I've mentioned Dave Keon hundreds of times on this blog. Keon was a perfect player, the ultimate argument against needing size to play a strong 2-way game. He was a splendid penalty-killer, a quality scorer and durable. I can't imagine a Leafs fan my age having an unkind word about Keon as a player. I can still recall his penalty killing: he'd rub out the right defender who had just passed the puck, and often do the same to the other guy (he was so fast) to the delight of the crowd. He was all about the pressure and the forecheck, but he could also anticipate extremely well. Dave Keon was born to intercept, and he did it for many, many years.
--
Last night during the tsn broadcast, Pierre McGuire stressed several times that Andrew Cogliano is a responsible defensive player. This is untrue. I can see why anyone might believe he could be and in fact Tom Renney has talked about it a few times this season. Cogliano does have some tremendous tools that could make him into a quality 2-way player, but he isn't there yet.
- Renney: "He certainly hasn't ever taken a day off. He's also a guy now who is processing the game as quickly as he can skate it. That's important because some people can't do that. Anybody who comes out of junior or their amateur experience in college and has put up big numbers expects to do that in the NHL. Sometimes, it takes a little longer. He'll find his level. The big thing for him is the investment he's put into being a good two-way player and the effort he puts into that."
--
Last night was a fun game for Oiler fans. I enjoyed it as much as any game this season, outside G1 which was an incredible debut for the kids. Tonight's tilt could be amazing too, but we need to remember that young teams often careen between great and poor performances. There is a feeling about this team, though. They're gaining confidence every game and good things are going to happen.
It's a little like Cogliano's progress away from the puck. You can see good things happening, just not in a straight line. Hey, they're kids. That's what they do. It's a part of the process. Patience.
--
I wrote a post over at ON awhile back about Sam Gagner. Those who have read my thoughts on 89 previously wouldn't be surprised by my views, but it did seem to send a ripple through parts of the blogosphere. Specifically David Staples (a decent and rational man) seemed to take issue with my views (his post is here) and with Gagner's defensive play:
- What do I see in terms of Gagner's defensive play? That he appeared to improve a bit from age 19 to 20, but has taken a step back so far this year at age 21. As a 19-year-old in 2008-09, he made mistakes that contributed to 1.47 goals against per 60 minutes of even strength play. Last year, he improved to making just 1.24 goal-causing errors per 60 minutes. But this year he's been part of the blood bath in the Oilers' own zone, making 1.78 errors per 60. He's done so while often playing with two of the Oil's best defensive wingers, Hemsky and Dustin Penner.
I think David Staples is a fine person and a great writer. My interaction with him has always been positive and I wish him no ill will. I don't think "errors" moves the conversation forward in a rational way. What am I missing?
What am I missing?
ReplyDeleteNobody knows. The inherent belief in the current iteration of the error is that mistakes only matter when they lead to goals against... which means that the player who hacked it up at the blueline and then skated to the opposite blueline to wave his stick in the air did nothing wrong, because Gilbert lifted a stick to prevent a tap-in and then a deflected shot went up into the netting for the whistle.
The stat only becomes useful if you take the current scoring chance view and assign errors based on each chance. Even then, I feel like you end up with a lot of scorer bias (some people will blame the winger for being weak on the puck in the offensive zone, others will blame the center for not breaking up the pass, etc.)
(this is a lot like what Quain said, but since I typed it out before reading... I'll leave it there.)
ReplyDeletere Staples and his error-checking machine... I think he's on to something, but like Cogliano's two-way game, isn't quite there yet. He's cherry picking plays that result in goals and while goals are, well, the ultimate goal, of the game, they are only a small microcosm of the entire game.
As with any sport, there's a limit to the rational thought used in the analysis of the game. You would naturally think that everything evens out and Gagner's higher error-rate this season on goals scored against is indicative of his overall play on the other side of the puck. But, unless you know how many errors he's making that don't lead to goals, it's hard to say.
The two items that concern me are:
ReplyDelete1. WHEN these errors are made. If you make a mistake and Rypien's on the ice, that's a little different than when Sedin's on the ice. That's a huge punishment. I don't see a filter, although David Staples may have allowed for it.
2. We know that luck has an enormous impact on things, so not counting all of the chances for and against (as with Corsi, or Dennis' scoring chances) gives luck a heyday.
I don't really see much of a problem with using errors combined with other advanced statistics.
ReplyDeleteAny extra information is good information the way I see it as no statistic standning on its on gives an acurate picture of a player's worth.
Gagner is a decent young player, but I think even his most ardent supporters have to admit that he isn't exactly killing the ball yet.
Again this season he's been getting the best linemates, with weak quality of opposition, and the best zone starts and he hasn't been killing the soft minutes.
He's getting out corsied by the rookies Hall, and Eberle who have had a harder row to hoe if you go by the metrics.
If you think the Oilers need to strong two way centers to win as I do, then you have to hope Gagner takes a step forward in that regard sooner rather then later. Staple's "errors" stat doesn't seem at odds with what the other stats are telling us in this case. which is Gagner doesn't appear ready to be a bus driver at this time.
Well, just back from Ottawa/Montreal games. Fun trip. If anyone gets a chance to see a game in Bell center, it's a great atmosphere (other than the 1 fingered salutes I received due to my Oil jersey...harsh crowd..haha.)
ReplyDeleteCool tidbit: So I am on the plane to Moncton (had an Oilers cap on) and I notice this guy sitting in same row as me but across the aisle. He had a golf shirt with an Oilers logo on it. Figured he was at the game as well and one of us would start talking about the game but didn't happen.
Waiting in the baggage area , hardly anyone around, I went over and asked this guy if he was at the game, he said no , that he was in Shawinigan last night. I said it was a good game, good comeback. I got thinking geez I know this guy from somewhere. So I approach him again after a few minutes and ask him if he works for the Oilers, "Yep"....I said you are Stu aren't you? "Yep". I I made a comment how Oil Nation was liking his work thus far.
Chatted a bit, was pretty cool. He mentioned the kids are great. Talked about Hall starting to "get it" and should be fun to watch.
He also said (unsolicited by me) "we got some D comming" I said " so you are going to try and land Larsson?"...he chuckled and said "no we can't get too good", at that point my bag was in front of me so had to go, but man did I want to stay and talk some more.
During the flight he was talking to person next to him about Poti and all his allergy issues and he felt bad for him etc.
Anyway thought I would share that, was pretty cool, he must be scouting the Q this week.
The error stat has a bunch of problems that make it near worthless.
ReplyDelete1. Sample size. Since errors are only assigned on goals against, and there are only 3 or so on average of those a game, the error stat only tells us what's happening in about 240 events over a whole season. (Note the difference with Corsi or Dennis' scoring chance numbers.)
2. No control group to check for accuracy. If Staples spent a year looking at other team's errors and we found a correlation between the best defensive players and the lowest errors, the stat would be more confirmed as being more useful. But it hasn't been. (Corsi is recorded for every team. In general the best players have the best Rel Corsi and the goons, AHL tweeners and drowning rookies have the worst number.)
3. Biased sample. There's no check in Qual comp here. I guess you could be divide errors over qual comp.. But my guess is, if you had the Staples error number for every player in the league, you'd find it's almost entirely determined by qual comp. I think if a player faced the easiest qual comp. (and easy zonestart) over multiple seasons and had a worse error number than the players facing the tough opp., then it might tell you that player is struggling.
4. The great thing about Corsi and scoring chances -especially in conjunction- is that they measure who is controlling the game and who is most likely to win. (Assuming equal goaltending.) In the long run, you win hockey games by winning the chances battle and by winning the shots battle. If you were doing that, but your team made more Staples-errors, you'd still have the better team. So it's not clear if Staples-errors are a measure of something that leads to winning.
Oops, two things I forgot:
ReplyDeleteSince errors occur on only ES goals, there are fewer than 240 events per season measured.
And Dennis-scoring chances are now being measured for a number of teams. I believe at least the Flames, and the Panthers. Not sure who else. Ideally, we need more teams to be measured their too.
"We know that luck has an enormous impact on things"
ReplyDeleteThat's a matter of perspective LT. I'd hold that luck has nearly nothing to do with the outcome of a hockey game, there are just so many interactions going on at once that it is insanely difficult to comprehend in any other manner.
