EPTFINALLY! - The 2013 NHL Season - Page 13
| Forum Index > Sports |
|
57 Corvette
Canada5941 Posts
| ||
|
Flaccid
8917 Posts
| ||
|
FaCE_1
Canada6184 Posts
| ||
|
Flaccid
8917 Posts
These player-usage graphs by Robert Vollman are really fun to look at: ![[image loading]](http://nhlnumbers.com/uploads/Image/toppairing.jpg) To explain what's going on there: You have zone start on the bottom - the further to the right, the easier the zone start this particular player gets. A player with a high offensive zone start is going to have an easier time doing good things than someone always forced to start from his own end. On the left, you have quality of competition - obviously measuring the opposition this player routinely faces. It goes without saying that if you're playing against the opponents weakest set, you're going to be well fed. The bubbles represent shot differential, blue being positive, white being negative. A blue bubble means a player is actively pushing offense in his team's favor. A bigger bubble means a bigger differential. This differential is often the most measurable, stable and consistent way of seeing a player's overall effectiveness, long-term. So what does this all mean? It means that a player in the top-left with a big-ass blue bubble is a robot sent from the future to murder John Connor and win Norris Trophies. For obvious reasons, it's vacant. A player in the bottom right with a white bubble would be a horrible NHL defender, and so on. Now look at PK Subban - pushing positive offense with a below-average zone start against stronger-than-average competition. That's a hell of a player right there. Were any Predator fans wishing and praying that the team would deal Ryan Suter? According to the math, Sutter and Subban are similarly effective hockey players for their respective teams. Oh and PK is 23 years old, busing doing stuff that no other 23 year old NHL defender is doing not named Ryan McDonagh. And he'll be doing it effectively for at least another 10 years. You gotta lock that shit down. Young defenders of his ability are as common as a modern playoff game in Toronto. But this is hockey. And it is Canada. Someone has to be the whipping boy for each franchise. And Montreal loves to run its best player out of town. edit: What this graph also tells us is that it sucks to be Eric Brewer. Damn. Worst zone starts on the planet against the oppositions top lines. That's a tough assignment. edit 2: And if Ryan McDonagh keeps this up, wow is he ever gonna get paid. | ||
|
Flaccid
8917 Posts
![[image loading]](http://i.imgur.com/r49Ac0B.jpg) Holy shit, right? Just look at that barren, post-apocalyptic wasteland about the x-axis. The Habs got slaughtered when facing anything worth calling 'above-average' competition. The only guy on the team able to push positive offense while facing the toughest comp? One PK Subban. And this is the guy Montreal wants to play scapegoat with. Unbelievable. Trade him to the Avs where he can go win some cups. Also hilarious, check out Scott Gomez. 10% better zone-start than any other offensive player while up against weak competition. Just goes to show how hard of a push the coaching staff was giving him while he proceeded to lay an egg of a season. Brutal. | ||
|
QuanticHawk
United States32154 Posts
Man I really hope Slats finds a way to get McD on a steal of a deal like Montreal did to PK. That 2 year still make shim an RFA when it is up too, correct? Fuck. I guess I can't complain because PK will be McDonaugh's comparable. Bouwmeester being highly ranked by that is odd. I'll take a look in a bit since I should have some free time, but there was a real cool article up reently using the advanced stats proving how much PK owns. Though that graph might be from it! ROFL at the Gomez chart. hahahaha | ||
|
Sub40APM
6336 Posts
| ||
|
QuanticHawk
United States32154 Posts
On January 31 2013 05:46 Sub40APM wrote: So under this chart Tyutin is a better defenseman than Lidstrom because he plays the same level of opponent but gets half as many offensive zone starts? And Nikitin is better than Chara. Seems legit and totally believable. it's context on a player, not the end all be all of the discussion. All it says is that Lidstrom got way more cupcake zone starts than Tyutin, and they had roughly the same shot differential while facing similar competition. In that instance, it explains to you why Tyutin's +/- is low, and why Lidstrom's points are higher it just tells you much more about what scenarios a player is being used in that you could ever see by watching or looking at a traditional stat sheet. | ||
