The Noon Number
by J.P. on Jan 4, 2012 12:00 PM EST
in The Noon Number
Email
Print
89 - Number of blocked shots with which the Caps have been credited during their current four-game win streak (blocking at least 20 in each game, something they did once under Bruce Boudreau this season), a pace which would easily lead the League over the course of the season. And while that might sound like a good thing... it isn't, necessarily. As Colby Cosh put it:
The counting of blocked shots has a problem similar to the counting of double plays turned by a team in baseball; they correlate negatively, if at all, with the winning of games. An individual blocked shot might have a positive value—though even that’s certainly not true in every case—and you want players who are willing to block them. But it’s better not to give up lots of opportunities for blocked shots.
(For more on the correlation between blocking a lot of shots and success, check out this ESPN Insider post by Neil Greenberg ($), and the discussion in the comments on this post, among other sources.)
139 comments
|
0 recs |
Do you like this story?
Comments
Why are people now saying we are getting TOO many blocked shots? I’ve heard for years we don’t block enough shots, that our players aren’t willing to do what they need to do to win (like blocking shots). now its too many?
It’s not that the team is trying too hard to block shots. It’s that opponents are getting too many shot attempts that can turn into blocked shots (or saves, or goals, or misses).
Being on the receiving end of that many shots for too long is playing with fire.
"You do that, you go to the box, you know. Two minutes, by yourself, you know and you feel shame" -Denis Lemieux
by leacha on Jan 4, 2012 12:23 PM EST up reply actions 1 recs
This is it.
When you block a shot, it’s like tossing a many sided die into the air and leaving your fate in the hands of a cruel and fickle God. Does it carom out of the zone and force the attacking team to regroup? Does it pinball into the net? Does it deflect to an unmarked player for an easy tap in? All these outcomes, and many others of course, are possibilities.
But spending too much time in your own zone is not a recipe for winning in the long haul. Although if you are going to spend a lot of time in your own zone anyway, better to block shots than not, I suppose.
But best of all would be to spend more time in the other guy’s zone, where his poor bastards can prove their manhood by diving in front of pucks and utter fervent prayers to the merciless God of random chance.
Patron saint of quality footwear.
by fat_daddyo on Jan 4, 2012 12:59 PM EST up reply actions 1 recs
When you block a shot, it’s like tossing a many sided die into the air and leaving your fate in the hands of a cruel and fickle God. … Does it pinball into the net?
Japers' Rink: Hockey blogging from the most powerful city in the world
That whole series is a great example of why it’s bad, bad, bad to spend too much time in your own zone.
Patron saint of quality footwear.
And, on the flip side, the MTL series was a pretty good example of a team getting dominated territorially and getting really pretty lucky.
Japers' Rink: Hockey blogging from the most powerful city in the world
Right – Montreal was lucky, just as the Caps have been lucky lately. When you’re out-Fenwicked, out-possessed, out-chanced, you’re usually out-winned.
Japers' Rink: Hockey blogging from the most powerful city in the world
I’d encourage you to read the stuff I linked to in the post. Good discussions on the topic.
Japers' Rink: Hockey blogging from the most powerful city in the world
NHL on the Fly—the fount of great analysis and a bastion of objectivity—said that the Caps are finally doing all the little things right. Like blocking shots.
Knuble: "I am what I am. I play well with good players."
On his milestone: "It's going to be like unwrapping a birthday present. Then the day after you're kind of like [sighs]....Now you just have to keep going."
by capsyoungguns on Jan 4, 2012 12:33 PM EST up reply actions
I was also referring to last nights recap:
“Now that Vokoun is stopping pucks like fans had anticipated, maybe the Caps can leave some of those shots to the ’keeper and save their ankles for another day.”
that would say let them go on net. nothing to do with amount of shot attempts
Thanks for the link. I just submitted a post on this topic to WaPo, so if you don’t have ESPN Insider more info will be up in a day or so.
"Shots aren't the important thing. Scoring chances are way more important than shots." - Bruce Boudreau
See my work in the Washington Post and on ESPN Insider.
Follow me on Twitter @ngreenberg
Looking forward to it—speaking as someone who bangs her head against that pay wall. Eventually I will cave in.
Knuble: "I am what I am. I play well with good players."
On his milestone: "It's going to be like unwrapping a birthday present. Then the day after you're kind of like [sighs]....Now you just have to keep going."
by capsyoungguns on Jan 4, 2012 12:34 PM EST up reply actions
I thought Neil’s legal name was “Greenberg ($)”…
J.P.: You might be the king of all geeks here…
by Alz Well That Ends Well on Jan 4, 2012 1:06 PM EST up reply actions
No, it’s Billy Beanberg.
Japers' Rink: Hockey blogging from the most powerful city in the world
by J.P. on Jan 4, 2012 1:07 PM EST up reply actions 1 recs
After the Bourque incident last night and with the number, I was expecting this to be about someone currently with a buyout cap hit.
On the topic, agree about too many blocked shots not being a good thing; prefer to see the guys clearing the puck out of the zone.
My therapy must be going well – I see 89 and think Mogilny now.
Japers' Rink: Hockey blogging from the most powerful city in the world
But it’s better not to give up lots of opportunities for blocked shots.
Don’t like the way that reads. It is either a “no shit” kind of empty statement, or it makes it sound like the caps are trying to give up shots so they can in turn pad their blocked shots. I get that teams don’t want to give up a lot of shot opportunities – but if they do, I want my team blocking them. Period.
Keep blockin’ those shots, boys!
___
But that’s the point. If you’re blocking a lot of shots it’s because you don’t have a lot of possession and if you don’t have a lot of possession, you’re not in a great position. In other words, blocking a ton of shots correlates negatively with something that is strongly associated with winning. Per an article that seems to have died on the Internet somehow:
Having a lot of shot blocks has zero relationship with goals against and penalty killing percentage, and has a very, very slightly positive relationship with save percentage.
There’s a chart that went along with that post:
So what you see is shot blocking negatively correlated to winning hockey games. Pretty straight forward.
Japers' Rink: Hockey blogging from the most powerful city in the world
by J.P. on Jan 4, 2012 12:47 PM EST up reply actions 7 recs
This is my favorite hockey chart.
"You do that, you go to the box, you know. Two minutes, by yourself, you know and you feel shame" -Denis Lemieux
That makes sense and that is a very informative chart.
I think my reaction to this is based more on the fact that it seems for all the whining that has gone on for this team not doing the little things, and failing to form good habits… consistent shot-blocking is a good thing to be doing – and the focus should be turned to reducing chances, not to try and make shot-blocking look like a negative practice.
consistent shot-blocking is a good thing to be doing
No, not it’s not. See above.
The little things that are discussed (I wouldn’t say “whining,” since desiring improvements in the fundamentals of the game doesn’t seem bitchy to me) are things like outlet passes, gap control, etc. – things that actually help you win.
I applaud the effort, the guts, etc., but I’d rather the Caps be reducing chances, as you say, via possession than doing a collective Dikembe Mutombo impression.
Japers' Rink: Hockey blogging from the most powerful city in the world
(“No, no it’s not,” that is.)
Japers' Rink: Hockey blogging from the most powerful city in the world
Your chart shows correlation because with the increase in shot-blocks, there is an increase in shots. Let’s decrease the shots, but keep the shot-blocking train going. That’s what I’m saying. Shot-blocking, on it’s own, ISN’T bad – in fact it’s good. It’s just the increasing number of shot-blocks means there have been more shots attempted – which is the bad.