I think that Staples hasn't followed the thinking around baseball errors; otherwise, he wouldn't be trying to track them in hockey. It's a century old concept in baseball that is fortunately finally being abandoned by people who know better.
ReplyDeleteErrors in baseball are fraught, of course, because it means that you were at least close enough to the ball to make a play on it. And, as Kris pointed out, the sample size is very small.
"I believe counting only the events that surround goals misses most of the game and in fact punishes those who play against the best opposition too severely."
ReplyDeleteHear, hear. The statistic is overly refined and focuses on a very narrow part of the game. I think it is fun to look at these advanced stats, but with a grain of salt. The more refined a statistic, the more subject it is to the effects of luck which you have duly noted plays a significant role in any hockey game. Luck plays a larger role in hockey than many other pro sports where hands are used (basketball, football) or ample space is provided to allow skill to have a greater effect on the game (soccer). The other sport closest to hockey in terms of luck would be baseball, but only parts of it, and not to the same degree.
How many times have we seen a team outskill, outwork, and outchance another team only to lose the game 3-2? If anything, I find goaltenders are most honest about the luck effect. The winning goaltender on an outmatched team will often end up being one of the stars, heralded as the skilled player who got the win for the weaker team, and you often hear words from the goaltender like "well, the puck was hitting me tonight, it felt good" Translation: I got some lucky breaks tonight on pucks I couldn't see and Ovechkin, uncharacteristically, couldn't get the puck up on those two point blank chancees when I was a yard sale in the blue paint, but I'll take it. I'm not making any promises for tomorrow though."
Fortunately for fans, over the balance of an entire hockey season, the more skilled, harder working, better chancing team will win more games because a game is a complex event made up of lots of little events that will eventually outweigh luck with a large enough sample. That goes out the window when you select out a specific event and highlight them with a statistic, especially without controlling for confounding influences like quality of competition, linemates, goaltending and others. You are more likely measuring the chaos inherent in any hockey game than anything meaningful.
I think any statistic including the basic boxcars and +/- should be viewed with a jaundiced eye. They give a rough idea of what kind of player you are looking at, but they alone tell only a small part of the story. Jason Smith would be an exteme example. Watching a player tells so much more and that is why I like to hear stories about Stu MacGregor making his rounds preparing for 2011 draft week.
He's getting out corsied by the rookies Hall, and Eberle
ReplyDeleteSo is everyone else on the team. Hall and Eberle are #1 and #2 in CorsiON.
Same thing for RelCorsi.
BTW, this is a fantastic sign about these two kids. The counting numbers should come soon.
Hall-Horcoff-Eberle could be a legit #1 line by the end of the season. Jesus Christ that's awesome.
Looking at the numbers is making me rethink Wanderandveer. He's not great, and his game looks like shit, but he's been reasonably effective. Best RelCorsi and CorsiOn of all D-men. And he's facing decent comp. I've been too hard on him.
Smid is struggling a bit. Strudwick is the anchor we all know he is. Gilbert's numbers are coming back up. Whitney is a pretty nice player at his salary and is still just entering his prime.
Unlike you LT I do not like Staples but that is mostly attributed to his slavishly support of a $350 million dollar public subsidy of a private business. Bad idea with money the City of Edmonton cannot afford predicated on proposed commercial development that is very very optimistic at best. Too busy, apparently, developing "assigned errors" to investigate some of the feasibility of the ancilary deveopment attendant to the new arena. Too bad, it would make for a GREAT READ!!!!!
ReplyDeleteI will have to go down to the basement and look for some of the early Bill James Abstracts. I remember being absolutely crushed when empirical evidence showed that Tim Foli and Chris Speires had range factors of about 10-12 feet (and were good shortstos within that range) but Ozzie Smith and Garry Templeton had RF of about 35-40 feet and got to 150-200 more balls a year. Templeton's significantly higher errors came on about a 150 more outs.
Gagner is doing just fine. May top out as a 2nd line center that scores but there is nothing wrong with that. Staples wants Gagner to be an all around #1 center. I want Staples to be a balanced journalist. Guess we are both gonna be somewhat disappointed
And as great as Peckham sometimes looks, he's struggling. I think the predictions that he will be anything more than a third pairing guy with toughness are a little premature. (I know he's a rookie, but he's getting killed.)
ReplyDeleteSmid might be proving himself incapable of handling top 4 minutes without a much superior partner, a la Visnovsky. (That sucks, if it's true. Though he's still young-ish.)
It looks like our forwards are very good now. (See the Corsi numbers for Hall and Eberle.) We need some D because we're burning ELC's, and value deals to Penner and Hemsky.
Wasting this season is unwise. It's going to get harder to tank if we stay healthy.
Yeah I don't think Staples stat paints a complete picture. Many other metrics (stats) are also guilty of this however. He is onto something when he focuses on the plays that matter the most (the goals) but overall I'm with the majority here.
ReplyDeleteAs for Cogliano, I give the Oilers credit for the patience they've shown with him. Back when I used to write (I don't know how you do it LT) I noted the following quote from Mike Babcock:
I guess you put them out there in those situations over and over again until they do it right. You don’t let them off the hook and you make them check their best players. We went through the process a couple years ago we'd alternate who match up against the other team’s best, Pavel would match up, Z would match up and then Draper would match up and we'd go night to night. We'd give them the hardest matchup. Because in our opinion we feel that’s just what you have to do. We do the same now with Fillipula and with Franzen. We think your best players gotta play both ways, that’s the philosophy we have
I like how the Oilers are rolling lines. The clock is ticking on Cogliano however, there are many viable options in the pipeline for his job.
Last nights game was a beaut. Love the fist pump celebrations in the last couple games (Hall, Gagner, Penner)! The team is trending away from a lottery pick but I've thought since the end of last season there is no need to burn another season. Development of the 8 other youngsters is more important.
I think it would be pretty cool if Cogliano has recognized his abilities and adjusted accordingly.
ReplyDeleteUsually guys have to join a second or third team before that sinks in. Would be a real coaching coup if Renney can remake Cogliano into what he is rather then who he might think he is.
LMHF#1 said...
ReplyDelete"We know that luck has an enormous impact on things"
That's a matter of perspective LT. I'd hold that luck has nearly nothing to do with the outcome of a hockey game, there are just so many interactions going on at once that it is insanely difficult to comprehend in any other manner.
So this is basically searching for really complicated non-sensical answers when the simple one is staring you right in the face. It sounds vaguely religious to me.
So this is basically searching for really complicated non-sensical answers when the simple one is staring you right in the face. It sounds vaguely religious to me.
ReplyDeleteI think this sort of depends on how we're viewing "luck". If we're using it to mean "random chance", then there's not a lot of luck in hockey: the puck, hit a certain way, in a certain set of circumstances, will be subject to a certain set of forces and will behave in a certain way. That's not random chance, that's physics. Of course, so is a coin toss.
But if we're using "luck" to mean "factors that are, for all practical purposes, out of the players' control", then there's an enormous amount of luck in hockey. That a player happens to make contact with a puck in one way, leading it to go into the net, rather than another almost identical way, that would have made it hit the goalie's shoulder, is something that the player can't fully control at the speed at which hockey is played. That a gouge in the ice happens to be right in the puck's path rather than an inch to the right of it is not really within a player's control, even as its presence there isn't a matter of random chance.
Riv - "really complicated non-sensical answers"? I fail to see how you got there. The answers themselves would be both uncomplicated and reasonable. The fact that getting there is very complex does not allow for a circumvention through use of the term "luck".
ReplyDeleteAlso, I have no interest in a holy war, but I find the stat-producers are often engaged in pseudo-religious devotion to their particular gathering method and what it supposedly means. There's a difference between what it shows in the literal/statistical sense, and the conclusions that are then made. The first is certainly valuable; the second I often find is done either badly or to achieve a certain goal.
Maybe this is why Tambolowe hasn't been criticized by the MSM:
ReplyDeletehttp://www.tsn.ca/blogs/bob_mckenzie/?id=343662
This is a few steps further than being blocked in by the zamboni.
Wangs World! Wangs World! Party Time! Excellent!
Steve - Well put; those varying descriptions would certainly be an impediment to clear discussion.
