The Puck Stops Here
Next entry: Worst 20 Fenwick Numbers
Previous entry: NHL Pulls Out All Stops To Keep Balsillie Out
Top 20 Fenwick Numbers
by PuckStopsHere on 08/27/09 at 11:18 AM ET
Comments (18)
In my look at sabermetrics and hockey, I have recently introduced the Fenwick Number. The Fenwick Number is alternative to the Corsi Number. It is an attempt to show puck possession and track players similarly to +/- ratings, but with a smaller statistical uncertainty since it includes more events. Fenwick Number is essentially Corsi Number with blocked shots factored out. It is valuable to try these sabermetrics formulations from various starting assumptions because we want to be certain that any conclusions are not merely artefacts of a certain set of starting assumptions.
Today I present the top 20 player Fenwick Numbers from 2008/09 and compare them to the top 20 Corsi Numbers.
Rank | Player | Team | Fenwick Number | Rank in top 20 Corsis |
|---|---|---|---|---|
| 1 | Niklas Kronwall | Det | +303 | 4 |
| 2 | Johan Franzen | Det | +280 | 6 |
| 3 | Zach Parise | NJD | +256 | 13 |
| 4 | Pavel Datsyuk | Det | +255 | 1 |
| 5 | Eric Staal | Car | +252 | - |
| 6 | Dan Boyle | SJS | +247 | 20 |
| 7 | Brian Rafalski | Det | +244 | 2 |
| 8 | Henrik Zetterberg | Det | +242 | 7 |
| 9 | Alexander Ovechkin | Was | +229 | 3 |
| 10 | Martin Havlat | Chi | +227 | - |
| 11 | Nicklas Lidstrom | Det | +226 | 5 |
| 11 | Marian Hossa | Det | +226 | 7 |
| 13 | Mikael Samuelsson | Det | +225 | 14 |
| 14 | Joni Pitkanen | Car | +223 | - |
| 15 | Ryan Getzlaf | Ana | +220 | 11 |
| 16 | Brian Campbell | Chi | +213 | 12 |
| 17 | Duncan Keith | Chi | +208 | 19 |
| 18 | Jonathan Toews | Chi | +202 | 18 |
| 19 | David Moss | Cal | +200 | 10 |
| 20 | Ryane Clowe | SJS | +199 | - |
For the most part this is a shuffling of the Corsi top 20. Only four players appear in this Fenwick list that do not appear in the Corsi list. They are Eric Staal, Martin Havlat, Joni Pitkanen and Ryane Clowe. The four players from the top Corsi list who are missing on this list are Mike Green, Nicklas Backstrom, Corey Perry and Scott Gomez.
Since blocked shots are not included here and they are with Corsi, the main movers are players who either block a lot of shots (this moves them upward - as shot blocks are not counted as shots directed against them) and players who take a lot of shots that get blocked - typically players who take a lot of slapshots (this moves them downward). Since these numbers count all shots when a player in on the ice in 5 on 5 situation, this may be blocked shots by teammates instead of the player himself, although in most cases that would be a lesser factor.
I like the fact that the Corsi list had two of the three Hart Trophy candidates in the top three positions (Datsyuk and Ovechkin). They have been moved downward in the Fenwick rankings. It is Niklas Kronwall who leads this list as he was among the top shot blockers in Detroit (he finished second to Brad Stuart). There is no claim made that Kronwall is the best player in the NHL based on this, but he had a good season on a good team. Both benefited his Fenwick rating. The fact he is a shot blocker moves him up the list relative to his position on the Corsi list.
The Fenwick list gives a group of 20 players who succeeded well in their roles in the NHL last year and had good puck possession. They are all from teams that had good team Fenwick rankings. It is possible to do a similar adjustment as was done to Corsi rankings to attempt to remove team contributions.
Filed in: | The Puck Stops Here | Permalink
Tags: Alexander+Ovechkin, Niklas+Kronwall, Pavel+Datsyuk,
Comments
Whereas in baseball you can you sabremetrics to generate a virtual season whose results will closely match the actual season results, in hockey that is currently impossible to do.
Whereas in baseball you can USE sabremetrics to generate a virtual season whose results will closely match the actual season results, in hockey that is currently impossible to do.
Typo, sorry.
Posted by VooX from Behind the Bar in the Hasek Club Car on 08/27/09 at 03:12 PM ET
You are 100% wrong when you try to argue that sabermetrics is used to predict the results of individual games. In any given game, anything can happen. Anything that can be predicted reliably is something that is true over the long term.
