Wednesday, October 31, 2007

Corsi Numbers

During the lockout Darcy Regier was a keen advocate of rule changes, you probably remember the groovy shaped nets that the Sabres organization were promoting. I heard Regier on the radio a few times then, he is one hell of an interesting guy. He was making compelling arguments using statistics I had never heard of; where the pucks went in, which percentages in which part of the net, number of goalposts and crossbars hit, and so on. Sabres goalie coach Jim Corsi also spoke on the radio, just once that I heard, a very technical guy with the demeanour of a scientist speaking on CBC radio, and surely the source of all the wonderful data that Regier was bandying about. In any case I gave him credit for it.

He also used "shots directed at net" as a measure of the activity in a game. I had no idea they were even tracking this stuff. Since then the NHL has added missed shots, blocked shots, goalposts and crossbars to their play by play sheets. They also show you who was on the ice for each event.

The Corsi numbers for the Oilers are to the side here. I don't include blocked shots here, it's the total number of missed shots, saved shots, goalposts, crossbars and goals directed at the opposition's net while a player was on the ice ... minus the same at his own end of the rink. These have been corrected for ice time, every Oiler's numbers have been prorated to Staios' EV ice time, because he's the Oiler with the most of it.

As a point of reference: Zetterberg is +91 by this exact metric, and looks poised to lead the league again by this measure, in spite of playing tough opp. Granted Babcock teams are always a little shot happy, surely that helps. Lidstrom will lead all defencemen at this, he's +60 already (after the Staios' ice time level correction) and probably by some distance, in terms of just shots +/- he's been a world beater ever since the NHL started publishing shift charts, probably longer. Really, in a fairer world he'd have won 10 Norris trophies in a row by now.

Unstoppably, player's results will gravitate towards the Corsi numbers as the season goes on. A few good bounces here and there can make the early stats misleading. To pick on an old favourite, Lupul is EV+/- +3, and Corsi -41. The drop is inevitable, just is.

As an aside: Scott Hannan is an abominable -60 by this measure. A country mile worse than any of his team mates. Matt was right, something is wrong with Scott.

And as always, common sense should be liberally applied. And as with all hockey stats, without applying the context (i.e. who these players were playing with and against) it will be of lesser value.

On the Oilers:
- Stoll is off the rails completely.
- Hemsky has arrived.
- Gilbert is just flat-out too good for school. Rookies should suffer at this, especially defenders. Imagine where he'd be if he wasn't playing tough minutes with Grebeshkov recently?
- Maybe Reasoner really is back on from, he has looked it.
- Penner and Pitkanen have managed to produce better Corsi numbers than my eyes tell me they should have. I'll trust the numbers.

42 Comments:

Blogger Kish said...

Too many good points (and too few bad ones) to comment on individually, so; great post.
The corsi list is pretty much how i would have ordered the oilers so far. if i were to rank Roli, it would definitely be in the bottom third. at best.

10/31/2007 2:56 pm  
Blogger PDO said...

This comment has been removed by the author.

10/31/2007 4:09 pm  
Blogger PDO said...

Is anyone else really surprised at how poor Tarnstrom's numbers are?

He's honestly looked pretty damn good in the games that he's played by my eye...


Vic, what steps do you go through to get these numbers? Ditto the EV+/- numbers? Do you have a program ripping the information from nhl.com, or is it all available at some nifty site developed by some marvelous bastard?

10/31/2007 4:11 pm  
Blogger Vic Ferrari said...

kish:

Yeah, Roloson hasn't been great. He'll bounce back though. I mean if Kipper keeps going as he is it will be one of the worst save% seasons ever by an NHL starter. And Biron is on track to kick the ever living shit out of Hasek's high water marks in that regard. Neither will last long.

As I'm in a mathy mood: If you look at the bonafide starting goalies in the league, guys who will play 50+ games and have some track record: sum up the differences between their save% now and the average of their last 3 or 4 years. Next week it will be smaller, and the same the week after, and so on through the year. I'd bet that almost every week, with maybe one aberration, they will collectively gravitate towards their career averages.

Some will have good years, some bad. Same thing happens with slots players ... i.e. take away the noise and there's not much left. EVsave% even stronger this way.

Roloson will come around I'm sure. He isn't getting any younger though, hopefully they spell him with Garon for 25 games or so.

10/31/2007 5:07 pm  
Blogger Vic Ferrari said...

PDO:

I'm a bit surprised by Tarnstrom's numbers too. Maybe it's coincidence, maybe I haven't been giving MacTavish enough credit on that count. I dunno.

As for the numbers, unfortunately nobody is publishing these, or at least not that I can find. As you suspect, i just scrape them off of NHL.com with some the the most simple, heavy handed, CPU-taxing php script writing on Earth. The hockeynumbers guy does have some good stuff along a similar vein though.

I can't imagine that a whole bunch of people would want to look at this sort of thing. But eventually I'll get around to leaning out the script a bit and putting it online.

10/31/2007 5:16 pm  
Blogger Vik said...

Good post. That's a nice little stat there. I wouldn't use it alone to judge anyone's performance though but with it and some of the other stats out there, statistics are at least starting to paint a bit of a picture for hockey.

10/31/2007 5:43 pm  
Blogger mc79hockey said...

How is that Zack Stortini is getting minutes? I'm kind of "meh" on the value of fighting anyway but when the guy is a lousy fighter and every metric says that he sucks...what's the point? Why are we doing this to ourselves?