ReplyDeleteWhen it comes to discussing why a game unfolded as it did, I'm concerned with those who would dismiss an event such as the impact a big hit in the first period that not only made a defender's arm sore but also intimidated him to some degree has when; in the third the same forward wins a race to the puck against the same defenceman and sets up the game winning goal. There is no stat to show either that this happened or the impact it had, but anyone who's played the game knows this kind of thing happens. A combined picture of this type of game analysis with detailed statistics could certainly be powerful for scouting and analysis, but neither itself is complete.
Very excited for the game tonight. LT's rational warnings be damned I think all the Ontario linked kids take it up a notch in the center of the hockey universe and it is a 7-2 drubbing.
ReplyDeleteThe most entertaining part will be the uproar after the game as another arrow is added to the argument that Big Purple Head Burke has further chained the Laughs to eternal mediocrity (or less). I'm almost starting to feel for the guy....almost.
I'm obviously a believer in the SC's but they have to be measured along with QualComp.
ReplyDeleteI think you take that and sprinkle in some Corsi and you've got what you need.
As for MacGuire, that was a time when I liked the guy and really appreciated his enthusiasm. But then I heard him on a radio station being way off the mark on something about Jarrett Stoll and it dawned on me that the guy was a bit of a poser when it came to knowledge.
now, should we expect that even a paid professional would know the in's and out's of all the teams? Maybe not. But at least they should know about the canadian teams.
a funny thing has happened as basically every game is now available for all fans of all teams: you really don't need reporters to tell you anything about what happened on the ice because you can see it yourself. and I'd rather have someone talk to Bruce or LT on the hot stove about oilers matters rather than milbury or francise. what do they know exactly? they might have sources for the off-ice stuff but on-ice? Not worth talking about.
so, macguire's a bit of a paper tiger in that regard.
no 83 for tonight so fuck that.
ReplyDeleteno 83 for tonight so fuck that.
They should IR him and bring 23 up.
Let Hemsky heal properly dammit!
Wouldn't mind a look at Omark as well.
Khabby starting and no 83?
Not feeling good about it.
Go Oilers!
Still hit them for a unit at +150 as I am a homer and not smart.
*clap,clap*
WG: you don't need to win many games to get your juice back with those odds; Oil at +250 at Ott on the 3 way $line? I mean I know we're young and we can suck but these are the Sens.
ReplyDeleteHere's something Micheal Grange wrote on the Globe's site today about Heatley. Hey, I totally agree but I can just imagine if it was writen in Edm.
--- Don’t get me wrong. Heatley is a tough guy to get behind. The Senators took him in when he needed out of Atlanta after he killed his friend in his Ferrari; paid him well for his impressive performance and watched him bolt when he lost some time on Cory Clouston’s power-play, but not before taking $4-million of the fan's money as a parting gift. It’s hard to dress that up (and by the way, when Heatley says he's moved on, and it's just another game for him, that's one guy who probably means it, given his past).
Paience with 13...ugh. As I've said before, mine has run out. If he can't win faceoffs his use is limited in a checking role.
ReplyDelete89 looks like he's still on his way to becoming a very good player. He's learning the ins and outs of the game and he still has his off nights. It will come with time. His powerplay work hasn't impressed at all this year, though. How many shots has he taken on the powerplay?
Considering we're talking about new fuzzy stats, I figured I'd check out the PowerScoutHockey.com site and check in on their fandangled statistics for the Oilers so far....
Points: L 25th C 14th D 4th
Goals/gm: 20th
Shots/gm: 29th
Hits/gm: 18th
Blocked Shots/gm: 18th
Special Teams: 30th
Team Intensity: 30th
Team Defence: 30th
Dominance: 30th
Team Speed: 26th
Team Performance: 22nd
At Center: 26th
At Wing: 19th
At Defence: 12th
Goaltending: 30th
Super.
And Dennis-scoring chances are now being measured for a number of teams. I believe at least the Flames, and the Panthers. Not sure who else. Ideally, we need more teams to be measured their too.
ReplyDeleteCanadiens, Capitals, Flames, Maple Leafs, Oilers, Panthers, Rangers
in the third the same forward wins a race to the puck against the same defenceman and sets up the game winning goal. There is no stat to show either that this happened or the impact it had, but anyone who's played the game knows this kind of thing happens. A combined picture of this type of game analysis with detailed statistics could certainly be powerful for scouting and analysis, but neither itself is complete.
ReplyDeleteThat picture is not possible for humans to have. But we don't need a picture with that much detail to make reliable, accurate but not perfect predictions about which player will succeed in which circumstances.
Here's an analogy. If we knew which way every drop of water in a cloud-system was moving, we could predict with certainty when it would rain. But we can't do that. We don't have that picture.
Nonetheless, we can make pretty accurate, quite reliable predictions about when it will rain by looking at the big picture and observing trends.
That's what the so called "advanced stats" do. There's absoultely nothing wrong with them, and they tell us a lot about what's going on on the ice. Just like meteorology tells us a lot about what's going on in clouds, even though some of the movements of individual cloud particles can be described as random.
I feel loathe to comment because I am not a natural math guy and can only appreciate the broad strokes of each picture painted. Nevertheless...
ReplyDeleteI do have an appreciation for Staples' system based on the fact that it is a half-way measure between traditional +/- and Corsi/RelCorsi/Fenwick/etc. I find the latter measures harder to get my head around (not a math guy, right) as much as I find traditional +/- too limited and affected by unimportant events (stepping off the bench, skating to the middle and watching the rush score a goal).
I am actually happy to see all the different stats and stats makers try to claim their stake in explaining on-ice events because the more systems fight it out, the better the systems become (comptetition good, monopoly bad). Plus more data is rarely a bad thing when examining a relatively unimportant thing like hockey games.
Hey Bar Qu,
ReplyDeleteThat's an interesting point.
But the way I see the "error" stat, it's really a kind of counting stat analogous to goals and assists, but it's about defense, not offense. (And it's a lot more subjective, more like "hits," than goals. Who's counting hits determines a lot. Goals are objective.)
As such, it suffers from all the problem the counting stats suffer from: e.g. someone could have the "errors" analog of a high SH%. High SH% inflates goals beyond what should be expected for a player. That is, we say a high SH% is just luck (in most cases which aren't Ovechkin) which will run out, i.e. the player will return to a career mean SH%. By analogy, some players will have a bad run of luck being in the wrong place at the wrong time because their teammate screwed up, or because a puck took a funny bounce. These players will have inflated error numbers. But unlike goals -where we know the SH% number- we don't know the "errors" analog of SH%, i.e. we don't know how many of a player's errors are largely due to bad luck.
This is particularly damning of the error stat because the sample size is so small.
Jake that's a cool story. I remember sitting on a plane once from here to Vancouver witha Rangers scout, who had a cup of coffee with the Rangers when Giacomin was still there. It was great listening to his stories and to hear what goes on in the mind of a scout. At that time he said they got a substantial bonus for each kid they found that got drafted.
ReplyDeleteDennis, love that story about Heatly, especially the last line.
I don't know why pro scouts always sit at my seat during pre-game practices. And then they disapear in the hallway to the locker room. It's funny some times the note they write. Remember some guy took a note about Garth Murray who ''Couldn't skate for shit''.
ReplyDelete"The Edmonton Oilers have a lot of problems. Sam Gagner isn't one of them."
LT: You post in a dark dark place.
This made me laugh incredibly. Btw I support the fact, Gagner's not the problem at all.
''Lowetide yes the Oil do have lots of problems and yes Sam Gagner is one of them. You can pull out shayne corsi numbers or whatever to convince yourself otherwise but sometimes the truth hurts. We got unlucky that the 07 draft was so poor.''
Riiiight.
I'm not a big fan of the Staples Error for the reasons that you've stated and I acknowledge that Corsi et al have the best rpedictive power. That said, I do think that there's something that the stats don't show but I have o idea how to quantify it.
ReplyDeletePeople rave about Penner's numbers and I'm not going to dispute them. That said, it's pretty obvious that on some nights he's floating through the game. Sometimes for long strretches. Yea ,maybe his numbers are still there, but they'd be even better if he were actually skating.
I feel the way about Gagner. By my eye his defensive positioning has been better, though is mistakes have had bigger consequences. I don't think that Corsi looks at whether the SA while he's on is line are good SA, but that's what the scroing chances are for.
Long story short, while I don't thik the Error stat matters in any way other than to place blame and evaluate whether or ot a player is making the same mistake over and over, I don't think that there's a good system for saying who a player is or what their benchmark is without also using your eyes.