I meant whining more as a constant, well-intentioned drone of the masses that wanted this team to finally commit to the little things. It was meant in an endearing way – not as an attack.
I get what you’re saying. To paraphrase Gretz, 100% of the shots that never reach the net don’t go in. But here’s the thing – those shots aren’t going in anyway. Per Neil’s link:
The reality is the majority of shots being blocked are of low quality, and teams that block a lot of shots do so by necessity because they are spending a majority of their ice time in their own end without possession of the puck.
It’s quite possible – probable, perhaps – that the benefits of attempting to block a given shot (favorable deflection) are outweighed by the negatives associated with the attempt (unfavorable deflection, injury, screening goalie on a shot that makes it through, etc.).
It’s similar to the proverbial “good penalty” in some ways. To be a “good penalty” for a team that kills 80% of its opponents’ PPs, that penalty would have to negate a scoring chance that is 21% likely to succeed (i.e .the chance is at 21% or better, while scoring on the PP is 20% likely). Those 21% chances are exceptionally rare. In many cases – with regards to taking penalties and blocking shots – the defending team, statistically, would be better off letting the play happen without the action.
(Granted, on the “good penalty” point, that action would have to be a 100% guaranteed penalty, which doesn’t really exist in practice, or the odds shift around.)
Japers' Rink: Hockey blogging from the most powerful city in the world
That makes sense – even from the past few games, 20 blocked shots seem high I would guess because only a handful of them were of the “good quality” that prevented a decent shot and stood out during the course of play.
I guess I just keep reverting back to the playoff series with the Rangers were all the announcers were talking about was how Torts forces his guys to block shots and that’s why they’re so hard working and hanging around in all the games…
And how many playoff series have they won?
If you think about it, what kinds of shots get blocked? Almost exclusively those from outside the “home plate” area, no? By definition, those are non-scoring chances. So lots of blocked shots means a) you’re likely getting crushed in possession and b) you’re risking a decent amount to stop non-quality chances.
I like the dedication it shows, and I do think that matters some. But ultimately, it’s probably not a great stat for a team to be piling up.
Japers' Rink: Hockey blogging from the most powerful city in the world
No way, Orlov blocked a shot from inside the crease last game. Your argument is flawed.
Please, call me F&B.
The reality is the majority of shots being blocked are of low quality
Is that empirical? Based in logic? Does it consider the fact that most of the shots that lead to rebound goals are also of “low quality”?
Please, call me F&B.
Is that empirical? Based in logic?
It has to be, I don’t watch the games.
In all seriousness, it is based on Sh% of the shot in question.
Does it consider the fact that most of the shots that lead to rebound goals are also of "low quality"?
Yes. Curious, how many of those shots do you think actually exist? Per game?
As for rebounds:
"Shots aren't the important thing. Scoring chances are way more important than shots." - Bruce Boudreau
See my work in the Washington Post and on ESPN Insider.
Follow me on Twitter @ngreenberg
Yes. How many goals per game do you think are scored off rebounds?
Also, how many shots do you think are rebounds on the whole?
Let’s use the widely accepted definition of a rebound being within two seconds of the shot in question, without any interfering events.
"Shots aren't the important thing. Scoring chances are way more important than shots." - Bruce Boudreau
See my work in the Washington Post and on ESPN Insider.
Follow me on Twitter @ngreenberg
I’d say that teams play for rebounds, and rebound chances are generally dangerous, so blocking a potential rebound-creating shot has some value in it.
Please, call me F&B.
Answer the questions, counselor!
Japers' Rink: Hockey blogging from the most powerful city in the world
OK, great, but I asked how many goals per game do you think are scored off rebounds?
"Shots aren't the important thing. Scoring chances are way more important than shots." - Bruce Boudreau
See my work in the Washington Post and on ESPN Insider.
Follow me on Twitter @ngreenberg
Man, it really sucks when someone dodges a question, doesn’t it?
Please, call me F&B.
by Rob Parker on Jan 4, 2012 1:58 PM EST up reply actions 7 recs
I’ll bite. I’d say, finger to wind, one every two or three games. So .3 to .5?
Occasionally wrong, never uncertain.
This is good conversation all over and I am not sure where to jump in, so I’ll do it here. This sounds like a group kinda talking past each other. I posit that it’s not that “Shots Blocked” is a “Good” or “Bad” thing. Clearly you want to have them, but a lot are a sign of something wrong. I think it is because it’s not a singluar event in the data. That is, for there to be +1 Shot Blocked (Good), there must first be a +1 Shot Against Attempt (Bad). It’s not like a goal (good) which has as a prerequiste Shot and Shot Attempt (Good and good).
I suggest it’s the ‘wrong’ stat. As has been suggested before perhaps a percentage of shot attempts against blocked is better, or to address quality percentage of scoring chances against blocked would be the better metric to address here. In the limits, if my team blocked 100% of the scoring chances generated by the other team, that is oozing in awesome. If they block 0% of them then proper cap management dictates replacement with cones. By making it a percentage it would seem to help factor out the “Yeah, because you are getting shelled” aspect, and could normalize to measure well for a team that does dominate the puck too.
Second thought: Blocked shots never go in the net directly or ‘go through’, those are accounted for in the stat boxed labeled “Goal against”. right?
Occasionally wrong, never uncertain.
by VaMedic on Jan 4, 2012 1:45 PM EST up reply actions 1 recs
I would think that seeing the correlation between % of total shot attempts blocked and winning would be interesting to know, yes.
Japers' Rink: Hockey blogging from the most powerful city in the world
I also think this is pretty fertile ground for the distinction between Fenwick and Corsi and the qualitative differences inherent in the use of the respective stats in any analysis. Isn’t Fenwick more predictive? Wouldn’t that imply that blocked shots have some impact?
Please, call me F&B.
It’s my understanding that Fenwick is more predictive, yeah. And that certainly would imply that blocked shots have an impact, if adding blocked shots in makes the metric less predictive. But isn’t that, at the very least, saying that blocked shots aren’t strongly correlated to winning (and, at the most, saying there’s a negative correlation)?
Japers' Rink: Hockey blogging from the most powerful city in the world
I don’t know how there’d be a negative correlation. Isn’t it basically saying that if a shot is blocked you don’t even count it as a possession metric? So the only shots that a team is “punished” for in the possession metric are the shots that get through cleanly.
But yes, the Corsi and Fenwick predictive values are close enough that you still are left with the conclusion that it’s not a huge factor. I’m more curious about where that divergence is coming from. Obviously you are giving up possession if you are getting blocked shots, but why is that less bad than a shot that goes wide?
Please, call me F&B.
I’d think it’s because shot-blocking is more about the defending team (despite being not-entirely-skill-based at the team level) and therefore separating it out results in a truer picture of what the team in question is actually doing.
So if the Caps attempt the same shots (say, 50) against Team X and Team Y, but Team X is a better shot blocking team and blocks more shots (say, 10 to 0), the Caps’ will have identical Corsis (50) in the two games (holding what happens at the other end constant at 0), but a better Fenwick in the game against Team Y (50 to 40). The Fenwick is likely to be more predictive because more shots are getting through against Team Y, so likely there are more scoring opportunities and ultimately more goals against Team Y, against whom the Caps have the higher Fenwick.
No?
But they’re really close. And IIRC score effect is may explain the difference (trailing teams do better in Corsi relative to Fenwick for some reason).