Stoll is an interesting cat. I've never been his biggest fan, for largely aesthetic reasons but his performance has never really blown me away either. He's always struck me as a guy who can score at a so-so level playing soft minutes whose big advantages are on the PP and the PK. With the acquisitions of Souray and Joni, the Oilers kind of stole the spot he needs for what he does. How much do you pay a guy who doesn't fill a hole on the PP for you, brings what Stoll brings at ES and is a good PK guy? It's a headscratcher to me that he's much more valuable to the Oilers than Raffi Torres and yet the numbers I hear tossed around for him make Torres money look small. If I was Lowe, I don't know that I'd be eager to get Stoll at a big price tag, even before I started to wonder about his headaches.

10/31/2007 6:21 pm  
Blogger RiversQ said...

Very cool stuff. (Why are the blocked shots left out anyway?)

It does trend reasonably well with the results right now and as you say, that'll probably shake out and end up being much closer.

As far as results go, I've been dabbling with a modified Wilson number. Basically just to make it look better. If you go by EV Pts/hr divided by EVGA/hr and then just multiply by the ATOI, the Oilers shake out like this going into last night:

Hemsky 19
Horcoff 17
Brodziak 13
Torres 11
Cogliano 10
Gagner 9
Souray 8
Penner 7
Pitkanen 5
Nilsson 5
Smid 4
Tarnstrom 4
Gilbert 4
Reasoner 3
Staios 3
Sanderson 3
Greene 1
Stoll 0
Jacques 0
Roy 0
Grebeshkov 0
Pouliot 0
Schremp 0

As always, this is muddied by puck luck, the dmen get pounded and there are a few really small sample sets in there. However, for the guys who have played a fair amount, this jives reasonably well with your Corsi list.

Aside: Detroit really does throw a silly number of low quality shots at the net - the scoring chance gap last night was probably 2-3 in favour of Detroit while the shots were more than double. Detroit's a little skewed here for sure, but there's no doubting their quality, or Lidstrom's or Zetterberg's.

10/31/2007 10:08 pm  
Blogger Jonathan said...

I really, really wish that the NHL would step out of the dark ages in terms of statistical information made available to the general public.
I also wish that, failing the above, the mainstream media would make up for the shortfall by doing indepth statistical analysis on their own (as a happy byproduct, Al Strachan would no longer be able to find a job that pays him to talk about hockey).
Since neither of those are happening anytime soon, a very big thank you from me to Vic (and of course, the other number crunchers of the Oilers blogosphere)

11/01/2007 11:21 am  
Blogger MetroGnome said...

Very interesting metric.

11/01/2007 12:53 pm  
Blogger Vic Ferrari said...

mc:

By all accounts Stortini is a really nice guy, but it is impossible to defend him as a hockey player. Is anyone even trying (family aside)?

Here's my best effort for Zack, using an intersting stat, stemming for Riversq's comment:
The Oilers are more likely to block a shot, and less likely to get a shot blocked, when Zack is on the ice.

That's the best I can do. I won't defend it though.

.
.
.

CinOstyle-HalfFullGlassGuy sez:
"Zack's determination, grit, and all around goodguyness inspires his teammates to block shots! AND motivates them to get their shots through at the offensive end of the rink!"

CinOstyle-HalfEmptyGlassGuy sez:
"The Oilers are always in their own end of the rink when Stotini is out there".

11/01/2007 1:00 pm  
Blogger mc79hockey said...

I only noticed what "CinOstyle-HalfEmptyGlassGuy" had to say when I came down here to write that exact thing.

11/01/2007 4:02 pm  
Blogger Vic Ferrari said...

Riversq:

Yeah, just my own thinking that if you block shots you must be doing something right, because chances are you are at least back and in the shooting lanes.

I'm starting to think that my thinking on that is off. Given that every player that's effective at keeping the puck in the good end of the ice seems to have a lot of their own team's shots blocked, and few of the opposition's shots blocked.

So Iginla is on the ice for only 25 Flames shot-blocks. And Robert Nilsson plays less than a fifth of the EV icetime and is out there for 17 Oiler shot-blocks. Hmmm.

At the other end of the rink the Flames have their shots blocked 50 times when Iggy is on the ice, and the Oilers 8 when Nilsson is out there.

I probably should have done it that way in the first place. For what it's worth, most guys say near enough the same anyways. Especially on D, where only Greene moves rank, inexplicably he improves a bunch (I thought he was the shot blocker? Damn). Up front Pouliot, Jacques, and Penner roll up the rankings a bunch and Sanderson and Cogliano drop a lot.

11/01/2007 7:31 pm  
Blogger Slipper said...

This Corsi expression seems like a keen way of eyeing exeptionally strong or extraordinarily ineffective forwards. I think there's too many variables to make accurate assumptions on players who fall in the middle ground. In my opinion forwards would drive these results, and defenders, minus perhaps the rare exception (see: Lidstrom), would be passengers on the bus. By that I mean that I don't often see defencemen forecheck, pressure for turnovers in the neutral zone, or cycle down low

11/01/2007 8:27 pm  
Blogger Black Dog said...

Another goal tonight for Zetterburg - anyone want to make the case that he is the nest player in the league and the Wings the team to beat.

Vic and Riv - how about a post in the near future about the kids on the back end? I would be interested to hear your take on the youngsters. I really think Gilbert is the real deal. Grebeshkov had a few games there where he was making the big mistake but seems to be improving. Smid and Greene have improved, imo, although marginally.

Thoughts?

11/01/2007 8:35 pm  
Blogger Slipper said...

Regarding Lupul. I've been watching alot more hockey this year since Shaw got the CI package. When watching the Philly games I've noticed Lupul is usually the first man heading out on the rush on the right wing and the last man back when the puck sxwings. The Flyers use the LW side as their strong side when they're trying to exit their end, kind of like Detroit. So you rarely see Lupul fighting for the puck on the wall, an area where he failed so often as an Oiler last year. So obviously Lupul's still a floater and I can envision his EV +/- eventually plummeting, but I can also see his production being steady, as long as he's left in that position, and isn't taken to task for it.