In baseball, where sabermetrics started and are best understood, there is no math I can do whatsoever that will tell me that player X will strike out or hit a home run in his next at bat. Largely, what can be done is to look at what has happened in the past to determine how well the players and teams played. One wants to find undervalued players and avoid those who are overpriced. This is the basic philosophy behind Moneyball by Michael Lewis - which is written about Billy Beane of the Oakland Athletics (and is one of the important books of baseball sabermetrics). The object is to find mathematical tools that allow you to better analyze past results in order to better value the players in those games and then attempt to project them into the future. Projection into the future is at best a black art. A player who looks like he projects well can ruin his arm, knee or back in a freak injury and everything changes. There is no reliable way to predict individual games as betters attempt to do.
Bookies really don’t care to predict games anyway. They are interested in balancing the books. They want a situation where as much money is better for a team as is bet against them and adjust the odds to try to achieve this. Their odds should not be confused with anything sabermetric. Its the line where they expect to get half the people betting on either side of it. If we know that a big market like New York will get more bets on its behalf than a smaller market like Kansas City or Milwaukee, then bookies will set the line such that it is tougher on the New York team. Bookies make their money on the “vigorish” which is the extra money one pays in a losing bet or by having odds that favor the house (add up the Las Vegas odds for each team at the start of a season and typically there is a roughly 30-40% chance than none of the 30 teams in the league win the Stanley Cup).
Your claim that football is understood sabermetrically is wrong. It isn’t. There are too many players on the field doing too many things at once. It is less understood than hockey.
The understanding of hockey is imperfect. Lots of work needs to be done in the future if it is to ever catch up to where baseball is (if that is even possible). That said, studying where we are now is very valuable. If you do not have the background or interest to understand the posts, that is fine. Then they are not written for you. There are many reasons I write this blog, but to entertain VooX is definitely not one of them. If I entertain you, that is great, but its hardly a goal in my mind.
These posts are a resource for people who are interesting in hockey sabermetrics. I have many more at my old blog that you may check out if you wish. They address goalies, +/-, ranking careers of long retired players and other topics.
It is 100% wrong to say that sabermetrics is meant to predict the outcome of future games. That is one attempted application, but usually a failed one because on any given day anything can happen. All that can be reliably predicted is longer term trends and even they come with a lot of fudge factors. If a key player gets hurt or if a young player makes a big step forward everything changes. The best application for hockey sabermetrics is to make sense of what happened in the recent past and then attempt to use that to predict the future. It is an imprecise tool to predict the future and will be used alongside other tools, which basically come down to educated guessing.
Applying sabermetric methods to look at last season is probably the best application for them at this point. Most of your comment is backwards and wrong.
Posted by PuckStopsHere on 08/27/09 at 03:41 PM ET
It is valuable to try these sabermetrics formulations from various starting assumptions because we want to be certain that any conclusions are not merely artefacts of a certain set of starting assumptions.
That convoluted nonsense is all you need to know about TPSH.
As a hockey fan, ask yourself when it was that you last worried about how valuable “sabermetrics formulations” or the “artefacts” that might affect them are to you. Right. Never. Me either.
And by the way, knucklehead, it’s “artifact” not “artefact”. Geezus, what an idiot.
Posted by OlderThanChelios from Grand Rapids on 08/27/09 at 05:01 PM ET
Artifact and artefact are both correct spellings.
Artifact is the more common American spelling.
If you want to be be a knucklehead and flame somebody about spelling, make sure you got your facts right.
As a Canadian, I often use non-American spellings. If you really find this interesting, read this.
Posted by PuckStopsHere on 08/27/09 at 05:19 PM ET
First off TPSH, I applaud you for sacking up and defending yourself and your work. Now, on to the issues cause you have my interest now, not just my disdain.
Secondly, baseball (and football) are games composed of discrete “events”. The game stops then a single play occurs. The game stops then a single play occurs. Rinse and repeat. This style of game is perfect for statistical analysis.
Hockey is wonderful (and more perplexing to bookmakers) because it is a game that is composed of non-discrete events. The game is fluid and play only stops under specific circumstances.
This is one of the best articulations of the difference between the major sports. How would you account for this within your sabermetric analyses?
That said, studying where we are now is very valuable. If you do not have the background or interest to understand the posts, that is fine.
Articulate their value. You get hated on a lot, and I’m guilty, because you make either obvious claims (Babbles is a good coach for Team Canada) or lists that include a bunch of numbers then a narrative that says “he had a good season on a good team” about the top player on your list. Is that something anyone who watched Kronwall play this season would not be able to tell you? Along with a lot more specifics about when he played well and not, along with a handful of good reasons for those peaks and troughs.