|
Zorkmid
4410 Posts
| ||
|
Sub40APM
6336 Posts
On January 31 2013 06:05 QuanticHawk wrote: it's context on a player, not the end all be all of the discussion. All it says is that Lidstrom got way more cupcake zone starts than Tyutin, and they had roughly the same shot differential while facing similar competition. In that instance, it explains to you why Tyutin's +/- is low, and why Lidstrom's points are higher it just tells you much more about what scenarios a player is being used in that you could ever see by watching or looking at a traditional stat sheet. If that was the case it would be easy to normalize it. Put Lidstrom next to another Red Wing defender who shared his starts -- like Rafalski before he retired or White last year -- and compare. Ditto with the Columbus guys. But dont show me one chart that says something to the effect that "If Tyutin was on the Red Wings he would be nearly as good as Lidstrom" and expect a normal hockey fan to say "Of course, math!" | ||
|
QuanticHawk
United States32154 Posts
| ||
|
Flaccid
8917 Posts
On January 31 2013 05:46 Sub40APM wrote: So under this chart Tyutin is a better defenseman than Lidstrom because he plays the same level of opponent but gets half as many offensive zone starts? ... Seems legit and totally believable. Look at it for what it is - the 2011-2012 NHL season. By Nik Lidstrom standards, Nik Lidstrom didn't perform up to par in 2011-2012. He only got 30-some points where he typically gets 60-70. He also retired. By his own admission he was not the effective hockey player anymore that we all remember him as. Look at the same style of data for a year where Nik Lidstrom had a more Lidstrom-esque year, and you'd see his bubble significantly larger and the comparison wouldn't be made. Still, it's Lidstrom and he's amazing. He dominated against the tough competition. He just didn't dominate them quite as much as he might have used to ;-) Don't cherry-pick a comparison where Nik Lidstrom had a sub-par year and make the case that the data is shit just because Lidstrom is historically great. That perception is meaningless and why these kinds of unbiased stats are useful. What the data says, specifically is that Tyutin, while given less of an offensive push than Lidstrom, was able to cover the spread. By account of being a top-pairing guy on an awful team, he did quite well. The beauty of shot differential is that it helps remove the factors of shitty goal tending and superior scoring power. So by this metric Columbus Blue Jackets actually have the ability to look like decent hockey players. Go figure. It also tells us that the Red Wings are a far better hockey team. Tyutin had shitty zone starts because the Blue Jackets don't know what the other end of the rink looks like. Lidstrom got a good zone start against tough competition on a very good team as an 80 year old. And look, he did quite well. Like any data, it warrants a reasonable interpretation, not a "oh this part doesn't immediately make sense so let me dismiss it in its entirety" reaction =P. Hockey fans are some of the most throw it at the net, see what happens; saw him good; 110%; lacks hustle bunch of mofos out there and for whatever reason tend to shit a brick whenever people start looking at the game from a different angle. Hockey, moreso than any other professional sport, is influenced by luck and chance. As such, the statistics form less of a complete whole than they might in other sports. Still, they tell an interesting story and an important one at that - much better than we can hope to perceive through our natural bias and limited ability to see everything. I mean, I typically get lectured on how stats are dumb by people who then bring up the plus-minus. The bloody plus-minus. The stupidest metric in sports. | ||
|
Sub40APM
6336 Posts
| ||
|
Flaccid
8917 Posts
![[image loading]](http://i.imgur.com/0md8VkN.jpg) That shows how well the Detroit Redwings control a hockey game. Holy shit. Their zone-start graph starts at like 53%. That's insane. Not Lidstrom's fault that his team was too good to put him in a position to over-achieve ;-). It's not so much about comparing two players, it's about showing the effectiveness of players based on the roles in which they are placed. For example, we all know Ovechkin was awful last year, but do we know just how awful? Dude received a Gomez-ian sized push and still wasn't able to push the puck forward. And if you want to compare players, here are the graphs for each team: hockeyabstract.com/testimonials/playerusagecharts2011-12/Player Usage Charts 2011-12.pdf Kaberle probably won't tell you anything useful since he bounced around a bit. Overlay them or whatever, but read the analysis and put things in perspective. I'm not gonna do the work for you just so that you can be like "looooooook t3h mathz are wr0ngzzzzz put moaR pukz on n3t!!!" or dismiss things entirely. I'm willing to waste time, but not that much ;-). | ||