The Oilers seem to use the RW boards as the dominant side and thier LW's lead the rush out of the zone. I think it's often why Smyth appeared to be cheating so much last year, but he didn't suffer much for it and that was probably because he had a better eye and sense for the handle of the puck his winger or defenceman had when he decided to take off.

I just thought I'd strike this up as conversation because it seems every team has a strong side in their own end that they rely on, no matter the line combos. This is the side they'll ring the puck around to, where the winger corrals the puck and either tries to simply get it the F out of the zone, make a pass to the centerman who waits in support, or skate it out and make the headman pass to the opposite wingers who's heading North. It makes sense that the side never changes because it would just creat confusion with the mix and matching of the forwards and defensive pairs.

11/01/2007 8:45 pm  
Blogger Matt said...

Vic, on the Blocked Shots thing, I think you were right the first time.

It sounds like from your 7:31 comment that not counting BlkShts "compresses the field" as a whole a bit. As a number that is designed to illuminate, that doesn't sound like too bad a thing.

The other way of looking at it is I guess that blocking shots, or not, or having your shots blocked, or not, is either flukey or it is not. While I'm sure there is an element of luck in there at times, half the point of the Corsi number is that it's a ready-made large sample size, no?

Since (1) the sample size is huge, and (2) it is self-evidently better to block a shot than not, or to get a shot to the net than have it blocked, then excluding blocked shots makes perfect sense. The alternative is to argue that, what, you're punishing guys who have a disproportionate number of their shots blocked over a long period of time? That sounds like something that should be punished in a metric like this.

For further common sense evidence of this, just look at your Oiler examples. Is there someone who wants to argue that Greene is more effective than he appears because he (and his partner, linemates) blocks relatively few shots?

11/01/2007 9:04 pm  
Blogger choppystride said...

Yeah, just my own thinking that if you block shots you must be doing something right, because chances are you are at least back and in the shooting lanes.

Yes, the blocker is perhaps doing something right, but it's harder to make the same argument for his other linemates.

IMHO, if a credit - or, in our case, the nullification of a demerit - is to be assigned for a blocked shot, it should only be given to the blocker himself. Theoretically, one or more of the defending players (perhaps even including the blocker) has committed a prior screw-up that lead to the shooting opportunity, which is very much like the scenarios for MS & SOG.

Of course, if one plays with uneven weighting, then all kinds of (perhaps interesting?) worms may be opened. For instance, in the converse case to the "shot blocker", should a "shot misser" be penalized for (theoretcially) wasting a scoring opp? Should all the weights be simply -1 and 1?

11/01/2007 9:16 pm  
Blogger Slipper said...

Isn't a blocked shot often a real low percentage shot in the first place? I guess it is indicative of puck possession. Then again, blocking a shot, even taken from 20-30 feet, is preventing a rebound chance, which is a higher percentage chance. Therefore, a shot blocker is preventing good scoring chances against.

11/01/2007 9:23 pm  
Blogger Dennis said...

Absolutely great stuff, Vic. It's been great what guys like you, Ty and Riv have come up with over the years.

Slipper, in thinking about who plays on the Oilers left side, doesn't it seem like 14's always the guy in charge of either chipping the puck from the wall or making a breakout pass? That's the way I remember things. But then I think about our other primary portsider, 27, and sometimes they go to his side and sometimes they dont.

11/01/2007 11:02 pm  
Blogger Slipper said...

Dennis: Since being paired with Hemsky and Horcoff, I've noticed Torres as being the one to first break into the neutral zone. That sets him up wither for the head man pass, or as the first forechecker in the case of a dump-in. If the break out is slow the LW stalls right at the attack zone blueline, along the boards. That line is a little exceptional compared to the others, as Hemsky or Horcoff can often retreive the puck and skate it out of their end; as opposed to the other crews who often get pinned deep and ring it around to the RW wall- 14-10-83 seemingly do not need to revert to that very often.

That's not to say that they never play it to the LW wall in their own end if the play dictates it. But when that's the case often it catches the weak side defenceman by surprise (Staios reversing it due to pressure and catching Smid out of position comes to mind), and often there isn't a winger waiting there either.

11/02/2007 12:07 am  
Blogger Slipper said...

Dennis: As an addendum, I feel I need to add that this was a thought that occured to me during the Oiler PPV amidst one of their player under the microscope segments. The player in question was Raffi, and I watched him bail the Oilers' zone as the puck was rung around "the far wall on your telly". He kind of cruised the neutral zone and when nothing immediate happened he positioned himself against the wall at the attack zone blue line. That's what got me thinking about Smyth (shooting for the North, just ask Vic) at the LW last season and Lupul being beaten on RW the half wall aswell, albeit it was two different lines.

In the same broadcast I decided to pay attention to how Detroit transitioned from their own zone, and from my perspective, unless extreme situations dictated otherwise, it was like clockwork: Datsyuk covered the LW wall, as Zetterberg bantied
(depending on the situation) between offering puck support or breaking into the neutral zone. Meanwhile, Holstrom had sprung himself deep. But that makes sense, you know? Holstorm is probabley the weakest skater amongst that set of forwards, he's also one of the league's most tenacious forecheckers, and if things go south in transtion, who would you prefer as the two forward behind the play? Datsyuk and Zetterberg, of course.

Apply that to the Oil. Even though it's not as sexy, the same principles apply. My only reasoning behind thinking that a team, when faced with real pressurw from the fore check, has a set principle or, better yet, a set approach to pressure, is because every team seems to rely on one simple play in their own zone. Ecspecially when things get harried. Either ring it out left or ring it out right.