But please do reply; I’m honestly interested.
Posted by Osrt on 08/27/09 at 05:47 PM ET
Artifact and artefact are both correct spellings.
Artifact is the more common American spelling.
I actually knew that jackass because I looked it up beforehand. So please tell us that you actually chose to use this version of the word…
artefact: noun, British spelling of artifact .
But, hey, anyone who spends every day, all day, screwing around with obtuse statistics to justify whatever BS postion he’s taking at that particular moment just might do that.
I stand by my original statement: Geezus, what an idiot.
Posted by OlderThanChelios from Grand Rapids on 08/27/09 at 06:57 PM ET
Dear “OlderthanChelios”
If you know both spellings were correct than why do you immediatly assume he should write the word in the american spelling? after all you’re the linguistic “geniuses” using “YO”, and “Y’all” and “ain’t” and “Geezus”
Anyhoo, this blog is about statistics, the guy who writes this likes statistics. I couldn’t care less about statistics, obviously, neither could you.
I’m interested by this guy’s love of statistics (which is why i read this blog), you think he’s an idiot… Then why are you here reading and reacting to something you don’t care about, or which you belief is flagrantly wrong? That’s just baffling to me.
(If I misspell, don’t worry, I’m Belgian, English isn’t my maternal language, I learned Dutch, French, English and even German in school. Looks like you didn’t even bother to learn English.)
Posted by fish on 08/28/09 at 02:54 AM ET
after all you’re the linguistic “geniuses” using “YO”, and “Y’all” and “ain’t” and “Geezus”
Well, you’re right one out of four times, fish. And I used Geezus instead of “Jesus” for obvius reasons.
Posted by OlderThanChelios from Grand Rapids on 08/28/09 at 10:43 AM ET
And it is equally “obvius” that Older_Than_Chelios lives in a glass house when it comes to spelling and typos…
To save time, Chelly, please be aware that TPSH (and the rest of us Canadians, for that matter) just might spell “colour” with a “u”, and “defence” with a “c” and “organisation” with an “s”. Of course, we also reserve the right to type “color”, “defense” and “organization” if the mood strikes us…
Posted by Matthew McCallum from Redding, California on 08/28/09 at 12:09 PM ET
...please be aware that TPSH (and the rest of us Canadians, for that matter) just might spell “colour” with a “u”...
No matter how TPSH spells out his nonsense, he’s producing (as John points out) the worst blog ever…at least in the history of KK. His “everything can be explained with numbers” silliness belongs over at Statistics Central where guys with pocket protectors spend their lives analyzing minutia.
Posted by OlderThanChelios from Grand Rapids on 08/28/09 at 01:46 PM ET
The idea that everything can be explained by numbers is nonsense. It’s nonsense that older than Chelios made up as a strawman to attack. It has no basis whatsoever in anything I have ever written.
Posted by PuckStopsHere on 08/28/09 at 01:48 PM ET
It has no basis whatsoever in anything I have ever written.
Right. Because you never write about how statistics explain everything we need to know about who’s the best and the worst at anything that’s hockey-related. You just never stop being an idiot, do you?
Posted by OlderThanChelios from Grand Rapids on 08/28/09 at 02:59 PM ET
Right. Because you never write about how statistics explain everything we need to know about who’s the best and the worst at anything that’s hockey-related.
That is entirely correct. I never do write that.
What I have done this summer is look at a few metrics that can be used in place of +/- (which I don’t claim is a way to find out who is the best or worst hockey player either) and see who ranks at the top or bottom of those lists. For the most part, this is done to explore the other metrics. I have drawn some conclusions from this but NEVER have I suggested statistics explain everything we need. I have many times said that they don’t explain everything (for example look at the sabermetrics and hockey post I frequently link to.
I think that many aspects of hockey can be better understood by sabermetric analysis and many cannot. I think it is valuable to try to understand where that line is drawn today and see if it is possible to advance the line so that a little more can be understood. I doubt that hockey can ever be fully understood with statistical analysis and that is a big part of its appeal. If it truly was so simple that everything could be solved it wouldn’t be worth our time to watch and discuss it.
Posted by PuckStopsHere on 08/28/09 at 03:08 PM ET
What I have done this summer is look at a few metrics that can be used in place of +/- ...