|
Sub40APM
6336 Posts
So here you are making a judgment that someone closer to the top left -- like Tyutin is a superior player to someone on the top right, like Lidstrom. I am not even talking about 'intangibles' and 'leadership' and '110%' don cherry esque mysticism that the old guard uses to overhype certain kinds of players over others. The reason why I want to find players who moved around and their comparable performances is to see whether this type of analysis generates reasonable predictive power -- was Kabrle being mediocre on the Leafs in his last year mean something for him being mediocre on the powerful Bruins? Did Ian White magically learn how to be a better dman in SJ or was it the system? Why did Jay Bouwmeester become so awful in CG -- it cant be just the quality of opponent because he year he was moved the NW was just as incompetent as the SE -- and so forth And if this model does not have predictive powers but is primarily descriptive then all you can say is "Last year P.K. Subban was the best d-man on his team by a wide margin" which isnt the same as "P.K. Subban will be an all star" because maybe Markov or Diaz are going to start getting the positions he was taking and cutting into his points and toi. | ||
|
Sub40APM
6336 Posts
I cant imagine the Leafs owners being cool with another team showing up there, especially if its owned by someone who actually wants a successful franchise and not just a giant money maker that sucks in billions from the foolish citizens of Toronto. | ||
|
Flaccid
8917 Posts
| ||
|
Stratos_speAr
United States6959 Posts
Wild upset Blackhawks in a shootout! 2-2 after OT, Wild win SO 2-1. | ||
|
Essbee
Canada2371 Posts
On January 31 2013 05:07 Flaccid wrote: Well, if people like "the maths," then it's easy to suggest that 29 other NHL teams would gladly take Subban off of the Habs' hands. These player-usage graphs by Robert Vollman are really fun to look at: To explain what's going on there: You have zone start on the bottom - the further to the right, the easier the zone start this particular player gets. A player with a high offensive zone start is going to have an easier time doing good things than someone always forced to start from his own end. On the left, you have quality of competition - obviously measuring the opposition this player routinely faces. It goes without saying that if you're playing against the opponents weakest set, you're going to be well fed. The bubbles represent shot differential, blue being positive, white being negative. A blue bubble means a player is actively pushing offense in his team's favor. A bigger bubble means a bigger differential. This differential is often the most measurable, stable and consistent way of seeing a player's overall effectiveness, long-term. So what does this all mean? It means that a player in the top-left with a big-ass blue bubble is a robot sent from the future to murder John Connor and win Norris Trophies. For obvious reasons, it's vacant. A player in the bottom right with a white bubble would be a horrible NHL defender, and so on. Now look at PK Subban - pushing positive offense with a below-average zone start against stronger-than-average competition. That's a hell of a player right there. Were any Predator fans wishing and praying that the team would deal Ryan Suter? According to the math, Sutter and Subban are similarly effective hockey players for their respective teams. Oh and PK is 23 years old, busing doing stuff that no other 23 year old NHL defender is doing not named Ryan McDonagh. And he'll be doing it effectively for at least another 10 years. You gotta lock that shit down. Young defenders of his ability are as common as a modern playoff game in Toronto. But this is hockey. And it is Canada. Someone has to be the whipping boy for each franchise. And Montreal loves to run its best player out of town. edit: What this graph also tells us is that it sucks to be Eric Brewer. Damn. Worst zone starts on the planet against the oppositions top lines. That's a tough assignment. edit 2: And if Ryan McDonagh keeps this up, wow is he ever gonna get paid. Ekman-Larsson and Pietrangelo are both on very defensive teams. They are also both miles ahead of Subban. | ||
GTR
51637 Posts
| ||
| ||