Japers' Rink: Hockey blogging from the most powerful city in the world
I follow your logic, but I was under the impression that Fenwick and Corsi were proxies for puck possession. The results of the shot are irrelevant compared to the fact that your team is in the offensive zone with the opportunity to take the shot. So as you lay out the argument it makes sense that Fenwick would be more meaningful, but wouldn’t hitting the net also be more meaningful? If we talk about which shot attempts have a higher likelihood of going in then hitting the net is obviously a part of that. I guess a response would be that there is luck in whether a shot hits the net, and I don’t know that I’d agree entirely with that claim.
My concern is that there seems to be a breakdown from “these are proxies for puck possession, which is what really matters” to something more qualitative going on in the offensive zone. It just seems as though when you start taking out blocked shots for reasons that aren’t related to the stats’ ability to serve as a puck possession proxy you are moving the goalposts somewhat, and inching ever closer to that dangerous discussion of shot quality.
I would guess that trailing teams do better at Corsi relative to Fenwick because they are throwing everything at the net and the team with the lead is content to collapse and keep everything to the outside and block what they can. (Related, I think, this Noon Number discusses games the Caps won, so maybe some of that was going on with CGY. Despite the raw shot attempts, I didn’t think CGY had many great scoring chances… until the third period… when the Caps should have been locking it down and blocking perimeter shots. So I don’t know.)
Please, call me F&B.
Yeah, I don’t disagree with any of that.
I think it comes down to doing the best we can with the information we have.
Goals-for and -against correlate perfectly to winning – you will win 100% of the games in which you outscore your opponent.
But what underlies goal-scoring? Scoring chances and luck (and “luck” here includes fate/chance/whatever as well as shit that’s not fate/chance/whatever, but we can’t explain, so we call it luck). And what underlies scoring chances? Shots attempted… and luck.
But there is a bit of a leap from shot attempts to puck possession – we refer to Corsi and Fenwick as “possession,” but that’s based on the assumption, like you say, that “your team is in the offensive zone with the opportunity to take the shot.” If it was true “possession,” though, it’d be somehow quantifying cycling the puck, time with the puck in the offensive zone, etc. – possession in the definitional sense of the term.
So it’s not really “possession” that matters as much as shot generation, is it? Logically, one could imagine three shots on goal in a 10-second span easily producing more offensive opportunity than a 20-second game of catch at the blueline that was true “possession.”
As to hitting the net being more meaningful than missing it (or “why include missed shots if you’re going to exclude blocked shots?”), my guess is that there’s a good bit of luck (or “luck”) there and that plenty of missed shots are as close to being goals as actual shots on goal. For example, which is closer to a goal – a Dmitry Orlov bomb from the top of the circle that whistles just high and wide or an Eric Fehr floating wrister to the bread basket from 40 feet? Obviously one goes in and one doesn’t if there’s no goalie, but the percentages probably say the Orlov shot (based on where it’s taken, not where it ends up) is more likely to score, and where it ends up is a matter of chance.
Japers' Rink: Hockey blogging from the most powerful city in the world
I agree, the Orlov shot definitely has the better chance of going in, but at the same time, the ability to hit the net is more skill than luck, I’d wager.
Please, call me F&B.
I think that gets into a bit of a selection bias issue – most of these guys could put one into the middle of the net with every shot if they wanted to. Some do. But goal scorers try to pick corners, and tend to cut it too fine (see Semin, Alexander) and the miss high and wide more often. Putting a shot exactly where you want is a skill (they even have an event for it at the All-Star Game!), but it doesn’t take much chance to throw that off by a few inches, and it doesn’t take more than a few inches to send it high and wide.
Japers' Rink: Hockey blogging from the most powerful city in the world
Fair enough, I wonder how many shots are really the “pick the corner” variety. I think a lot of the guys are just trying to get it off and hit the net, but your point is well taken.
Please, call me F&B.
Yeah, no clue. My guess is it varies greatly from player to player. Jason Chimera probably just wants to make sure he hits the net. But I doubt Alex Semin has ever taken a slap- or wrist-shot that he wasn’t trying to put in a spot the size of a “win a car at intermission” hole in the upper corner.
Japers' Rink: Hockey blogging from the most powerful city in the world
Oh no doubt with Semin. I wonder what AO tries to do. It looks like he’s just trying to shoot a hole through the goalie, but I think he probably does try to pick corners pretty routinely.
Please, call me F&B.
Not that it is definitive on this point, but…
Ovechkin: 146 SOG, 70 misses
Semin: 73, 49
Chimera: 93, 22
So of their shots that aren’t blocked, Ovechkin is getting 68% on net, Semin 60%, Chimera 81%.
I’d buy that the three try to pick corners with the frequency that those numbers might imply.
Japers' Rink: Hockey blogging from the most powerful city in the world
Regarding the predictive ability of Fenwick versus Corsi for a given timeline:
Money quotation:
Therefore, while corsi has more predictive validity with respect to goal ratio at the within-season level, fenwick and shot percentage appear to correlate more strongly with goal ratio over a sufficiently large sample of games. In other words, in theory, both fenwick and shot percentage seem to serve as better measures of team quality than corsi does.
One caveat: the differences between the values here are small, and we only have three seasons of data. It may very well be that all three variables correlate equally well with goal ratio over the long run. This subject may require further study in the future when more data is available.
"Essentially, all models are wrong, but some are useful" George E.P. Box
I have $50 to Olie's autism charity or So Kids Can and a beer on the under for 50/57/107 as AO's final stat line for 2011-'12
And on Pekka Rinne's PK SV% dropping under .920 by the end of the '12 season
by Knee high to a duck on Jan 4, 2012 3:52 PM EST up reply actions
Yes! Knee High is here! Been hoping he’d join in…
Japers' Rink: Hockey blogging from the most powerful city in the world
I think you’re gettting at something I’ve wondered about but had a hard time putting into words. I’ll try one more time.
Take last night as an example. The Flames had:
19 shots on goal
24 blocks against
9 shots that missed*
That’s 52 attempts at goal, with roughly 46% of them getting blocked.
It feels intuitively like a stifling defensive performance might correlate positively with winning, and that a stifling defensive performance might include blocking a large percentage of shots, or blocking a large percentage of shots in conjunction with a low raw number of attempts.
In other words, giving up, say, 75 attempts while blocking 25 might be a much bigger correlation that drags down the expected points %. Or something. I’m not clever enough to sort out all these issues coherently. But maybe this can be a starting point for one of the smarter guys to start looking at the issue.
*at ES, according to the Corsi/Fenwick report, so might be undercounting missed shots on the PP here
Patron saint of quality footwear.
By way of contrast, last night Washington had:
21 SOG
7 blocks against
3 missed shots (again, at ES only)
That’s 31 attempts at net with 22.5% of them blocked.
Not sure what you might make of that, to tell you the truth.
Patron saint of quality footwear.
And going with Neil’s point, if the Caps did block a ton, and shots that are blocked are generally low quality, then it also means that CGY was sacrificing quality for volume (which is anathema to the stats crowd, I’ll note). Does that make the raw numbers look better for the Caps?
Please, call me F&B.
Right. In the extreme, if you block 100% of your opponents’ shot attempts, you’re going to win. If you block 25 of 50, the possession battle is likely different looking than if you block 25 of 75.
But I might be wrong, but I seem to recall someone (Mehta?) noting that shot-blocking on an individual level is a skill (Volchenkov, etc.), but on a team level is luck-driven. Essentially, the Caps can’t count on blocking 46% every night – they’re obviously much better off fixing what’s leading to the possession discrepency.