Again Dennis, this is only a theory. It's somethng I arrived at and then continude to watch for, to either assert or debunk my opinion. Last night I watched the Habs vs. Flyers game. Lupul had no battles inside his own zone on the wall. He was the first forward out of the zone and the last man back. Though seemingly the rest of Philly's forward units seemes to migrate towards that trend: that the RW headed the rsuh and the LW was the responsible side.

Atleast it's something we can watch for and communicate about after tonight.

11/02/2007 7:32 am  
Blogger jadeddog said...

great post.... this is the type of stat i would absolutely LOVE to see online... where you could run the numbers for each team etc.

if you could do that, well, ill be very happy :)

11/02/2007 10:56 am  
Blogger Dennis said...

Yeah, I'll try and watch for that tonight, Slipper. It's an interesting theory and one worth checking out.

11/02/2007 1:08 pm  
Blogger namflashback said...

Slipper,

I think you are close in your analysis, but it isn't quite as static. Weak-side, strong-side is changeable, depending on which side of the ice they surrender to the opposition forecheckers.

The reason HHT has been working so well is that:
1) Torres on the d-boards rarely fails to make either
- chip out
- chip to Horc
- reverse to D to reset
2) Hemsky is getting a bit better on the boards to do the same off that side.

Since they are also getting Gilbershkov as their D-pair, this group seems to be able to key off of either side D. To my eye -- the two G-D-men will need a little more time to work their trap breaking play -- but seem to have a plethora of options ahead of them. That's a nice 5 player combination.

Although you are right that LD to RW seems to be the favourite tendency, all 5 players need to make themselves available for the option plays.

11/02/2007 2:56 pm  
Blogger Vic Ferrari said...

Great comments all over. I'm convinced that keeping blocked shots out is best, so that's what I'll do. I take choppystride's point about crediting the shot blocker individually, but that would detract from the simplicity of the thing.

To me it's a pretty solid indication of zone time and scoring chances at evens. Of course if your team can't bury their chances, that isn't necessarily enough. And if you can finish like Pavel Bure, well to my mind it's not quite as damning to have poor zone time and have let your goalie see a bunch of pucks.

11/02/2007 3:00 pm  
Blogger Vic Ferrari said...

slipper:

Very cool, I love that stuff. I haven't noticed that before, but I'll be keeping an eye out for it. Are you sure that teams just aren't tending towards moving the puck out of the zone from the board side that the player bench is on? I dunno, just thinking out loud, you do see that penalty box side winger way out ahead a lot, for all teams I think. Obviously they wouldn't (shouldn't) be up there if there is a puck battle on their boards.

When I think of the Oilers clearing from the sidewall, Hemsky and Pisani mental replays do come to mind the most. As well as some difficult shifts by Georges in that regard. Hmmm.

With Detroit I have replays in my head of them clearing the zone by putting the puck from the corners right back to the low slot, sticks with me because few teams dare to throw that play into the book.

11/02/2007 3:10 pm  
Blogger Vic Ferrari said...

jadeddog:

Try these links, some are for lists of game numbers, some are for the shots metric discussed in this thread. I'm sure you will be able to figure it out, change the variables in the URL to suit yourself:

http://timeonice.com/shots.php?gamenumber=20167
http://timeonice.com/xshots.php?first=20169&last=20171

http://timeonice.com/yesterday.php
http://timeonice.com/today.php
http://timeonice.com/byteam.php?team=NYI

11/02/2007 3:19 pm  
Blogger RiversQ said...

I'll stick to my guns on this one. I still think the blocked shots should be in.

Personally, I think the blocked shots are reflective of where the puck is and serves to increase the sample size, which is only a good thing Matt somehow used against the idea.

I don't care how well you block shots, or how well your teammates block shots, if you're spending all that time in your own end it'll probably catch up to you. Conversely, having your shots blocked a lot isn't totally a bad thing - it's hard to get scored on while the other team is stuck blocking your shots.

IMO, this measure has zero value in describing the quality of opportunities. Excluding blocked shots is almost like trying to impose some element of shot quality into the data. Like I said, JMO, but I think Corsi had good reason to have it in there in the first place.

11/02/2007 9:58 pm  
Blogger choppystride said...

I think it really depends on your usage for the Corsi #.

If it's to be used as a proxy for scoring chances, then I guess you can make a case of excluding BS. Although, personally, I've always regarded BS as a sub-category of MS. The only diff being that the NHL decides to give you the reason why the shot missed - it was blocked.

However, if it's to be used as a proxy for zone time, I agree with RQ that it should be included. In fact, I would even go further to include faceoffs as well. TK & GV might also be considered except that there might be a question of subjectivity in the way they are tallied.

11/03/2007 12:45 am  
Blogger The Forechecker said...

Great post and even greater discussion, guys. I saw we grab the pitchforks and torches, storm the NHL offices, and get stuff like this made publicly accessible. It's a pain in the a$$ having to construct workable data out of the play-and-play and other reports currently posted.

Yes, they make money charging for the real-time info that supplies news agencies, but setting up a data repository that's updated weekly or something could be a huge PR win. They could sponsor initiatives aimed at high school math education, for example, with sample exercises on how to wring meaning out of raw data, much like the TV show "Numbers" does.