Wow. Sorry about the typos. Crappy keyboard and lack of proofreading.
ReplyDeleteYou raise a good point kris (sample size), and I think that is what ultimately will pass or fail his method. If others can adopt his stat and use it for their team, then develop a good estimation of an 'error mean.' If it remains his particular stat and no one else can adopt it, then that tells the tale too.
ReplyDeleteCC: yeah that was the kicker, wasn't it?
ReplyDeleteI was reading the whole piece and that line prompted me to post it here.
Dennis that's what always bothered me about Heatly's actions post-Ferrari. He just seems aloof for a guy who was given a second chance that very few people in this life are given.
ReplyDeleteCC: I have been saying that for years. The fucker wanted out of Atl and they obliged him and then he wants out of Ott as well.
ReplyDeleteThis is one of the times when Dany Heatley can thank his lucky stars that the NHL isn't overly popular because can you imagine the scrunity he'd get if he were an NFL or MLB player?
It would be unfuckingreal.
I'm not saying the guy has to go around with a fucking permanent millstone but at some point you quit being entitled while you're ahead.
CrazyCoach:
ReplyDeleteYeah it was quite neat, not every day you get to talk it up a bit with your favorite NHL team's head scout. Was a great way to end the road trip (and RiversQ was asking why the _________ I would fly to see the Oilers :) .
Another Ottawa defenceman on waivers. I was never really impressed with the guy...
ReplyDeleteOTT puts D Hale on waivers. NYI J Colliton clears.
I do think that there's something that the stats don't show but I have o idea how to quantify it.
ReplyDeleteThat's true of everything we can predict with some degree of probability. No? I can predict that it will be cold in Calgary, but sometimes an unexpected Chinook happens. Even so, we're still pretty good at predicting the weather.
People rave about Penner's numbers and I'm not going to dispute them. That said, it's pretty obvious that on some nights he's floating through the game.
That's not obvious. I'd say it's false. He's a big guy so it's harder for him to accelate and easier for him to gas, so it sometimes looks like he's slower than other players or than he was at the beggining of a shift. But it's also his size that allows him to dominate puck battles. Faster is less fat, but less phat, too.
Yea ,maybe his numbers are still there, but they'd be even better if he were actually skating.
Or not. And he is skating.
Ender...
ReplyDeleteThat was a helluva pic you placed on the Magnus post on ON today. Well done, sir!
Dennis,
ReplyDeleteI was telling Brian I was only taking the Oil 3way ml because they can't win in OT or SO this year.
Forgot about that last night and took them 2 way and BANG!
See, luck in hockey.
No way they should be +150 even without 83 and 35 starting.
That's what gambling is all about, taking every small edge you can find and having the bankroll to deal with the variance.
I used to make a reasonable portion of my income playing poker, the variance in hockey gambling is easy peasy comparatively.
SpOILer: I'd love to take credit but I,m not that Ender
ReplyDeleteKris: We'll have to agree to disagree. I really don't think that it's relative speed. I know he takes longer to accelerate and that's fine. But I still think it's pretty obvious when he's skating and when he's not. Underlying numbers be damned.
Love 14.
ReplyDeleteI was skeptical that Eberle would be that good, especially right away.
ReplyDeleteI was wrong.
And 14 has the good stick to relieve the pressure with tired teammates.
ReplyDeletePut the A on that kid.
Get Macintyre off the ice where he belongs.
ReplyDeletesmac retires orr. Nice
ReplyDeletezzzzzz.... fight
ReplyDeletehopefully the end of Smac for tonight.
I think even the most casual readers get that you don't like fighting Kris.
ReplyDeleteYou can give it a rest now.
Dennis it slays me even more about Heatly when I work hard, play fair, and then get that CRA reassessment due to an innocent mistake. HA HA HA
ReplyDeleteGagner just avoids a Staples-error there.
ReplyDeleteI don't mind fighting. I don't like SMac and I think having a fighter who does that and can't play is useless to a team that needs depth and PK'ers.
ReplyDeleteSo...
zzzzz.....
I think even the most casual readers get that you don't like fighting Kris.
ReplyDeleteI kind of do like fighting, though I try to suppress it. But what I like is when, after a cheapshot, one hockey player takes it upon himself to fight the other one, to avenge his teammate, or when a hockey player is frustrated with his team's unemotional play and tries to provide a spark by fighting, or any of the other things that might cause two hockey players to fight.
This ritual "my team's designated non-hockey-player is going to fight your team's designated non-hockey-player, because this is the only reason that either is in the lineup" is just dumb.
I bet my heart rate rises 50% on an Oilers PK.
ReplyDeleteGood thing it usually doesn't last too long. :)
This comment has been removed by the author.
ReplyDeleteLeafs Homer announcers think Jones actually knocked Gustavsson down there.
ReplyDeleteSteve I can get that what you're liking is the show of emotion. I readily get that you could look at that Smac/Orr fight and see it as a purely mechanical thing.
ReplyDeleteI think every hockey fan loves when a guy on your team absolutely pastes that guy whose been a shitbag all game.
And I bet Orr and Mac will fight again.
Teddy Peckman!
ReplyDeleteteddy pecks rulz
ReplyDeleteIs Bruce here tonight?
ReplyDeleteStortini has looked awful so far. ;)
Ender,
ReplyDeleteYou seem to be saying the stats don't tell us everything, because they don't tell us that Penner could skate harder and is lazy, but our eyes do tell us he is lazy?
I just want to be clear here.
1. No one can possibly contend stats can tell you about laziness.
2. The claim that Penner is lazy has no objective evidence to support it.
Can we agree on 1 and 2 and that 1 doesn't say anything relevant about underlying non-boxcar stats?
Helluva sports night tonight. Heatley back to Ott. LeBron returns to Cleveland and of course, the Oil and the Laffs. Remote's going need new batteries by the end of the night.
ReplyDeleteWhat's lebron?
ReplyDeleteNice road period, especially considering G2 of b-t-b. The boys are going to have to weather a storm here, all those free agents that have been dying to sign in Toronto will take it to them.
ReplyDeleteFricking murderer's row. Dammit.
Can we have SMac fight guys after the game?
ReplyDeleteThat way he can be "intimidating" and won't take up a roster spot.
If they record the fights so I can see them I'd be fine with that.
ReplyDeleteStortini had a funny moment in that period. The puck was zipping toward him in the neutral zone and it looked like he was going to make a play on it. But then he kind of went weeble people and it went through his legs.
ReplyDeleteNot putting him down, just a funny moment.
Fair enough Smarmy,
ReplyDeleteI don't mean to sound like a jerk, but whay don't you just watch UFC fights?
They're more entertaining, no?
A large chunck of hockey fights are about as interesting as the fights on old episodes of Jerry Springer.
Smac! Smac! Jerry! Jerry!
This just in. Ottawa hates Heatley more than we do.
ReplyDeleteCoen Brothers new movie is titled 'True Grit'
ReplyDeleteIs that the same as team toughness?
In regards to Staples' metric, as others said, it creates too small a sample size.
ReplyDeleteTake the exact same scenario as a goal 50 times and how many times is it a goal? 5, 10, 20, 40?
You don't know, that where the short term luck comes into play.
How many times out of 50 (or 50,000) does Price get his toe on 6's shot yesterday?
Its unknown, so you count items that can result in goals.
Shots are good, Corsi is good.
Fenwick is better because it removes some low percentage shots.
Scoring chances is even better than both because they are counted with a critical eye.
With Staples system you only get penalized when the result is a goal, and in hockey that happens far too infrequent to be counted with meaningful results.
Fuck, that was a bad period by the Oilers.
ReplyDeleteIt's not that I'm overly surprised because I've seen it a lot this year and yes I know it's BTB - I don't think it doesn't matter as much as you think though given the distance travelled and the age of our team - but that was a bad period given the comp.
The odd thing was how few chances the Oil PK gave up given all the goddamn visible seams; the saving grace was the Oilers had some good sticks.
Speaking of the PK, Oilers are running with 10/89; 13/27 and have put together the pair of 16/28 as the third pairing which means 14's off the PK - I guess he was scoring too many SHG;) - and i guess that means HEH is coming off the bench for the first post PK shift so that's not a bad plan.
Nice sweeping stick by 91 at the tail end of that breakway by 91; the kid just gets it from the perspective of his own end and does seem like the kind of fellow that will end up at some point on the PK.