Japers' Rink: Hockey blogging from the most powerful city in the world
It was Mehta and if you have a guy like Volchenkov, you’re going to see an increase in your team block rate. What’s almost totally random is your block-rate against.
"Essentially, all models are wrong, but some are useful" George E.P. Box
I have $50 to Olie's autism charity or So Kids Can and a beer on the under for 50/57/107 as AO's final stat line for 2011-'12
And on Pekka Rinne's PK SV% dropping under .920 by the end of the '12 season
by Knee high to a duck on Jan 4, 2012 3:55 PM EST up reply actions
What about in the playoffs? Does anything change when you weed out the really weak teams that rack up blocked shots because they are pinned for 82 games? I know TBL blocked a ton of shots when they won the Cup and I think CAR did as well. Maybe it was just random, I don’t know.
Please, call me F&B.
Here’s what Neil had to say about the playoffs:
During the postseason, over the same time frame [post-lockout], we see almost no correlation between blocking shots and winning games. The negative effects of having to block too many postseason shots are evident at the upper end of the spectrum, most likely because these teams are continually on the losing end of puck possession.
Admittedly it is a small sample size, but having to block more than 20 shots per game in the playoffs could limit your chances of winning the Stanley Cup. Last year the Montreal Canadiens were able to ride that strategy to the Eastern Conference finals but ended up losing to the Philadelphia Flyers in five games, while both the Colorado Avalanche and Ottawa Senators couldn’t make it out of the first round. This year the New York Rangers and Canadiens were two of the top three teams in blocked shots per game, both losing in Round 1.
Tampa Bay, the playoffs’ top shot-blocking team, made the most of the chances it got in dispatching Washington in the second round. However, if the Lightning continue to lose the battle of puck possession and rely on blocking shots as an equalizer, it is doubtful they make it much further after that.
Japers' Rink: Hockey blogging from the most powerful city in the world
Heh, doubtful they’ll make it much farther than the penultimate round. They suck!
Please, call me F&B.
They only made it past the anti-penultimate round.
Japers' Rink: Hockey blogging from the most powerful city in the world
My favorite part is the little triangle for P% all the way up in the upper left. That there is what you would call a statistically significant correlation.
Patron saint of quality footwear.
by fat_daddyo on Jan 4, 2012 1:03 PM EST up reply actions 2 recs
Could you provide a source please?
I’m really curious about where these data points come from. Would you please be able to provide a source/link?
I say that because while I think it’s easy to correlate that blocking more shots means you’re in fact giving more shots up, what I think isn’t entirely represented in these data points are the quality of shots and scoring chances to blocked shots. Much like all shots aren’t valued the same, the value of a blocked shot can’t be the same in each instance.
My point being, I think as fans we’d be better served if we were to be able to correlate the following data points to accurately assess whether or not that data is conclusive. Those points being:
- average total number of scoring chances conceded
- average total shot attempts yielded
- average blocked shots
As it pertains to the Capitals, I think what would be more valuable is to see a breakdown of those data points above from BB’s tenure to DH32’s. Of course the sample sizes would be vastly different, but it’d give us more insight as to the impact of DH32’s system. Even if the data isn’t entirely accurate due to the small sample size. Within the context of last night WSH gave up 44 total shot attempts. 24blocked, 19 on net, and 11 missed. Then 6 of the 19 shots on net were from low scoring areas near the blue line or beyond. It starts to look a little more favorable. Of course the Caps were able to put 32 shot attempts on net. 7 blocked, 21 on net and 4 missed with 6 being from low scoring areas near the blue line or beyond. Now 44 > 32, but 15 > 13 and unfortunately the ICETracker doesn’t track blocked shots or misses so we’re not able to accurately discern who had the higher quality scoring chances as that stat isn’t readily tracked. I feel that by considering these additional data points it really adds more context to the impact of blocked shots.
http://www.nhl.com/ice/icetracker.htm?id=2011020576
As we begin to use stats more and more, which I think is fantastic, I still feel we’re only topically hitting their values and what meaning/role they play in hockey. It’d be nice if we could provide even more context to these data sets.
Also please correct me if I’m wrong but if a shot block is attempted and the puck strikes the defender but still continues in, isn’t it recorded as a shot and not a blocked shot?
It’s from here (sorry I didn’t link to it above – I thought the link was dead).
Japers' Rink: Hockey blogging from the most powerful city in the world
Interesting!
Thank you very much for the link!
It actually confirms my thoughts. The author’s method uses the teams relative point percentage within the standings to calculate the overall value of shot blocking to wins. The one thing that I will say is flawed with this data is that it’s from 2008. Because of that fact, it’s actually not applicable for today’s NHL. A quick glance at the Real Time stats page on NHL.com for team statistics show that there are a few good teams close or in the top 3rd of the league in Total shot blocks. (Unfortunately I don’t have time to calculate avg blks and I imagine the rankings would shift slightly because of the difference of GP). Minnesota leads all teams and while they’ve struggled of late, they were at the top of the heap for some time. They are also averaging 5 more shots against than for. After that you do have some struggling teams MTL, CAR, NYI but then have TOR, DAL, NYR, SJS, TBL, EDM, PHI, NSH. You could throw TBL, EDM out of the list as they’re not really in “playoff contention”, but NYR, PHI, & NSH? Those are solid clubs. TOR, DAL, SJS are sitting on the bubble. What’s interesting is that only PHI has a positive S/G v SA/G.
My point being, Mirtle’s myth bust in the context of today’s NHL is actually wrong. Now, I believe that due to the lack of context from quality scoring and recent data points more so than anything else. Looking at current data points you’ve got the NYR as the best team in the NHL averaging in the top 3 in blocked shots, then you have 3 playoff teams, and 3 teams on the bubble among the leaders. Interestingly enough, on the opposite end of the spectrum is CBJ who are the worst team in the league and block the lowest shots.
To me, what this says, is that a quality of shot statistic is vital to this data set. This is also confirmed by the Out-Shoot/Out-Shot by statistics where you have BOS, MIN, NSH having high W% while giving up more shots than taking.
While I agree that there is definitely a correlation of shot blocks to shots against, we’re still not tapping into the quality of those shots being given up. As we don’t have that data I feel it’s a bit presumptuous to assert that in most cases having more blocks has a negative impact. Man, I really hope that scoring chances/opportunity are actually tracked and accessible. It’d really shed some light on which teams are actually stout defensively and which teams are just perceived as such.
by paid in full on Jan 4, 2012 2:32 PM EST up reply actions 3 recs
Let’s look at shots blocked on the road from last year (it gives us a full year of data and eliminates home scoring bias).
Top 10:
ANAHEIM
NY ISLANDERS
ATLANTA
NY RANGERS
PHILADELPHIA
BOSTON
COLORADO
TAMPA BAY
FLORIDA
BUFFALO
Bottom 10:
EDMONTON
PHOENIX
CAROLINA
SAN JOSE
OTTAWA
NEW JERSEY
VANCOUVER
NASHVILLE
PITTSBURGH
DETROIT
Of those top-10 teams, six playoff teams and some shitty teams. Of the bottom-10 teams, six playoff teams and some shitty teams. At the very least (from that small sample), the correlation between blocking shots and winning doesn’t seem to exist, does it?
Japers' Rink: Hockey blogging from the most powerful city in the world
I think I failed to communicate my point effectively.