11/03/2007 7:51 am  
Blogger rananda said...

i love the metric. one concern though is that certain players, those with high possession and who tend to create low events in terms of quantity but high events in terms of quality, may get shafted here. a god example is the jaromir jagr/michael nylander duo last year. this tandem drove the bus and was absolutely dominant for the rangers (especially in the playoffs). however, they tend to hold and hold onto the puck, hesitant to relinquish it even for a shot attempt unless such attempt has a reasonable chance of success. thus their shot attempts will undervalue their possession and probability of creating a goal (or drawing a penalty) especially compared to players like ovechkin or heatly who like to shoot from everywhere. furthermore, jags/nyls often face a checking line who is more than happy to fire the pill towards the goal from anywhere in the offensive zone, inflating the shots allowed attempt without really describing how close the opposition was to goal. this effect will not apply to most nhl players but to players like jagr and nylander who play so uniquely distinct (and awesome) when together, it could be a problem.

i would love to see the corsi numbers for last year's rangers (as well as every other team of course) and see what we have. im certain that jagr, nylander, roszival, and malik played great in terms of possession, creating chances, limiting chances, drawing penalties, or any other terms you would use in helping a team win hockey games (this was especially true in the playoffs). i would love to confirm this through the corsi numbers and more importantly, test the validity of the metric. conversely, i would expect numbers to reflect shanahan getting abused at even strength most of the year (esp in the playoffs) despite potting a goal here or there (and doing it against much weaker opposition than the big guys).

i agree with rivers q and dont see why we wouldnt include block shots in the calculus.

also, i think we should add in penalties drawn as well. forcing the opposition to take a penalty, even if a shot is not attempted, seems to be an effective way to not only increase the sample size a bit, but to factor in an event that is indicative of strong play and which leads to an increased chance of winning the game. conversely, we can subtract penalties taken.

11/03/2007 10:12 am  
Blogger Vic Ferrari said...

Riversq:

I'm not usually this wishywashy, but I think you're right. I've added blocked shots back on at timeonice.com.

I'm not sure how long I'll leave this shots stuff up there though, I'm running into account suspension problems because of CPU usage. At some point I may just have to put that stuff into a password protected directory and give you, MC, etc the code.

Either that or lean out the code ... which frankly is unlikely for a while at least.

11/03/2007 12:23 pm  
Blogger Vic Ferrari said...

rananda:

Thanks for commenting, and don't worry about the Rangers, they'll be fine.

And it's Shanahan, Drury and Avery who are playing the tough minutes for the Rangers, at least whenever Renney can get his way. I think that's probably why they don't look so good. Also probably why their Corsi numbers are no hell.

They'll come around, all of those guys have had success in the same role in the past.

I take your points, Jagr's line does wait for the better scoring chance, at least more than some, but he is at +25 right now, which is damn impressive. Prucha and some guy named Brandon Dubinsky are the only ones ahead. Probably because Renney is shit scared to play them against anybody good. Call it the "Wolski effect".

You can check out the H2H icetimes, and shiftcharts at timeonice, you'll find that's the case I'm sure.

Bear in mind that when you're playing against DET the tough minutes are against Zetterberg (who Jagr would see a lot of in Detroit) not against Draper (who Jagr would see a lot of in MSG). So when Crosby rolls into Colorado, Quenneville loads Smyth and Sakic onto a line and they play every shift against Sidney excepting a few after PPs and the last shift of the game (LaPerriere was rightly rewarded for playing the shift after COL PPs vs Crosby, and doing well. LaPerriere is a great guy, popular player, and did a good job with the "shift-afters" on the night. And Q is a smart coach.

And the heralded COL rookies are still looking forward to playing against Sid one day.

So any argument that uses "Staal's line had to play against the top checking unit of COL!" (LaPerriere's line) doesn't make sense to me.

And don't pine for Nylander, when Jagr was apart form Nylander (about a third of his icetime or thereabouts iirc) at evens in the past two seasons, all his numbers (already damn impressive) got a shade better. Counting numbers and underlying numbers alike. Conversely Nylander fell off the map in all regards.

Isbister put up stunning numbers, by all measures, when playing with Jagr last year, much better than Nylander. And I can tell you for a fact that he's no hell.

11/03/2007 12:39 pm  
Blogger rananda said...

hmm, im not really sure why you think it's shanny, drury, and avery getting the tough mins, at least at es, for the rangers. first off, that line has only played parts of 2 games together this year. (drury started the season with jagr and straka, was moved between shanny and avery before avery got hurt). more importantly, renney generally does not employ a straight up checking line but rather matches top line (jagr's line) v top line when he can get away with it (the thinking being that jagr and nylander/gomez [in theory for the latter at least] can control the puck in the offensive zone and neutralize the other teams' scorers). he certainly did that all of last year (to great success towards the end of the year and in the playoffs, jagr and nylander dominated the briere line for long stretches). this year, it hasnt really been any different. my memory of the games thus far this year is that renney is more than happy to put jagr and gomez against the other teams' top lines. looking at the shift charts for the last two rangers home games, this is basically clear. jagr got most of the work against lecavalier and ovechkin at es. (fairing quite well against vinny, getting the raw end of it against ovehkin, kozlov, and gordon). my memory and understanding of renney (the latter is a bit better than the former at this point) leads me to believe this has been the case since the beginning of the year, though i did not go back and check all the shift charts.

interestingly, i think you have it the other way around. jagr would see a ton of draper in det (tho renner will try some things to get jagr away from checkers on the road from time to time) and seeing tons of datsyuk and zetterberg at msg.

as for jagr fairing better at evens w/o nylander, i find this interesting, shocking at first glance, but i think there's an explanation. if you isolate just last year, it certainly cannot be the case. they were together at es all year and the few times they were not would be the rare occassions when renney double shifted jagr or did something strange to get away from pandolfo or get jagr against a 4th line. i think the majority of the jagr-nylander separation had to be in 05/06 when jagr played w straka and rucinsky for a long stretch, a line that was just as dominate if not more so that any jagr nylander combo, so that probably explains the result im thinking. jagr was also alot healthier that year (bum shoulder for much of last year, straka was also playing hurt for much of last year) so i think that skews the numbers against nylander. you cannot have had the privilege of having nylander on your team (the rangers are only my team in terms of geography and access to hd coverage, i go with the players) and not pine for him, he's that kind of player.