Four or five great chances by the Leaves and good reflexes and positioning by Khabby but poor rebound control.
I'm looking forward to seeing if the boys bounce back or if they're feeling satisfied.
lastly, one thing TO seemed to have figured out early was the way to get this team is to hold onto the puck and wait for the coverage to break down. Last night CTL seemed more preoccupied with scoring nice goals off the rush.
WG: don't ask me why this amount but I put 17.50 on a Oilers and CTL in reg parlay: odds +797
LeBron is the prototype of the modern day primadonna athlete and tonight he gets some serious comeuppance.
ReplyDeleteAnyone know why the building looks half-empty?
ReplyDelete13 makes one more step there and it's a legit dump but instead a blind backhand pass and an icing,
ReplyDeleteFair enough Smarmy,
ReplyDeleteI don't mean to sound like a jerk, but whay don't you just watch UFC fights?
I don't like watching mostly naked men on the ground humping eachother. (Uh I mean wrestling eachother)
And yeah Dennis. The Oilers have been really awful. They'd be less tired if they didn't spend entire periods skating around in their own end.
What a goal by Hall. Keon backhand.
ReplyDeleteI love 4 as well.
ReplyDeleteHoly cow.
ReplyDeleteTHAT is quality.
ReplyDeleteHoly shit. AND the other 4. Beauty.
ReplyDeleteJesus Christ, now THAT was a backhand.
ReplyDeleteI openly admiringly cursed in reference to the speed when 4 cut that corner but that backhand was equally as impressive.
God damn that was a lovely backhand.
ReplyDeleteYou'd rather they hump each other standing up with their clothes on while hugging each other. You PRV.
ReplyDelete:>
we made Brian Burke's goalie cry.
ReplyDeleteYou'd rather they hump each other standing up with their clothes on while hugging each other. You PRV.
ReplyDelete:>
I like boxing too. Probably a joke in there somewhere.
Paajarvi took that puck away like he would take it from a scrub like me.
These kids are going to be beasts by the end of the season.
Kane and Toews.... Kane and Toews... I mean Eberle and Hall... Eberle and Hall.... definitely a sick backhand and burst of speed.
ReplyDeleteHall just has that look in his eyes... The guy's here to compete.
WG: don't ask me why this amount but I put 17.50 on a Oilers and CTL in reg parlay: odds +797
ReplyDeleteThat's some quality degeneracy there.
ATL and TBY were also giving great odds tonight.
If Brian had a house, it was on TBY tonight. I prefer ATl because betting Pavelec has bought some beers this year.
Hall just has that look in his eyes... The guy's here to compete.
ReplyDeletePart of the reason I leaned to wanting him in the draft. Not many players have his talent and that intensity.
Oilers definitely look like a team that's playing 3 in 4 nights. Maple Leafs have gotten away with a bunch of slashes tonight; guess we're paying for the Montreal reffing tonight -- not that the penalties we've taken were anything but.
ReplyDeleteWatching the Leafs broadcast at Alpha 2 Zulu Sports. Man, those guys are brutal.
ReplyDeleteKris:
ReplyDeleteI'll give you 2 and 1 in principle. I was really just saying that there are some things that don't readily show in stats, but can be pretty obvious by sight. Laying everythung on the numbers may help you gamble, but it's not guaranteed to tell you whether Gagner's defensive positioning has improved or whether or not Penner would likely have higher numbers with more consistency (and don't take that the wrong way -- his consistency that way has gotten a lot better over the last few years.)
So Errors will tell you if a player keeps making the same mistake and Corsi will tell you how much you give versus how much you get. But since there are 4 of your teammates on the ice, your numbers aren't necessarily indicative of your defensive ability. Especially against elite players like the Sedins. Scoring chances are likely a better measure in this regard than Corsi since they'll at least give you some indication that the shots are in or out of danger areas. All of these bits of math are great and predictive and all. But relying on the math will cause you to miss things like positioning, and when you're talking aboutu improvement in defensive skills, sometimes you care more about the raindrops than the storm.
I agree with LT that RelCorsi is the best numeric method we have for this. I just wonder if the disconnect when people are talking about this is the difference between improvement by the numbers or improvent by positioning.
Or maybe it's just a disagreement as to whether those are two different things.
Would love to see MPS pot one tonight. Hall & Eberle shouldn't have all the fun.
ReplyDeletewould like a pan up behind wilson there...
ReplyDelete-'whatcha been feeding that thing hauser?'
-'blondes.'
Bah, my feed died just before that rush.
ReplyDeleteGaggles got a nice garbage goal there.
ReplyDeleteMost hockey fans must agree with Kris that fighting is boring otherwise 18,000 of them wouldn't be all getting up to stretch at the same time whenever a scrap breaks out.
ReplyDeleteYou know what really makes me zzzz?
You know what really makes me zzzz?
You know what really makes me zzzz?
...repetetion...
hope there's another segment of 'FUTURES'. now that's a great name for a segment!
ReplyDeletespeaking of futures, we're kicking the shit outta the leafs.
Watching the game and perusing Leafs boards at the same time is like triple the fun.
ReplyDeleteGood god this guy Gene is interviewing is really old.
ReplyDelete1st period chances: 3-8
ReplyDelete2nd period: 7-8
So, a much better period by the Oil and I'd like to see them get close to the breakeven level just to keep this nice streak going of at least being in the ballpark chances wise.
Is it just me or has 89 been scoring a few goals like that? I remember one at Chi for sure and the little fucker isn't afraid to dive in there and get after it and good on him.
It's been such a long time now where we're watching kids and stupes that you forget how goals can be scored sometimes, you know? That last rush you see 91 pound a low slapper off the pads and thus creating the chance for a rebound and 89 digs in and digs it in.
That's like an old Dallas Stars goal scored by Mike Cunt Keane.
Stuck listening to Ched tonight and Jesus Christ I dislike Replacement Rod more every time I hear him.
ReplyDeleteAnyway, does anyone have a working video feed? ATDHE has been getting pretty patchy lately.
Vandermeer +3 through 2 periods?
ReplyDelete"Is he that good or are we that bad?" (Shoot 'Em Up)
We've got a few Winnipeggers and Manitobans in the house, right?
ReplyDeleteOKC are there on Fri and Sat so it would be nice if we got some eyewittness reports.
Looks like atdhe has changed the feed, its now a veemi feed that seems to be working.
ReplyDeleteTSN not showing a Leaves game? Another reason for Edmonton fans to hate Heatley.
ReplyDeleteAlso to add, someone started using Inebriated Oiler Fans again. Almost made the mistake of publishing the November prospect statistics there as a result.
Dennis:
ReplyDeleteWe're getting liquored and attending the Friday spectacle actually. Planned it and everything.
Not so sure how reliable our drunken rantings will be, but I'll definitely have something up over at Bubbling Under.
Dennis: Showerhead, Dorito and I are going to be at the game Friday. Looking forward to it as it will be the first time I've met either of them. Pretty stoked to see my first AHL game too.
ReplyDeleteI pity the Leafs fans, having to watch a bunch of teenagers schooling their doomed hockey team.
ReplyDeleteWait, is JFJ a healthy scratch? Or is he hurt?
ReplyDeleteI really shouldn't have to ask, but hope springs eternal.
Wait, is JFJ a healthy scratch? Or is he hurt?
ReplyDeleteIt would be a first if he is.
...Healthy, I mean.
I'm not normally a 'fire the coach' guy but it's hard to imagine the Leafs not being better without Wilson.
ReplyDeleteWas that Paajarvi that picked off that Kessel pass there and got it up ice?
ReplyDeleteWhen the feed is this grainy it helps when the play by play guy can name the players on the ice.
Also lordie the Leafs look lost. Could the Phrankeneuf and Armstrong injuries really be hurting them this much, or is this really Wilson being turned out?
That play was awesome for so many reasons.
ReplyDeletewhat the..
ReplyDeleteLeaf nation not happy. lol
ReplyDeleteLoving this road trip :)
Not quite Evers to Tinker to Chance, but Eberle to Horcoff to Hall rolls off the tongue pretty smoothly.
ReplyDeleteI've liked Paajarvi tonight, hah.. Jones scores. Wow.
ReplyDeleteBack to Paajarvi, he's looked hungrier tonight, as all the kids have tonight.