What you’ve actually done this summer is bore the pants off of every visitor to KK. I mean who can forget these 24 memorable posts from the past two months…
• Worst 20 Fenwick Numbers
• Top 20 Fenwick Numbers
• The Fenwick Number
• The Most Protected Player In The NHL
• Top 20 Offensive Zone Starters
• Top 20 Defensive Zone Starters
• Zone Starts
• Worst 20 Corsi Rates
• Top 20 Adjusted Corsi Rates
• Worst Corsi Rates
• Top Corsi Rates
• Jan Hejda’s +/- And Corsi
• Worst 20 Adjusted Corsi Ratings
• Bouwmeester And His Corsi
• Top 20 Adjusted Corsi Ratings
• Correlation Between Team Points, Corsi And +/- Ratings
• Team Corsi Numbers
• Corsi Numbers Worst 20
• How David Moss Got Such A Good Corsi
• Corsi Numbers Top 20
• Corsi Numbers
• Worst 20 Adjusted +/- (Rate Stat)
• Top 20 Adjusted +/- (Rate Stat)
• Worst 20 Adjusted +/-
I think “Jan Hejda’s +/- And Corsi” will live in the annals (or maybe in the anals) of hockey as a true testiment to your brilliance as a hockey blogger. I fully expect Ekturd to give you a call any day now begging you to bring your pocket protector and your calculator over to his little part of hockeydom. It really is where you belong.
Posted by OlderThanChelios from Grand Rapids on 08/28/09 at 05:08 PM ET
That’s relatively productive for a summer - given that I have had some time off and written about other stuff as well.
I hope you study it, the final exam on that material could be a killer 
Posted by PuckStopsHere on 08/28/09 at 05:24 PM ET
That’s relatively productive for a summer…
If by “productive” you mean wasting your time compiling 24 totally useless posts about minutia that’s completely irrelevant to the actual game of hockey, well, then I’d agree.
I hope you study it, the final exam on that material could be a killer
I’ll consider it a badge of honor when I flunk that exam. In fact, anyone who gets a grade of “D” or higher should be banned from ever watching hockey again.
Posted by OlderThanChelios from Grand Rapids on 08/28/09 at 06:59 PM ET
Add a Comment
Please limit embedded image or media size to 575 pixels wide.
Add your own avatar by joining Kukla's Korner, or logging in and uploading one in your member control panel.
Captchas bug you? Join KK or log in and you won't have to bother.
Commenting is not available in this weblog entry.
Most Recent Blog Posts
Missing Other League All Star Games
Top Defenceman So Far This Season
About The Puck Stops Here
The Puck Stops Here was founded during the 2004/05 lockout as a place to rant about hockey. The original site contains over 1000 posts, some of which were also published on FoxSports.com.
Who am I?
A diehard hockey fan.
Why am I blogging?
I want to.
Why are you reading it?
???
Email: .(JavaScript must be enabled to view this email address)
When learning from experts it’s best to learn personally from them, or from their blog. We can provide that with poker lessons blog, your home to learn poker personally.
Do you get shocked from the luck in the game of poker? Stop getting shocked and start being a Poker Shoker
Donate to Kukla’s Korner
TPSH is like a reverse-actuary. He uses “complex” statistical analysis to kill readers with boredom.
Baseball (and football) undergo sabremetrics and other stats crunching for a couple of reasons:
First and foremost, there is a heavy gambling industry. Bookmakers like to really understand who has an edge when determining odds. Giant casinos aren’t built by losing money. They like to be able to actually predict the result of the next baseball game with deadly accuracy.
Secondly, baseball (and football) are games composed of discrete “events”. The game stops then a single play occurs. The game stops then a single play occurs. Rinse and repeat. This style of game is perfect for statistical analysis.
Hockey is wonderful (and more perplexing to bookmakers) because it is a game that is composed of non-discrete events. The game is fluid and play only stops under specific circumstances. Line changes during play are unique and players have simultaneous duties on offense and defence, football and baseball don’t. This makes hockey a game not very well suited for meaningful statistical analysis for predicting outcomes or breaking down games.
For all your Corsi, Fenwick, or whatever numbers, you will have a extremely difficult time predicting the general events of any NHL game. Whereas in baseball you can you sabremetrics to generate a virtual season whose results will closely match the actual season results, in hockey that is currently impossible to do.
In my opinion, TPSH, most of your stat-based analysis is largely useless for hockey. Sabremetric analysis is meant to predict outcomes of games, not debate who had the best season last year. And these tools fail miserably when applied to hockey.
Posted by VooX from Behind the Bar in the Hasek Club Car on 08/27/09 at 03:10 PM ET