11/03/2007 1:51 pm  
Blogger rananda said...

also, as for the izzy quirk, it's pretty easy to figure out. izzy played on that line (nyls and jags) for about 10 or so games at the end of the year when the rangers were in do or die mode. the team hit their stride (started with the acquisitions of avery and mara and the injury to shanahan) and played great down the stretch. and the jagr line in particular was just dominating at es. they would go shifts at a time w/o seeing their end of the rink. this started with hossa as the third member, but didnt drop off too significantly (tho there was a noticeable drop, trust me) when izzy replaced him due to injury. make no mistake, this was all jagr and nylander (and roszy and malik). i think this illustrates nicely how even a stat like this has to be considered in a greater context for hockey. if you didnt have this info, you would be making the wrong conclusions, or unable to make any.

11/03/2007 2:00 pm  
Blogger Doogie said...

two G-D-men

Why not just call them the G-Men?

11/03/2007 4:32 pm  
Blogger Vic Ferrari said...

That's a lot of commentary, rananda. I don't know where to start. One day I'll write a post on Nylander, I'm still stunned by the fact that our GM, Kevin Lowe, was stupid enough to offer him that contract. Near miss there.

If you check the David Johnson EV stats on the sidebar for last season, Nylander played 237 minutes, near enough four hours, of EV icetime apart from Jagr ... and he bled goals against in that time, and his team failed to score to boot (puck has to be somewhere). Jagr got better.

If Johnson doesn't update these for previous seasons, sooner or later I will. It's the same for the season previous, almost exactly. And, with all due respect, there is a difference between providing context and rationalizing absent of proof.

Granted a lot of separation probably stems from the fact that if Jagr isn't tired after a shift, because the pucks been bouncing around his end the whole time, well he just stays out there for another shift, he is Jaromir F. Jagr after all. He is the Pavel Bure of this generation of players that way. What on earth was Slats thinking putting those two on the same roster at one point? Damn, if we assume that Sather isn't an idiot, then it was either a brilliant and diturbing social experiment, or ownership directed him. But I digress.

Drury's line was the one to play against vs BUF btw. I didn't see the BUF/NYR series, but that was the case last year. Especially if he had Hecht on his line, if Renney was running Jagr the shift after Drury then I would think that Ruff would drop Hecht back with Briere to help ease the pain, that was his usual pattern.

NYI (the eastern team I actually saw quite a bit of last year) loaded up Smyth/Sillinger/Hunter to go toe to toe with Drury's line. They did well, BUF just had too much scoring and D depth though.

If you look back through this blog, I have commentary on the first NYI/NYR game I saw after Smyth went to the island. It was at MSG, and it seemed to me that Nolan was running Kozlov and Satan at Jagr, which was mental, Jagr dominated. The next one at Nassau and Jagr didn't look to be a force, and he played damn near every shift against Trent Hunter's line. Turns out it wasn't Nolan being a fool in the game prior, just Renney getting his way.

Nolan moved Smyth with Sillinger and Hunter (who is all kinds of good at 5v5 hockey IMO) for the home stretch, and they played toe to toe against all the best players, I think there was a Rangers game in their too, I'll check when the mood strikes.

And the reason I assume that Shanahan, Avery and Drury have been playing the tough minutes (against the other teams best competitors, not necessarily the most famous. Defensive zone draws too) is because they are all good at it, all have Corsi numbers in the red, so chances are that spoke to that. And because it's so damn obvious that eventually all of these guys will end up with that gig from Renney. Still, I may very well be wrong, I certainly didn't check, it seemed to obvious to bother, but I'll happily be corrected with facts.

Islanders are set up in a similar way. Fedotenko, Comrie, and Guerin are out there trading chances with lesser players as much as possible. And while that only works if the rest of your team can carry the burden, and I think the Isles may be able to pull that off.

On WSH, what a train wreck. McPhee is the new Doug MacLean. They trap like an especially boring version of the Wild, except for Ovechkin, who is cherry picking like a young Rick Nash. (I'm talking pre-Hitchcock Rick Nash, only Hejda and Foote play tougher minutes for CBJ now).

I know it's off track, but I'll get back to this later.

11/03/2007 7:46 pm  
Blogger rananda said...

If you check the David Johnson EV stats on the sidebar for last season, Nylander played 237 minutes, near enough four hours, of EV icetime apart from Jagr

i cant find last season's stats on that site, nor can i see how i would determine how many min players played w or w/o other players, but i'll take your, or his, word for it. that seems insanely hight to me, though. that's about 3 min of es time per game. i watched probably 50 reg season ranger games and i cant recall any systematic occasions for nylander to be on the ic es w/o jagr. there could have been a stretch of games early on where they split the two, and that may account for the bulk of the time. actually, i do now recall a stretch where nyls played w shanny and dawes, so im assuming that represents a good chunk of the 237 min. gee, i wonder why his numbers were worse w that group. certainly, for the last 20 or so games when they made the push, and into the playoffs, there was very little es time for nylander away from jagr. any way to get stats post avery trade?

And, with all due respect, there is a difference between providing context and rationalizing absent of proof.

point taken. but what do your propose these stats are telling us? that nylander is a better player w jagr than w/o? thanks. that jagr is better w/o nyls than w him? no. that isbister is better than nylander? no. that isbister is better than hossa? no.