3.0's gonna be cracking heads tonight. This whole trying to tank thing is starting to look as successful as Edmonton's previous attempts to make the playoffs. They just can't win... at losing.
ReplyDeleteBoozeabulin SO ? MVP!?
ReplyDeleteWas watching the Habs game.
Are the Oil getting monster chance or simply the Leafs netminders horrid?
1st shutout in TO in franchise history!
ReplyDeletewe've NEVER shut out the leafs? hilarious.
ReplyDeleteListening to the game on radio here and they interviewed Kadri and methinks the Leafs figured they had it in the bag after P1 to hear him talk.
ReplyDeleteI get the impression, living here, that the Leafs are a) not very good and b) not very mature as a team and as individuals. I think they figured this to be a walkover.
The one guy I'd take in a heartbeat is Kulemin. He is terrific. But up front other than him they haven't much going on. I'd take the Oilers' forwards and that's with the holes we have in the bottom six.
No question.
I'm going to be honest: I'd rather finish 10th in the west than 15th. Not sure if 10th is a viable goal, but it would be nice to have hard evidence that the Oilers are on the upswing, especially if that would encourage the braintrust to plug holes in the off-season.
ReplyDeletewe've NEVER shut out the leafs?
ReplyDelete...in Toronto.
OK, great win and all that, but good lord, Jack Michaels is an embarassment.
ReplyDeleteHe calls a hockey game like it's the main event of WrestleMania.
agree with Kris
ReplyDeleteWell, that's all I care about.
BD, would you take the Leaf's D-Corp over the Oil's?
ReplyDeleteIt was more Oilers with some good finishing touches than bad goaltending.
ReplyDeleteHall in particular showed some excellent hands with the backhand and deflection goals.
Wilson may be gone soon. I was at the Oilers Leafs last year and the Oilers were just so poor.
ReplyDeleteHere we are just months later and they come in and wipe the floor with them.
Burke has to do something or its Kessel for two lottery picks.
Which is hilarious any way you look at it.
I'd rather the Oilers finish 25th to 30th than anything else.
ReplyDeleteIf you're not good enough to compete now, and you have this many holes, might as well get another blue chip. Any one of Couturier, Larsson, Landeskog, Nugent-Hopkins, Musil, Siemens would go a very long way.
Im with you Steve. Dont care about tanking anymore. Losing sucks hard and winning is awesome sauce.
ReplyDeleteHave only watched the 3rd, but man I don't see how the Oilers can finish last this year.
ReplyDeleteAnd I'm not saying the Oilers are that good. The Laughs are just pitiful.
Didn't Burke say something about seeing progress since he's ruled over the center of the universe?
Can Bettman please create a new Canadian division without Van?
Edm
Cal
Tor
Ott
Mon
Agreed, Dave. I just can't listen to that dude. I picture him tearing at his cheesie-stained shirt hulk-style and yelling "RAAWWWRRRR!!" every time anything more exciting than an offside happens. Though come to think of it, he even manages to over-emote the offside calls.
ReplyDeleteThe other thing is that a tenth place finish in the west would create the expectation that the goal next year is playoffs, and it would be really nice to be able to hold this management team to some expectations.
ReplyDeleteHall's hitting the target. What's going on? The fall for another Hall is about to be foiled by Hall.
ReplyDeleteLee - yeah I would despite tonight's results. Komisarek and Phaneuf are overpaid but Schenn Aulie and Gunnarsson are all young, Beauchemin is a nice vet and Kaberle is a nice playmaker.
ReplyDeleteBar Qu: I am a stats guy, of a sort. It's pretty simple to explain. Predictive power is overrated. If you oversift the data, all kinds of weird coincidences emerge.
ReplyDeleteYou CAN'T HAVE a dataset completely free of weird coincidence. When that happens, statisticians write a letter to the provost, and someone gets their research program suspended.
When people try to make up strings of random numbers, they fail to generate repeated digits as often as chance implies. We have a strange psychological notion of chance as coincidence free. In fact, there are lots of patterns in chance, but they don't mean anything, unless you have a beautiful mind in amplification mode.
It's a big problem. How do you tell the difference between weird coincidence and real effects?
In clinical research, it's by rigorously restricting how much you sift. You get to state ONE hypothesis, before you collect the data. Investigating fresh hypotheses AFTER you collect the data in order to get a publishable confidence level is severely frowned upon. You can ALWAYS find some weird thing at a putative 95% confidence in random data if you look hard enough.
In data mining, you accept any hypothesis that shows up, but your confidence level needs to be more like 99.99 and you're still suspicious until you collect more data. You don't approve new drugs on circumstantial evidence from data mining.
Even valid effects regress to the mean. Any spike large enough to notice is a fluctuation on top of a smaller spike that will hold up for longer, you hope, if it doesn't melt away like ice cream. Basically, anything that jumps out at you is boosted by chance nine times out of ten, and fairly often, is nothing but chance at all.
What's the use of statistics, then?
The sober application is to control for bias. This is done by working with large samples full of obvious chaff, like a goal scored 0.5 seconds after the power play has ended being counted as 5 on 5.
The problem here is that it is almost impossible to remove chaff without introducing bias. As soon as human judgement enters the picture, bias comes with it. Our intelligence is designed to seek significance. The world is too big and messy to cope with otherwise. Only a computer can flawlessly recall a billion observations in their raw state to reevaluate "saw him good".
A statistic such as errors on goals scored that comes from a small sample size pruned by human observation will be at least 50% bias. A statistic that is 50% predictive, and 50% bias is like taking two compasses on a ship and having them disagree. All mariners agree: take one compass or three.
The nice thing about weak, unbiased statistics is that you can add bias at any point in the process. Bias can wait until you roll the press. The only thing that's harder to do is to add bias while pretending you're not.
This doesn't have much to do with numbers. It mostly has to do with coincidence, bias, and human psychology.
(continued)
Which poster was it that said Hall doesnt know how to hold his stick?
ReplyDelete(continued)
ReplyDeleteWhat a numbers guy will tell you is that even when you're bending over backwards not to add bias (or the pretence of such), you can still mess up, especially if you're fond of the conclusion you end up with. That was the crux of the Stephen McIntyre hockey stick debate.
The nature of the filter Mann employed in the tree ring data caused the analysis to produce a hockey stick curve even when you fed it nothing but random numbers (red noise). Once corrected for this, 50% of the published warming effect went away.
Mann argues his work is still valid. Maybe it is. I think he's a statistical slob for allowing bias to account for half his original effect. It I wanted that kind of sloppiness in my analysis, I can get that from corporations and politicians. A scientist who is content to be merely right is not worth listening to.
Last night, some sucky anonymous pundit was accusing Cammallerie of not manning up after his gaffs.
http://montreal.ctv.ca/servlet/an/local/CTVNews/20101201/mtl_habshub_fever_101201/20101201/?hub=MontrealSports
On the climate debate, I sure wouldn't want Michael Mann in my dressing room.
Some of us take a hard line on bias, other prefer the scintillating amplification of chance. Statisticians tend to have low tolerance for drama queens. Ordering a black and tan is about as risqué as I ever get. Groundhog Day is a documentary. Once you correct for bias, every day is a slow news day, with rare exceptions.
I love this line from Wikipedia:
I’m about to give you one of my all-time favorite statistics: Rickey Henderson walked 796 times in his career LEADING OFF AN INNING. Think about this again. There would be nothing, absolutely nothing, a pitcher would want to avoid more than walking Rickey Henderson to lead off an inning. And yet he walked SEVEN HUNDRED NINETY SIX times to lead off an inning.
OK, now there's a stat with the power to jack the drama queen in the dullest of statisticians. This is not just some guy on the ice for 30 goals scored in the entire season, passed through the filter of human judgement, who is already suspect for having nothing better to do, anyway.
The Baby Oil snuck up on some teams in the East, a couple of whom aren't very good at the moment. BUT the PK is still historically bad, we're relatively injury free and getting outchanced last crazy nearly every night. Fun times right now but the lottery's still well within in reach.
ReplyDeleteThe kids looked GREAT tonight. Toronto looked good in 1st period and then quit. Time for a coaching change in TO?
ReplyDeleteIf the kids continue to get it, we are in real tough for the lottery. Eberle is one smart little player
The OIlers will be playing Western Conference games soon enough.
ReplyDeleteThey would make the playoffs in the East, by 14th is about the best that is possible in the West. They should be able to hang with the pack though.