What I’m saying is, looking at only blocked shots to wins doesn’t convey a complete picture. It leaves out, what I perceive to be a vital data point, which is the number of quality scoring chances given up. Using the total average blocked shots as a metric is similar to using total number of shots for as a metric. As evidenced in the WSH-MTL series, we know that not all shots are of the same value. Putting up 50 shots on goal while only generating 6 quality scoring chances is misleading. Wouldn’t you agree? It’s the same argument that both BB and GMGM presented after that loss, and I believe you had an article that opined the contrary.
I think it stands to reason that if you are blocking more shots, giving up fewer quality scoring chances, but giving up slightly more shots against, that there is a correlation there that defines the value of shot blocks. The implication being that fewer quality scoring chances have been prevented or blocked. The result being the team is stout defensively and the blocked shots are representative of that as you’d be reducing scoring chances.
To answer your question, what your data above conveys to me, is that shot blocking doesn’t have a negative impact as you presented in your article.
“And while that might sound like a good thing… it isn’t, necessarily”
Perhaps I misunderstood, but seeing that the eventual Stanley Cup winner blocked a lot of shots and their opponent VAN didn’t, it stands to reason that blocking shots from the data presented is neither positive nor negative.
I think the real value will become evident if we had access to the scoring chance data. It’s an assumption on my part at this point because we don’t have that data, but I believe that the logical conclusion were that data available is we would see that good defensive clubs will still block a lot of shots while giving up fewer scoring chances. Conversely, we would be able to assert that bad defensive teams can still block a lot of shots, but by giving up more shots and higher scoring chances the value of those blocked shots are diminished.
Again, I’m not trying to poo-poo on Mirtle’s work, as I’m sure it was right in 2008, but it’s evident that the methods he used to create his numbers were based on the points accrued for that season only. Changing the point totals would result in completely different data points. I think it would be interesting to see him do a follow up 4 years later and how those relationships changed.
My thinking is that we would see a significant increase in the value of the shot block based solely around the quick data points I gathered.
What I’m saying is, looking at only blocked shots to wins doesn’t convey a complete picture. It leaves out, what I perceive to be a vital data point, which is the number of quality scoring chances given up.
There’s definitely a stronger correlation between dominating scoring chances than blocking or not blocking a lot of shots.
I don’t think the fundamental premises upon which Mirtle’s post was written have changed much. And I’m sure that there are certain teams that can perform well and block a lot of shots – like we said above, % of total shot attempted blocked is likely informative.
Perhaps I misunderstood, but seeing that the eventual Stanley Cup winner blocked a lot of shots and their opponent VAN didn’t, it stands to reason that blocking shots from the data presented is neither positive nor negative.
I don’t think that picking out those two teams in a 41-game-each sample and then seeing what they did in the post-season tells us much of anything, to be honest.
My bottom line here is that there are a shit-ton of things that correlate more strongly with winning than blocking shots, whether the actual correlation between blocking shots is postive or negative (which it tends to be, since it tends to reflect a possession deficit and associated things that do correlate negatively with winning).
Japers' Rink: Hockey blogging from the most powerful city in the world
My bottom line here is that there are a shit-ton of things that correlate more strongly with winning than blocking shots,
I think this might be a carry over from a conversation elsewhere as I’ve not seen you make such a distinction in our exchanges or in the initial article. If you made that assertion elsewhere I did not see it. So we’re on the same page, I’m not disagreeing with your statement, in fact I’m agreeing with you because the statistic by itself is inconclusive. As I said above, it ends up being much like shots per game. We need additional data points.
whether the actual correlation between blocking shots is postive or negative (which it tends to be, since it tends to reflect a possession deficit and associated things that do correlate negatively with winning).
This is what I’m in disagreement about. Based solely on the data that you’ve provided, it can not “tend to be negative” nor can it tend to be positive. The Mirtle data is irrelevant as it’s 4 years old and based solely around the results from that year and even if it weren’t we would need additional data points to further extrapolate it’s value. So I am unsure as to how you can make your assertion that it “tends to be negative”.
Specifically, from your own data, 60% of teams from either spectrum made the playoffs, so those results are equal. Similarly, your data shows that the teams that went the furthest, BOS & VAN, were from opposite ends of the spectrum. Again, the results are equal. That to me, and as I said above, is significant evidence that to get the true value of those shot blocks you would need to establish a relationship between the number of quality scoring chances and the total blocked shots.
Based solely on the data that you’ve provided, it can not "tend to be negative" nor can it tend to be positive.
I’m not saying this is based solely on the data I’ve provided, I mean it generally. Within the context of a single game, it can correlate positively. Or within a win streak. But over extended, prolonged periods of time, it doesn’t. There may be teams that buck the trend, but the correlation, generally, is negative. And I don’t think Mirtle’s data is irrelevant at all. I think you’ll find in the work of plenty of other writers similar findings over the years.
My “own data” that you refer to is a miniscule sample of half a season with no real depth of analysis on the results, just a quick sample to show that blocking shots =/= winning. My “own data” doesn’t show me anything more, to be truthful – it’s most certainly not “significant evidence” of anything.
Japers' Rink: Hockey blogging from the most powerful city in the world
Further to the point on Mirtle’s data, yes, that chart was only for the current season. But he was asked about it in his comments and responded “I did check the SB/G vs. Points% figures for the previous two seasons and they were nearly identical.” So there’s that.
Japers' Rink: Hockey blogging from the most powerful city in the world
But over extended, prolonged periods of time, it doesn’t.
Perhaps you’re privy to information that I’m not, but I’ve not seen anything conclusive along those lines.
In regards to Mirtle’s work, I respectfully disagree. Those datapoints and correlations are based solely around that year’s data and are only relevant to the teams from 2008. The teams and game have changed significantly since then, wouldn’t you agree? As is the case with any method, if you change that data, the data points, along with the conclusion, change. Or is that not how you perceive it?
Not trying to be argumentative, just want to better understand your thought process. From my perspective a quick glance at the data from this year shows a different story, and were Mirtle to plug the current data in I believe that you’d see very different results.
My "own data" that you refer to is a miniscule sample of half a season with no real depth of analysis on the results,
I’m confused, why do you feel that Mirtle’s data taken from one full season is relevant, but data that’s half that scope minuscule and less relevant?
just a quick sample to show that blocking shots =/= winning.
Oh you and I are on the same page. :)
Please accept my apologies if it came across poorly, but I was only disagreeing with your comments about it having a negative impact because I don’t see that the information you’ve provided, (it could be elsewhere) is conclusive enough as it pertains to today’s NHL.
why do you feel that Mirtle’s data taken from one full season is relevant, but data that’s half that scope minuscule and less relevant?
In part because Mirtle actually looked at the teams’ records and shot-blocking numbers, whereas I just ham-handedly said, “These teams made the playoffs, these didn’t.” He also looked at 30 teams – I looked at 20, including the opposite ends of the spectrum.
And I don’t think I said it had a negative impact (which would imply causation), I said it had a negative correlation.
Japers' Rink: Hockey blogging from the most powerful city in the world
In part because Mirtle actually looked at the teams’ records and shot-blocking numbers, whereas I just ham-handedly said, "These teams made the playoffs, these didn’t." He also looked at 30 teams – I looked at 20, including the opposite ends of the spectrum.
It certainly is more in-depth, and yet when you compare the relationships we’ve been using, Top&Bottom 10, I think you’ll find that there is a difference between quality of teams over the past 3 years as opposed to the quality of teams from Mirtle’s data set.
