics for the static and integration with skating in the dynamic exercises. In sessions 7-10 I shot at 19%, 29%, 34% and 34%. If we include first time rebounds, these figures become 29%, 33%, 39% and 40%. Shows how important rebounds are to elevate relative scoring results!
My dynamic shooting percentage at 15' was 30% on goal on the initial shot. 12.5% of missed shots went in on the rebound. Which is a grand total of 39% shots/rebounds. Accuracy of type of shot/distance for remaining categories were: 15' Snapshot - 29%, Wrist shot - 26.1%, Full Face (snap) - 26%, and, Backhand - 15.1%. My 250 wrist shots from 30' were at 17.2% accuracy.
I think the sample size is large enough that the above figures mean something. I'm not sure exactly what that something is - beyond the fact that I appear to be getting more accurate despite the increased complexity of the exercises. When I mix this thought with the subjective knowledge that my technique, speed, power and stick-handling are also massively improving, I think it is safe to say that I am getting better and furthering my ambition to become a better player.
Which makes it all worthwhile.
The only stats I kept on my first 100 shots on 21 March were distance (15') and on target percentage (22%). So, I guess I've improved!
In sessions 2-6 I shot at 35%, 19%, 19%, 31%, and 14%. Overall, shot at 21.7%. All of these sessions include 50 shots at 30' (total of 250 shots at that distance. My accuracy at 30'was 17.2%. Session 6 totaled only 150 shots and was a really low motivation day.
In Session 7 I introduced the 'new routine' of 100 each of static and dynamic at 15', concentrating on shooting mechanics for the static and integration with skating in the dynamic exercises. In sessions 7-10 I shot at 19%, 29%, 34% and 34%. If we include first time rebounds, these figures become 29%, 33%, 39% and 40%. Shows how important rebounds are to elevate relative scoring results!
My dynamic shooting percentage at 15' was 30% on goal on the initial shot. 12.5% of missed shots went in on the rebound. Which is a grand total of 39% shots/rebounds. Accuracy of type of shot/distance for remaining categories were: 15' Snapshot - 29%, Wrist shot - 26.1%, Full Face (snap) - 26%, and, Backhand - 15.1%. My 250 wrist shots from 30' were at 17.2% accuracy.
I think the sample size is large enough that the above figures mean something. I'm not sure exactly what that something is - beyond the fact that I appear to be getting more accurate despite the increased complexity of the exercises. When I mix this thought with the subjective knowledge that my technique, speed, power and stick-handling are also massively improving, I think it is safe to say that I am getting better and furthering my ambition to become a better player.
Which makes it all worthwhile.
Thursday, 3 April 2014
Analytics - What makes a great player?
While surfing the outer region of the Habs' hockeysphere, in a 'fanpost' seeking to crowdsource raw passing data from NHL games for an amateur advanced statistics project, I came across references to the book/concept of 'Soccernomics'. Which in turn led me into the region of soccermetrics. Which still seemed a little amateur.
So, knowing that the Moneyball Concept seems to be most developed in American sport, I drifted to the NFLsphere and discovered a nascent, monetised version of advanced analytic concepts at work in relation to american football, and signs of the self reflexive awareness of this intellectual movement - where it ties in to the growing movement of industrial level analytics in the 'real world' in general.
Of course, the academic psychologists would have had to have been amongst the first on the field, and it is just like them and their sense of humour to have used basketball as the subject of their seminal study applying the science to the art of the sport. Needless to say, and staying within the psychological zone, their sub-industry has grown in a fractal bloom (eg this meta-analysis of the field of perception/expertise). Which is getting ultimately to the core questions, what makes a great player?
Moving back into the world of the hockeysphere, I passed through the more analytic oriented form of "sabermetric" moneyball logic in its native sport of baseball,before ultimately returning to one of my favorite hockey analytics sites for a good read.
So, knowing that the Moneyball Concept seems to be most developed in American sport, I drifted to the NFLsphere and discovered a nascent, monetised version of advanced analytic concepts at work in relation to american football, and signs of the self reflexive awareness of this intellectual movement - where it ties in to the growing movement of industrial level analytics in the 'real world' in general.
Of course, the academic psychologists would have had to have been amongst the first on the field, and it is just like them and their sense of humour to have used basketball as the subject of their seminal study applying the science to the art of the sport. Needless to say, and staying within the psychological zone, their sub-industry has grown in a fractal bloom (eg this meta-analysis of the field of perception/expertise). Which is getting ultimately to the core questions, what makes a great player?
Moving back into the world of the hockeysphere, I passed through the more analytic oriented form of "sabermetric" moneyball logic in its native sport of baseball,before ultimately returning to one of my favorite hockey analytics sites for a good read.
Subscribe to:
Posts (Atom)