Courturier is looking impossible now. The Islanders and Bruins and Devils need a defenseman more, but if Larsson is falling off, Courturier is a sure #1, regardless of need. Is that a blessing or a curse?
Maybe its time to offer Hemsky to the Canucks for Hodgson plus to get the right shot centre. The Canucks window is now.
Oilers O with Leafs D would be a pretty formidable squad. Throw in Montreal's G to put them over the top.
ReplyDeleteErrors would bed useful if they were measured for the entire game regardless of goals. Every short pass, every lost battle, evrey bad change, every loss of coverage. Everything.
ReplyDeleteLT said it best. Something like the error stat is equivelent to someone shouting out the score to you from another room in your house. It made sense to me, anyways.
Godot10, the Nucks may as well throw Bieksa in that deal too because that's the only way they can free up enough cap space for Hemsky that I can see.
ReplyDeletenice to see the boys lock it down for the SO for the vet.
ReplyDeleteFirst round's on Khabby!!
Too soon?;)
Hemsky for Hodgson would be an awful deal. Full stop. You KEEP 83, you don't deal him. End of discussion.
ReplyDeleteAnd I didn't want to be the over-optimistic jackass to draw the parallel, but here it goes - everyone's seen that game on ESPN Classic from 1980 when a young Oiler squad went into Maple Leaf Gardens and kicked the snot out of Ballard's Boys 8-3?
Someday, four or five years down the road, I hope Jim Hughson has to draw a parallel from that game to tonight's when counting down the final minutes of a blowout Oiler win to clinch a Stanley Cup.
I agree on the play by play guy. A couple things... he almost always calls the team "Edmonton". They are the "Oilers" pal... we are the Home team. What is more Canadian than "He Shoots, He Scores"!!! You don't say he "puts it home" or he "taps it in". I thought maybe I was ready to move on from Rod but when you hear him call that OT winner last night in MTL then listen to this guy...
ReplyDeleteWhy did they feel they had to go to an American boy to do this gig? That's why it sounds like wrestling.
In fairness - and it pains me to say this, because I love Rod Phillips - he referred to the 3-2 goal last night as "the tying goal". And that wasn't all that atypical of his last years behind the mic.
ReplyDeleteI'm tuning in to all the games Rod calls this year. But if your objective in listening to the radio is to know what's going on in the hockey game, Michaels is miles ahead of him.
Not comparing the new guy to Rod... he's a Hall of Famer and not fair to either one of them to compare. But.. I'm just saying...
ReplyDeleteThis winning is nice but if the draft was tomorrow we wouldn't have a lottery pick. Shit.
ReplyDeleteClarkenstein
ReplyDeleteNever worry we can start sitting the top 3 D men rotationally and dress Strudwick...play Fraser 15 minutes a night and dress and play Smac
Lottery lock......even if the kids "get it"
Pat: Bring back Souray and pick up two guys who can kill penalties - one of them must also be a FO monster - and this team's already good enough to be mediocre.
ReplyDeleteand then every chance they are in the playoffs next year.
Fun game to watch.
ReplyDeleteLove watching the kids get more confidence.
Love watching first passes out of the actually hit a guy on the tape.
Love watching 89 become a real NHL player.
They've beaten two teams who won't make the playoffs in the East and one team that wouldn't make the playoffs in the West in the last three games.
Next few games aren't against ball busters either.
Go Oilers!
*clap,clap*
The one guy I'd take in a heartbeat is Kulemin. He is terrific.//
ReplyDeletethe last 15YR in rsl/khl u22 (undeveloped core strength) forwards w/ 20 goals.
Ovechkin
Malkin
Frolov
Semin
Kulemin
Omark
Going thru the stats on Elite prospects this summer those are the ones i could find.
boopronger - I did. Because he was. He's improved that and is now carrying the puck around the opposition on the rush much better.
ReplyDeleteLets not get too excited. That was an awful game by the Leaves D. Smart teams stand us up at the blue and force the dump and retrieval. The leaves dropped off all night long and allowed us to rush the zone.
ReplyDeleteIt will be back to WC hockey soon enuff.
I remember when Mario came out of retirement and his first game back ESPN had a Mario-cam, where they isolated him onscreen whenever he was on the ice and you got to see his every move. If they had a Burke-cam during the Leaf games you could watch him go from smug and insufferable to exasperated and profane. I really can't stand the fucker and when the Leafs finish in the lottery again and the Bruins get another great young player, maybe the media will finally stop blowing this overrated tool at every opportunity.
ReplyDeleteI'd rather the Oilers finish 25th to 30th than anything else.
ReplyDeleteAgreed. But its fun to see them win (and be tied with Calgary!).
~Shouldn't someone be complaining about Strudwick right now? :-)
Another pitiful night for shots too. What would team corsi look like without young 18 yr old Hall firing attempts every shift?
ReplyDeleteThis team can still finish last.
Holy Crap - I miss the game because of a meeting, then realize I forgot to record it, so I just pop open Oilers.com to see the score expecting the short bubble to have burst and BOOM - there it is 5-0 Oilers.
ReplyDeleteNice
Late to the party.
ReplyDelete1) Zack sucked in his 3:27 tonight. Carrying around SMac may have had something to do with it, but it wasn't his best night. That line was brutal, and Renney didn't give them any ice at all even with the big lead. So it is; the sun don't shine on the same dog's ass every day.
2) Very nice win for the boys in the COTU tonight. 5-0, with the rooks lighting it up and the fans calling for Wilson's head. Good times.
Uppers:
ReplyDeleteHall with 2 goals. Eberle looking great. MPS shows up. Gagner and Horcoff look solid. The cup is ours! Khabi plays okay.
Downers:
-The leaves suck, but they outshot and outchanced the Oil, both early when the game was even and when they were trailing.
-The Oil didn't have to PK much, which could've resulted in the laughs scoring 10.
This comment has been removed by the author.
ReplyDeleteThis comment has been removed by the author.
ReplyDeleteThis comment has been removed by the author.
ReplyDeleteAll this piling on of Staples Errors, and not one person has mentioned that for the last two years now David has applied the exact same principles towards counting every scoring chance, for and against, as he has on goals for errors (or unofficial assists). The sample size isn't 5 events a game, but 50.
ReplyDeleteThis accomplishes several things, among them that the ratio of errors causing goals to individual scoring chances allowed can be calculated as a ratio to test for luck. There's your analogy to Sh%.
Furthermore, Sh% isn't just luck. For some it's a repeatable skill. For others, committing especially bad clangers that are more likely to cost goals is a repeatable trait as well. But you're never going to suss it out unless a) you evaluate and count, or b) you're quite content to rely on "seen him bad/good" as your guide. I'm pretty astonished by how many in these parts seem to prefer #2, let's put it that way.
To a degree errors are comparable to goals and assists which also by definition have a small sample size of ~240 events a year. Yet we place a huge proportion of our evaluation of offensive play on such outcomes. We've all seen a guy who created nine chances one night and get nothing, and the next night he creates one chance and gets two points. It happens, and such things will start to balance out over longer sample sizes like full seasons. Still there will be an element of luck in there. We all learn to expect that the data is always going to be corrupt to some extent.
Errors are subjective and points are not, but how often have we seen a play where some dude makes a D to D pass in his own end and gets a point, which the winger who drives the zone to open up space, screens the goalie and gives the shot a chance to go in, gets nothing. Both methods have their weaknesses, and no method is perfect.
David's individual scoring chances are to Dennis's team scoring chances as points or errors are to traditional plus/minus. Both are judgment calls in that the scorer decides whether or not it even is a scoring chance. This year at least, Dennis and David have had very similar net totals of chances game after game. The difference is in assigning them to everybody on the ice, or just those adjudged to be involved, pro and con, in the chance. So there's a second level judgement call in the one method that doesn't exist in the other.
As with traditional +/-, identifying everybody on the ice is useful for some applications, although not all players are involved in a typical play. Identifying those players who ARE involved, especially after tedious review of every play at net for an entire season, is a judgement call, sure. It's also a hell of an effort that deserves a lot more respect that it seems to get.
... (conclusion to follow)
Fun game. Love beating the Leafs. If only Mats Sundin were there to suffer it. He seemed to always kill us somehow in the past.
ReplyDeleteCogliano took 8 draws and lost 7 of them. Awesome.