And I don’t think I said it had a negative impact (which would imply causation), I said it had a negative correlation.
This comes across as though I’ve offended you. That was not my intent. Through what I’ve posted I think it was pretty clear what my meaning was, even if I chose the wrong word.
No, I take no offense at all. Just trying to clarify.
Japers' Rink: Hockey blogging from the most powerful city in the world
In regards to Mirtle’s work, I respectfully disagree. Those datapoints and correlations are based solely around that year’s data and are only relevant to the teams from 2008. The teams and game have changed significantly since then, wouldn’t you agree?
No, I wouldn’t. If anything changed significantly, it’s how the game is officiated. There’s been no great new tactical innovation, no great new goaltending technique, no big step forward in equipment, no major rule changes and three years of player churn. I’d bet on the league being largely the same as it was then.
. From my perspective a quick glance at the data from this year shows a different story, and were Mirtle to plug the current data in I believe that you’d see very different results.
Why do you believe that, other than a quick eyeball test?
I’m confused, why do you feel that Mirtle’s data taken from one full season is relevant, but data that’s half that scope minuscule and less relevant?
Because Mirtle used that entire season’s worth of data, whereas you’re glancing at a couple of teams and not actually running a correlation or any other statistical measure on it. There’s a difference in how thorough and rigorous those two analyses are.
"Essentially, all models are wrong, but some are useful" George E.P. Box
I have $50 to Olie's autism charity or So Kids Can and a beer on the under for 50/57/107 as AO's final stat line for 2011-'12
And on Pekka Rinne's PK SV% dropping under .920 by the end of the '12 season
by Knee high to a duck on Jan 4, 2012 4:12 PM EST up reply actions
I said great! The 1-3-1 sucks and should be hated.
"Essentially, all models are wrong, but some are useful" George E.P. Box
I have $50 to Olie's autism charity or So Kids Can and a beer on the under for 50/57/107 as AO's final stat line for 2011-'12
And on Pekka Rinne's PK SV% dropping under .920 by the end of the '12 season
by Knee high to a duck on Jan 4, 2012 4:21 PM EST up reply actions
No, I wouldn’t. If anything changed significantly, it’s how the game is officiated. There’s been no great new tactical innovation, no great new goaltending technique, no big step forward in equipment, no major rule changes and three years of player churn. I’d bet on the league being largely the same as it was then.
So you’re suggesting that officiating is worse, and that’s the reason for decreased shot totals, power plays and goal scoring?
Why do you believe that, other than a quick eyeball test?
If you don’t want to look at the data then I’ll provide it for you.
Unfortunately it will take a bit, but you will note that the “quality” of the teams within the top 10 shot blockers has changed. You can tell that simply by looking at the standings and real time stats for each respective year.
Because Mirtle used that entire season’s worth of data, whereas you’re glancing at a couple of teams and not actually running a correlation or any other statistical measure on it. There’s a difference in how thorough and rigorous those two analyses are.
Again, if you need me to present that information to you I will. It will take time to make it legible but using the exact same methodology on Mirtle’s data it’s evident that there is a difference and that using data from 2008 isn’t relevant.
So you’re suggesting that officiating is worse, and that’s the reason for decreased shot totals, power plays and goal scoring?
That’s very true. Refs have stopped calling interference the way they’re supposed to in most cases, and they’re allowing defensive players to get away with a fair amount of clutch-and-grab. It’s not just against any particular team, or with any particular refs – it’s league-wide. And It’s been true in the playoffs ever since the lockout.
Failure is always an option.
You’re complaining about sample size and then cherry picking the two teams that went to the finals as an anecdote to disprove the aggregate data?
"Essentially, all models are wrong, but some are useful" George E.P. Box
I have $50 to Olie's autism charity or So Kids Can and a beer on the under for 50/57/107 as AO's final stat line for 2011-'12
And on Pekka Rinne's PK SV% dropping under .920 by the end of the '12 season
by Knee high to a duck on Jan 4, 2012 4:07 PM EST up reply actions
Excuse me? Where do you perceive that I am complaining about sample size? I believe you’re mistaken as at no point did I complain about sample size. The only point that I referenced sample size was acknowledging DH32’s tenure.
As it pertains to Mirtle’s data, I was simply illustrating that using data from over 4 years ago which is based around those rosters, that style of play and those players, is not as pertinent as it’s being conveyed.
by paid in full on Jan 4, 2012 4:11 PM EST up reply actions 1 recs
We need additional data points.
That’s a complaint about sample size, in that you want it bigger or wider or whatever.
Similarly, your data shows that the teams that went the furthest, BOS & VAN, were from opposite ends of the spectrum.
Rebuttal based on a sample of 1/15th the size.
"Essentially, all models are wrong, but some are useful" George E.P. Box
I have $50 to Olie's autism charity or So Kids Can and a beer on the under for 50/57/107 as AO's final stat line for 2011-'12
And on Pekka Rinne's PK SV% dropping under .920 by the end of the '12 season
by Knee high to a duck on Jan 4, 2012 4:14 PM EST up reply actions
I think you’re confused.
Sample size: determination is the act of choosing the number of observations to include in a statistical sample.
A data set (or dataset) is a collection of data, usually presented in tabular form. Each column represents a particular variable.
There is nothing wrong with the sample size for the shot blocking data set, but the lack of a data set for quality of shot blocking is apparent.
Rebuttal based on a sample of 1/15th the size.
You’re taking that completely out of context. He provided 20 teams as point of reference, 10 in each perspective. Of each of those 10 teams 6 made the post season with 1 from each making the playoffs. There is nothing there that supports there is either a negative or positive correlation from shot blocking.
by paid in full on Jan 4, 2012 4:21 PM EST up reply actions 1 recs
You’re taking that completely out of context. He provided 20 teams as point of reference, 10 in each perspective. Of each of those 10 teams 6 made the post season with 1 from each making the playoffs. There is nothing there that supports there is either a negative or positive correlation from shot blocking.
I’m not taking that out of context, those are the complaints you’re making. Using that anecdote to support your point is doing exactly that.
Also, if there’s no correlation either way (and the chart shows a small negative correlation), the blocking shots has no effect on winning or losing, not so? Either way, that undermines the position that blocking shots is good.
"Essentially, all models are wrong, but some are useful" George E.P. Box
I have $50 to Olie's autism charity or So Kids Can and a beer on the under for 50/57/107 as AO's final stat line for 2011-'12
And on Pekka Rinne's PK SV% dropping under .920 by the end of the '12 season
by Knee high to a duck on Jan 4, 2012 4:32 PM EST up reply actions
I’m not taking that out of context, those are the complaints you’re making. Using that anecdote to support your point is doing exactly that.
You and I have some different definitions as pertains to certain words.
complaint:
A cause or reason for complaining; a grievance.
I said:
There is nothing there that supports there is either a negative or positive correlation from shot blocking.
Which is disagreeing with the assertion that there is a negative, or any for that matter, correlation. I then used the same data provided to explain/support how that’s not the case.
Correlation is a testable hypothesis – Mirtle found a small negative correlation in all three years. Hypothesis tested, hypothesis confirmed: correlation exists.
There is a full data set available to you via Mirtle. If you don’t compile it yourself, you have no idea as the veracity of that data set. It’s right there, the correlation between blocks/game and winning is -.12 or something like that. Not much correlation, but negative correlation.