Granted a lot of separation probably stems from the fact that if Jagr isn't tired after a shift, because the pucks been bouncing around his end the whole time, well he just stays out there for another shift, he is Jaromir F. Jagr after all.

ahh, the continuation of the popular myth. i understand the attractiveness of the easy to digest narrative, that of the brooding but talented winger, joyless and selfish, who doesnt play defense, doesnt come off the ice, doesnt work in the corners, etc. it's fun and easy to perpetrate, but it's not true, at least it certainly has not been true in ny. (well, maybe the part about not wanting to come off the ice is a bit. but i'd rather have a jaromir jagr decide, if he's feeling like something good could happen, to stay on the ice rather than basically any of my other players hop over). also, jagr extending shifts wouldnt explain nyls' es min w/o him. nyls' extending shifts, which he certainly does as well, probably accounts for a small part of the 237.

Drury's line was the one to play against vs BUF btw.

not sure i understand or agree with this. my understanding was that drury's line was used to check other teams' top lines whereas briere's and roy's line did all the scoring es (roy's against much weaker ocmpetition). if you have a betts, boyd gordon, pahlsson-esque shutdown center, you use that against brier's line and you try to get your scorers away from drury. unless you have a jagr-nylander type combo where you can throw that out against briere and they can shut them down and score.

And the reason I assume that Shanahan, Avery and Drury have been playing the tough minutes (against the other teams best competitors, not necessarily the most famous. Defensive zone draws too) is because they are all good at it, all have Corsi numbers in the red, so chances are that spoke to that.

shanny, drury, and dawes played against richards and hlavac against tb and backstrom and nylander against wash. betts takes alot of the defensive zone draws and did a lot of the checking of other teams' top lines (to the extent renney does that) last season. that hasnt happened yet this season (renney has drury if he chooses to go that route this year, but shanny will negate any positive effect there). it's jagr's line that always matches up against the heatleys, lecavaliers, jokinens. this is what i consider the tough min. renney doesnt send out drury and shanny to face malkin and sid so that jagr can hide against staal and sykora. if paddock want to use fisher to check him in ott (i dont know if that's what he does but i imagine it is), renney is fine w that. i understand the instinct to want to believe that drury and shanny are getting the tough min and being used how pahlsson and neidermayer were last year, but i do not think 's the case. they are being used at es in the same way that the avery, straka, shanny line was used at the end of last season. do you have the same motivation to say that that line was getting the tough min? if i get motivated enough i'll go through the shift charts and tell you who shanny has been playing against this year, unless there's a better way to go about it.

11/04/2007 11:15 am  
Blogger Vic Ferrari said...

Jesus, that's a lot of words, I'll start taking a kick at it.

I cant find last season's stats on that site, nor can i see how i would determine how many min players played w or w/o other players, but i'll take your, or his, word for it. that seems insanely hight to me, though. that's about 3 min of es time per game. i watched probably 50 reg season ranger games and i cant recall any systematic occasions for nylander to be on the ic es w/o jagr.
there could have been a stretch of games early on where they split the two, and that may account for the bulk of the time. actually, i do now recall a stretch where nyls played w shanny and dawes, so im assuming that represents a good chunk of the 237 min. gee, i wonder why his numbers were worse w that group. certainly, for the last 20 or so games when they made the push, and into the playoffs, there was very little es time for nylander away from jagr. any way to get stats post avery trade?



You just click on the team name, then the player name, rananda. This is data assembled from the NHL.com TOI sheets.

You're eyes have been deceiving you. You'll see that Nylander played just 13:28 with Dawes, there may well have been a proper shift with Nylander in there somewhere, maybe even a couple, though mostly this just speaks of noise and a some of post-powerplay icetime before they got of the ice.

point taken. but what do your propose these stats are telling us? that nylander is a better player w jagr than w/o? thanks. that jagr is better w/o nyls than w him? no. that isbister is better than nylander? no. that isbister is better than hossa? no.

Sorry for not being more clear. My point: That Jagr is driving the results bus here, granted with more opportunity and less responsiblity than most non-rookies.


ahh, the continuation of the popular myth. i understand the attractiveness of the easy to digest narrative, that of the brooding but talented winger, joyless and selfish, who doesnt play defense, doesnt come off the ice, doesnt work in the corners, etc. it's fun and easy to perpetrate, but it's not true, at least it certainly has not been true in ny.

It's not a myth if it's true. Still, Jagr is better than he used to be, that year in Omsk did wonders for him I think. Rheostatic's guitarist Dave Bidini created a brilliant documentary on Russian hockey during the lockout year, must viewing for any Rangers fan I'd think. The scenes from Omsk reminded me why I was a hockey fan. I think Jagr still travels back to that small Siberian city in the summers, no? And Kasparaitus' interview with the Russian radio DJ is one for the ages. I wish that I spoke Russian and knew the local colloquialisms for "how much American ass are you tappin'? Seriously", because I think that the translator was toning it down.

Back to point, I watched some NYR games last year, and while he has improved in this regard, leopards never completely change their spots. I'd bet the moon that the first NYI/NYR game after the trade deadline saw Jagr take longer shifts, in that one game, than Shanahan has in any one game in his entire career. Prove me wrong, there are about 400 shift charts available with Shanahan's name on them.

(well, maybe the part about not wanting to come off the ice is a bit. but i'd rather have a jaromir jagr decide, if he's feeling like something good could happen, to stay on the ice rather than basically any of my other players hop over). also, jagr extending shifts wouldnt explain nyls' es min w/o him. nyls' extending shifts, which he certainly does as well, probably accounts for a small part of the 237.

Let me get this straight, Jagr doesn't hover out shifts, it's a myth. But if he did, well then it's justified?

I don't know what to say.

.

The rest of the stuff seems pretty wild to my eye, but I'll get back to it.