As for
I then used the same data provided to explain/support how that’s not the case.
Let’s call it what it was: cherry picking in a way that does nothing to systematically rebut J.P.
The onus is on you to show that those correlation coefficients no longer apply in a meaningful way.
"Essentially, all models are wrong, but some are useful" George E.P. Box
I have $50 to Olie's autism charity or So Kids Can and a beer on the under for 50/57/107 as AO's final stat line for 2011-'12
And on Pekka Rinne's PK SV% dropping under .920 by the end of the '12 season
by Knee high to a duck on Jan 4, 2012 6:23 PM EST up reply actions
Actually, it’s closer to -.3 or -.4, which a much stronger correlation.
"Essentially, all models are wrong, but some are useful" George E.P. Box
I have $50 to Olie's autism charity or So Kids Can and a beer on the under for 50/57/107 as AO's final stat line for 2011-'12
And on Pekka Rinne's PK SV% dropping under .920 by the end of the '12 season
by Knee high to a duck on Jan 4, 2012 6:32 PM EST up reply actions
Also, as for complaint:
#2 A reason for dissatisfaction: “I have no complaints about the hotel”.
You had a complaint about the analysis.
Now that we’ve concluded that little exercise, did that cilia-fine parsing move the discussion forward any? No? Okay then, moving right along.
"Essentially, all models are wrong, but some are useful" George E.P. Box
I have $50 to Olie's autism charity or So Kids Can and a beer on the under for 50/57/107 as AO's final stat line for 2011-'12
And on Pekka Rinne's PK SV% dropping under .920 by the end of the '12 season
by Knee high to a duck on Jan 4, 2012 6:25 PM EST up reply actions
(and the chart shows a small negative correlation)
You’re absolutely right. It does show a slight negative correlation… with the data from 2008 . I wonder how applicable that data is to teams in 2012…
For example would you use power play percentages or goal scoring totals or win-loss records or shots for and against or any other metric from that year for that matter when discussing teams in 2012?
by paid in full on Jan 4, 2012 5:17 PM EST up reply actions 1 recs
And, if Mirtle is to be believed, from 2007 and 2006.
What makes you think the game has changed enough to invalidate that data?
"Essentially, all models are wrong, but some are useful" George E.P. Box
I have $50 to Olie's autism charity or So Kids Can and a beer on the under for 50/57/107 as AO's final stat line for 2011-'12
And on Pekka Rinne's PK SV% dropping under .920 by the end of the '12 season
by Knee high to a duck on Jan 4, 2012 6:07 PM EST up reply actions
Using Minnesota and their obscene (and as it turns out, unsustainable) sv% probably isn’t your best method for convincing people that shot blocking is now different in the NHL.
This is Fenwick, the same as Corsi, except that it excludes blocked shots. Minnesota is blocking a lot of shots because they have to – they’re in their own zone the entire game. When the other team is putting 58% of the non-block shot attempts towards your goal, you’re going to lose a lot of games unless you have stellar finishing or stellar goaltending. The Wild got outrageous goaltending, now they aren’t, and now they fall in the standings.
They’ve probably accumulated enough points on the back of that hot-streak to make the playoffs, but then again, maybe not. I’d bet on almost any other NHL team against them in a best-of-seven.
"Essentially, all models are wrong, but some are useful" George E.P. Box
I have $50 to Olie's autism charity or So Kids Can and a beer on the under for 50/57/107 as AO's final stat line for 2011-'12
And on Pekka Rinne's PK SV% dropping under .920 by the end of the '12 season
by Knee high to a duck on Jan 4, 2012 4:05 PM EST up reply actions
Well you’d lose. The Wild built around knowing how to win. They are winners. When it matters they will win. Win. Wynn.
Please, call me F&B.
We’ve seen with our own eyes that the Wild have gained 59% of all points available to them over the first half of the season. Until we see otherwise, there is absolutely no logical reason to believe that number will change over the second half. After all, you can’t predict the future, so we must all assume things will stay the way they are right this very second. QED
by HockeyGoalie29 on Jan 5, 2012 11:23 AM EST up reply actions 2 recs
Shots Blocked divided by Shots Attempted
How well does that correlate with P%?
I wonder what the avg Capitals goalie save percentage was and what the teams record was prior to Dale Hunter asking his team to commit to blocking more shots.
No, I don't think I'll ever get over Macho Grande
Terrible. And now it isn’t. But it’s a stretch to say that the blocked shots have had a meaningful impact on the increased in SV%.
Japers' Rink: Hockey blogging from the most powerful city in the world
Implementing this strategy may have had some effect on at least one goalie – to see his team giving up the body like that as an attempt to fix the whole not-winning thing. And may have given skaters an easier opportunity to show they are working hard and feel more involved more consistently over 60 minutes.
Maybe not the best long-term strategy to lead the league in blocked shots, but the gains in confidence and other factors in the short term may pay off in the long term even if they change the strategy or the blocks go down for whatever reason.
No, I don't think I'll ever get over Macho Grande
So the impact is on psychological intangibles like “confidence,” and “esprit de corps”?
Japers' Rink: Hockey blogging from the most powerful city in the world
Where are those triangles on the chart?
J.P.: You might be the king of all geeks here…
by Alz Well That Ends Well on Jan 4, 2012 1:13 PM EST up reply actions
I would like to see the statistic that tells us at what point trend stats/charts become applicable. After 3 games, 5 games? 20 games?
So when a team changes coaches often -but not always -there is immediate payoff in terms of wins. While in one case that may be attributed to better allocation of players, a better system, schedule, someone coming back from injury, etc, it’s commonly attributed to a sort of emotional honeymoon period in general.
In the same vein, things happen in practice, in meetings off-ice, in people’s heads when you suddenly are leading the league in blocked shots. In the short term, there is no stat for conversations between players about liking what we’re doing or not, or a goalie feeling like crap because he can’t stop basic shots and now his teammates are going to get drilled all game by slappers to try to cut down the shots he faces.
But eventually we know people get comfortable, people get into grooves, things become normal. And when things are normalized you then have viable long-term trend statistics. IMO
No, I don't think I'll ever get over Macho Grande
So when a team changes coaches often -but not always -there is immediate payoff in terms of wins. While in one case that may be attributed to better allocation of players, a better system, schedule, someone coming back from injury, etc, it’s commonly attributed to a sort of emotional honeymoon period in general.
But that’s all narrative crap, for lack of a better term. The reality is that coaches get fired when teams don’t get results, and teams very frequently don’t get results because they’ve had a bad run of luck. Often the new coach gets the benefit of that trend reversing.
Is Ken Hitchcock a great coach or did Davis Payne get crappy luck that evened itself out? I’d argue it’s more the latter than the former. Is Dale Hunter a great coach or did Bruce Boudreau’s goalies crap themselves and Dale’s been lucky enough to be here when they regressed to their true skill level? I’d argue it’s more the latter than the former.
Japers' Rink: Hockey blogging from the most powerful city in the world
bad run of luck
That best crosses the bridge I’m crudely making here at the office. Just the first alternate situation I came up with where stats are often hard to apply. I wasn’t trying to ride the thread off into the sunset on coaching changes.
The Noon stat is interesting, I’ve just seen a lot of stats here, and rather than just coyly say “sample size”, I wanted to mention other considerations which could be relevant in the short term, but I’m not in the locker room.
Just like I could say Knuble’s on the 4th line because Hunter felt he needed to take back the locker room and Knuble, who BB always made available to the media may have stepped on his cape, but that would be even greater conjecture, heading into the absurd.
No, I don't think I'll ever get over Macho Grande
After some consideration, I’ve become a little more careful about that. “Bad luck” reversing might mean that the new coach fixed a hole in the defensive scheme, or that a player is nursing a wrist injury or whatever. In the long term, we can’t predict those effects. That said, when it comes to coaching, I bet some coaches have a smaller average time to discovering new wrinkles, or how to counter new wrinkles, than others.
There’s quite a lot of gross luck in hockey, the bounces, but I’d mark a difference between that processes with variable completion time and variable completion quality.
"Essentially, all models are wrong, but some are useful" George E.P. Box
I have $50 to Olie's autism charity or So Kids Can and a beer on the under for 50/57/107 as AO's final stat line for 2011-'12
And on Pekka Rinne's PK SV% dropping under .920 by the end of the '12 season
by Knee high to a duck on Jan 4, 2012 4:20 PM EST up reply actions
I think that also gets to the role of adjustments and how those are so hard to measure because they are inherently short term (small sample) behavior.
Please, call me F&B.
Right. Gabe recently posted that players have longer-term hot and cold streaks than weighted coin flips would account for, which means that they’re hot, or that other teams have adjusted, or they’ve found a whole in most defensive schema, or whatever. It ain’t pure luck, otherwise it’d be distributed the same as pure luck.
"Essentially, all models are wrong, but some are useful" George E.P. Box
I have $50 to Olie's autism charity or So Kids Can and a beer on the under for 50/57/107 as AO's final stat line for 2011-'12
And on Pekka Rinne's PK SV% dropping under .920 by the end of the '12 season
by Knee high to a duck on Jan 4, 2012 4:40 PM EST up reply actions
Maybe I am simplifying it too much, but wouldn’t one except a decrease (or at least just staying the same) in SV% with more blocked shots, since fewer shots make it to the goalie at all?
I would think that would mean more screened shots and the higher quality shots are the ones getting through. So one would expect a lower GAA for sure…but I wouldn’t except a higher save percentage.
I like Laich, but I <3 Green
Keep the faith!!!!!!
by RockinRed4Life on Jan 4, 2012 1:20 PM EST up reply actions
Makes sense. I was just going off that quote – “Having a lot of shot blocks has … a very, very slightly positive relationship with save percentage.”
Japers' Rink: Hockey blogging from the most powerful city in the world
Someone – I think it was Greenberg – noted that the Caps under DH are getting better save percentages and better shooting percentages than they got under BB.
JP posted the PDO stats recently – horrible in the death throes of the BB era, much higher under DH.
That’s why they’re winning games. The narrative-driven shit about heart, character and playing the game The Right Way that you hear from the usual suspects is just that, shit.
Patron saint of quality footwear.
Have to reserve judgment until Don Cherry weighs in…and go the other way.
by ShootTheBullets on Jan 4, 2012 1:30 PM EST up reply actions
I’ve got to say, this noon number doesn’t surprise me from what I’ve seen the last four games. It certainly brings you back down to earth that this team’s problems aren’t solved yet despite the winning streak. Seeing the Big Four reunited and scoring again kinda makes you forget about all that, but Tomas Vokoun deserves a lot of credit for keeping the score so low, because the backcheck is still quite shoddy in my estimation. Way too many extended shifts in their own end.
I see flashes of good D-zone clearing throughout those games, but although it started well against Calgary, the ice began to tilt by about midway through the second, and it’s a sight I’m starting to get sick of seeing.
Capitals goal scored by #22, Mike...
I suspect there is a certain contextual granularity that affects this measure. Lot of blocked shots might be an indicator (taken over the population of games) of a lot of shots directed at your net. It might be (pulling a number out of thin air) a case of 40 percent of all shots attampts faced being blocked. If you give up 50 shot attempts in a game, 20 of them get blocked. But if you give up only 30 attempts, only 12 get blocked.
The key to this might be at the individual game level. If a team blocks a high percentage of shot attempts, it might indicate other things. Maybe we give up only those 30 shot attempts, but block 60 percent of them. 18 blocked shots on 30 shot attempts (at 60 percent) sounds a lot better than 20 blocked shots on 50 attempts yielded.
One is a case of sheer volume (high shot attempt totals, high blocked shots), the other a case of denial (lower shot attempts allowed, higher shares of them blocked).
If you've read this far...seek help.
I need to read this entire discussion in the morning when I have my wits about me. But you had me at:
“a certain contextual granularity.”Visual, poetical, and analytical in one small phrase.
Knuble: "I am what I am. I play well with good players."
On his milestone: "It's going to be like unwrapping a birthday present. Then the day after you're kind of like [sighs]....Now you just have to keep going."
by capsyoungguns on Jan 4, 2012 8:39 PM EST up reply actions
Agree with you, paid in full, and aaw6848. Essentially, the problem is that shot blocks do not exist by themselves. That is, by the time you are in position to block a shot, the shot attempt has already been taken. (When I say shot attempts, I’m including those that are blocked.) Reducing shot blocks by well, not blocking shots, doesn’t do anything to shot attempts against, so you may as well block them. Playing in such a way to reduce shot attempts against by holding possession has the effect of reducing shot blocks, but not because the team tries to block less shots. I think everyone understands that.
But it’s a key point when you try to start doing correlations. Shot blocks are not an independent random variable because they directly depend on the other team’s shot attempts. They may appear so, and with a large enough sample, it’s easy to fool yourself in to thinking of them as such, but they aren’t. So correlating shot blocks with point percentage is certainly going to be misleading unless you account for the shot generation on which they depend. (Arguably, you could raise a similar issue with many of the other statistics correlated versus P%.)
Awesome discussion. I feel the answer lies in multiple variable regression, but if you allow for enough variables to account for the spectrum of topics we have discussed then the analysis becomes useless because the correlation to any one variable is so low.
The reality is that while statistics have a hard time qualifying which shots are worth blocking, they also have a hard time qualifying which players are worth risking blocking shots. If fourth liners and their muck it up brethren want to lie down and take shots all day, I say go for it, but Backstrom and crew better be picking their moments pretty carefully – and Backstrom has had some beauties.
Neil has a nice related piece today: http://t.co/4tR05tEf
Japers' Rink: Hockey blogging from the most powerful city in the world
Comments For This Post Are Closed
Secondary Sidebar
User Tools
FanPosts
Community blog posts and discussion.
Recent FanPosts
FanShots
Quick hits of video, photos, quotes, chats, links and lists that you find around the web.
Recent FanShots
The Vault
In Association With
Latest Puck Daddy - NHL from Yahoo! Sports
- Who are the Non-Smythe candidates for the Devils, Kings?
- End is near for Rick Nash, Blue Jackets as trade discussions set to resume
- Why the Devils and the Kings are evidence the Capitals shouldn’t trade Alex Ovechkin
- Eulogy: Remembering the 2011-12 New York Rangers
- Stanley Cup Final: 5 keys for New Jersey Devils and Los Angeles Kings
Latest stories and headlines from NHL.com
- Marty McSorley hopes LA Kings reverse his curse
- Driven DeBoer delivers Devils back to NHL's elite
- Driven DeBoer, fired by Panthers, delivers Devils back to the land of elite in just one year
- Bruins acquire son of Hall of Famer Ray Bourque
- Bruins add a Bourque: Chris Bourque, son of Hall of Famer Ray Bourque, acquired from Capitals