A question, Rananda:

By your eye, which forwards are on the ice for the largest share of the own-zone draws at evens for the Rangers?

By your eye, which forwards are on the ice for a disproportionately large share of the offensive-zone draws at evens for the Rangers?

11/04/2007 4:01 pm  
Blogger rananda said...

You just click on the team name, then the player name, rananda. This is data assembled from the NHL.com TOI sheets.

got it. im almost sad i've discovered this as this could lead to a great deal of wasted time.

nylander played w shanny and dawes for a game in toronto last season. dawes scored a goal (his first ever in the nhl) off a nylander pass, but was sent down shortly thereafter. i think nylander ended up playing with shanny and somone else for a few games as half of his ev time away from jagr was with shanny.

That Jagr is driving the results bus here, granted with more opportunity and less responsiblity than most non-rookies.

clearly, jagr is driving the bus here. i dont understand why you suggest it's w lessened responsibilities though. jagr was consistently matched against either teams' top scoring line or checking line, along with teams' shutdown d pair in most case. according to the behind the net numbers, jagr had a negative quality of competition number, but i am very skeptical of that. how is the behindthenet rating of opposing player determined? (how the hell could this be determined anyway, with any kind of authority or useful objectivity). i am assuming it's done via points or some other offensive statistic, which would then undervalue the rating of players like madden, gordon, talbot, players that jagr saw a fair amount of last year, and who are difficult to play against. (which would also explain why the leaders in this category are always the checking guys, theyre going up against the offensive players, not the other checkers. shaone morrison has a high rating because he lines up against jagr, not because he lines up against madden). this has to be the reason for the negative quality of comp figure. how often do you see jagr out there against players like vanek or preissing? that to me would be lessened responsibilities, he would have feasted on those guys.

Rheostatic's guitarist Dave Bidini created a brilliant documentary on Russian hockey during the lockout year, must viewing for any Rangers fan I'd think.

sounds amazing. any idea how i could score a copy?

not sure if jagr goes back to omsk in the summers but he has repeatedly stated his desire to return there to play once his nhl career is over. if he doesnt pick it up he may not get the automatic extension for next year (based on attainable point totals and the rangers winning a playoff round - with the caps again picking up half the tab) and we could be seeing him in the red and black next year. otherwise, i think we'll see him in omsk wearing the cereal box helmet a la larionov in 2009 for his last year of competitive hockey.

I'd bet the moon that the first NYI/NYR game after the trade deadline saw Jagr take longer shifts, in that one game, than Shanahan has in any one game in his entire career.

this is interesting. shanny does take short shifts, in fact, he was taking 20-30 second shifts in the playoffs. it was insane. the reason? he absolutely could not keep up with the play. it was pretty sad. a puck around the boards on his side had maybe a 30% chance of getting out. he knew he couldnt hack it and he was getting off as quickly as he could. it didnt matter. he was directly responsible for the 3 goals in the final buff game (he got caught flatfooted after an o-zone faceoff in the 1st period (same thing happened in the atlanta series) something in the 2nd period i cant remember, and he lost hect on a harmless 3 on 3 in the 3rd. i would argue he was the most determining reason the rangers lost that series. i have no basis for analyzing his off the ice contribution to the team (by all accounts he is tremendous), but he was brutal on the ice in the playoffs short shifts or not. 0 second shifts would have been much better.

the long shift part is true about jagr. i guess the best counter is, who cares? wayne gretzky probably scored more points after the 60 sec mark of shifts than a lot of nhl players did their whole career. good players too. if jagr was getting burned, the coach would yell at him, bench him, and the team would lose. jagr doesnt like being benched or losing any more than anyone else.

By your eye, which forwards are on the ice for the largest share of the own-zone draws at evens for the Rangers?

let me preface any response to this by noting that renney, more than any other coach i have watched (other than ivan hlinka who was in a league all his own on this), tends to roll his lines absent matching or getting certain people on for certain zone draws. he is not at the level of the russian coaches who literally role 4 lines with out any thought as to situation or matching. i contend that it is that tactic, absence of tactic rather, that has resulted in poor showings (at least relative to ability imo) at major international tournaments recently. dave king is his recent book about coaching in magnitogorsk talks about how the russian coaches do this and, implicitly at least, how dumb it is. king talks about how he would be able to get malkin out against 4th lines and malkin would just eat them alive. btw, king led that team (not even close to being the highest salaried team in the league) to the best record of their modern era, since the old cska teams used to go a whole season and lose maybe once or twice. i dont know if it's laziness or stupidity, but that russian style is beyond baffling to me.

but back to renney. he's not as extreme, and i think it stems from a respect for and confidence in all of his players (the ultimate player's coach, which partly explain he gives shanny way too much on ice responsibility). renney for the most, does not play the situation. i have seen him use dubinsky on dzone draws, and ive seen him keep betts, orr, and hollweg out there for ozone draws. to wit, in the crucial buff game 5 with the rangers up by one with under a min to go, he had nylander, jagr, and isbister out there for the dzone draw to protect the lead. of course that didnt work out that well (tho girardi, who is becoming a fantastic player and combining w tyutin to form a bona fide shut down pair) would love to have that play back).

but i guess, and this is really not something i would guess w too much confidence, that it was primarily nylander and immonen/straka/avery in the ozone, and cullen and betts in the dzone. this year, it would be gomez and dubinsky in the dzone. i dont think this pattern is going to be nearly as clear for a coach like hanlon, where youre going to see gordon and steckel out there for the dzone draws, and kozlov and nylander in the ozone.

11/05/2007 10:10 am  
Blogger rananda said...

that last paragraph should read gomez and dubinsky in the ozone and drury and betts in the dzone.

11/05/2007 10:12 am  

Post a Comment

<< Home