I'll be the first to say I don't always agree with David's conclusions. Things like how many goals/60 Sam Gagner has been responsible for in the first 20 games are going to be just as subject to small sample bias as calculating how many points per 60 a player is getting over a similar period of time. I sure wouldn't be prepared to say the player is deteriorating based on the numbers cited in that one example. Over such spans it might identify trends and hot streaks but it sure in hell isn't enough on its own to identify talent. Not in 20 games. But sometimes it supports what the eye can see; guess what, folks, Jason Strudwick has a lot of errors!
ReplyDeleteThe last point I'll make in this sphere is that errors and individual scoring chances are not be-all and end-all stats in my view, but complementary stats that add to the overall body of data. I use them in conjunction with Dennis's scoring chance stats, with shots data, Fenwick, Corsi, QualComp, ZoneStart etc. They're just more arrows adding yet another dimension to the analysis. When all the arrows point the same way, it's confirmatory. When they don't, it's time for another look, at the player, at what we're trying to measure, at what we might be missing, and at the methodologies being used.
Face it, Bruce. You're just a stats junkie. They could come up with a stat that counts which shots come from the left and right sides of the players stick blades and you'd be all over it. ;)
ReplyDeleteWorst game of the trip but sometimes that's the way things go.Took a look at the standings tonight for the first time this season.Seriously.
ReplyDelete6 game home stand with (mostly) middling competition.I'm not sure what to cheer for-I have already come to terms with acceptance,I can't go back to hope...I just can't do it.
A house?
ReplyDeleteHa.
I only had two units on Tampa!
Thankfully made up for it with 2 units on Edmonton.
Sullying my good name Darcy....
Fun game to watch. Love that it was the kids who lit TO up too... rubbing their noses in it a bit.
gagner impressing again..
ReplyDeleteand Hall.. wowzaz..
Haven't read the entire thread, so maybe someone mentioned this, but part of LT's critique is about a few months out of date. ..
ReplyDeleteI agree that errors on goals against onl measures so much. More data is neeeded.
So this year, I'm also measuring errors on all scoring chances against.
You will find it in this document under the "scoring chances against."
So on these chances, the player made some mistake that contributed to the chance against. Not sure if that advances things or not. But when I criticized Gagner's defensive play, I also had this stat in mind. He's leaked more chances than Horc, at same level of as Gagner.
https://docs.google.com/Doc?docid=0ARMC4T3agcfmZDQ2cTRxOF82MWpnMmcyZGNn&hl=en
#1LMHF
ReplyDeleteI agree that some of conclusions drawn from stats are wrong-headed, and I'm guilty of this as the next guy, or more than the next guy.
That said, I've never been able to wrap my ahead around applying a team stat like Corsi plus/minus, goals for plus/minus and scoring chances plus/minus to an indvidual player.
I know this data is readily available, handy, convenient, objective all that good stuff, but there's way too many false positive and negatives for it to be fair and accurate when applied to an individual player.
So I'll stick to trying to winnow outt he false positives and negatives, for what it's worth, both on goals for and against and scoring chances for and against.
I've long accepted that this work isn't everyone's cup of tea and am content with that. At the same time, I was happy to learn that Billy Moores of Oilers and Rangers were doing similar individual plus/minus work on goals and scoring chances in the 1990s. So I'm not the first person to have this crazy idea, and not nearly the most hockey savvy.
Despite my concerns above, I have to say what a delight it is to cream the Leaves. Thats a H2H record that needs to tilt the other way in my lifetime. Looks good on The Fire Hydrant too.
ReplyDeleteBattin .500 last ten games... I'd be happy with that the rest of the way.
This comment has been removed by the author.
ReplyDeleteI don't know what it is about me, but when it comes to announcers, it's Shaft vs. the World.
ReplyDeleteI think Bob Cole has a legendary voice. I acknowledge that the last few years has featured many occasions when he has no clue about what's happening on the ice, and he loves the Leafs like I love warm cinnamon buns, but his voice still gives me goosebumps all the same. And to me, the voice counts for a lot. I'll never forget his call of Salt Lake 2002 Gold Medal Game, which was the day I went from hockey fan to crazed lover. I also learned that day that Joe Sakic was my favourite hockey player.
But I'm still surprised to hear all the anti-Michaels sentiment. Rod Phillips hasn't had a clue about what's going on the ice for some years now (and that's a lot more important on radio than TV), and his Oilers homer-ism is too much for me to handle. Jack Michaels is descriptive and succinct, is never at a loss for words, and gets me clenching the steering wheel like every game is SCFG7.
I'm not here to bash Phillips; he's a part of the Oilers as much as anyone. I just think Michaels has been terrific.
When I saw him at the Saddledome when the Oilers were in Calgary in October I spoke with him (orange wig and all), and he seemed like a genuinely nice guy who was really enjoying his job. That doesn't hurt either.
If they had a Burke-cam during the Leaf games you could watch him go from smug and insufferable to exasperated and profane.
ReplyDeletePBoy, you just made my night.
@Ribs: True dat, I'm a stats junkie all the way. More is better, as lon gas there's an actual point to what's being tracked. The ones under discussion here have passed that smell test with me long since.
ReplyDeleteGreat stuff anon. Never really understood the "only one hypothesis test allowed" thing. I had always expected the real tonic to hypothesis fishing was simply replication and data expansion. Always seemed like a more sensible use of hard-won data to me. I mean, you can't eliminate Type I errors, so insisting on single tests dooms so many data sets to the waste bin when interesting correlations published (with suitable caveats about the amount of fishing used to find the interesting correlations) might lead to further study and a wealth of data useful in meta analysis.
ReplyDeleteBut your post helps me at least sympathize with the practice.
Unfortunately I just ran a test to see if there is a relationship between published attendance figures and the number of goals scored in the third minute of the second period. Damn, I burned the whole NHL historical dataset and will have to wait another year to gather enough data to determine if Cogliano's relative Corsi depends upon his linemates! At least I'm living clean.... ;)
interesting correlations published (with suitable caveats about the amount of fishing used to find the interesting correlations) might lead to further study and a wealth of data useful in meta analysis.
ReplyDeleteExactimundo. No stat is an island.
Bruce,
ReplyDeleteDidn't realize that David was counting his metric for every scoring chance, and not just goals.
That adds a lot more units to the sample, which is always a good thing and I'm a fan of scoring chances being a key metric.
Anonymous,
ReplyDeleteyou lost me somewhere around hockey stick and Mann.
Though I did get the gist of the first post.
Bruce thanks for that explanation (and Mr. Staples for your clarifications too).
ReplyDeleteAs I stated before, a lot of good data is not a bad thing when we are looking at the reasons things are happening in our game of choice.
David Staples: Didn't know that and thanks for pointing it out. Do you also account for toughness of minutes/quality of linemates?
ReplyDeleteDavid - I like your stat more than a decent amount of others, mainly because there's a lot of explanation behind it, and the stat attempts to account for what and why in a decent amount of ways. A meaningless shot from the blue line that was effectively a giveaway is a non-event.
ReplyDeleteI still don't understand how other stats sort of "quality of opposition" and such with much discernable value, being that the people that this opposition plays against contribute to that quality.
@Woodguy: Yeah, every game I post David's counts of Individual Scoring Chances as part of our detailed game reviews/player grades at Cult of Hockey. I guess we're off the beaten path for some in the 'sphere (sobs uncontrollably), but I can assure you that we both take it pretty seriously and put a lot of effort into those.
ReplyDeleteI agree with you (and Dennis, and David) that scoring chances are a key metric, and stand on my own island that conversion rates of scoring chances (equivalent of Sh%) is also significantly important and not just random distribution, but that's a future project.
The best part from my perspective, and what serves as the control, is that I use both Dennis's "team scoring chances" (my term) and David's "individual scoring chances", both of which tally up all the chances but assign responsibility differently - Dennis gives a + or - to everybody on the ice, David just to those Oilers he adjudges to be directly involved in the play in a positive or negative way depending on which end of the ice is involved. So the individual numbers aren't necessarily very close to each other, but what is very reassuring is that the team totals very consistently come in real close, in ratio and in raw number. So that strongly suggests that both of them are seeing the same plays as scoring chances, even as they then apply their different methods for assigning responsibility for same. Which raises my level of confidence in both.
As for QualComp / QualTeam, they do not get applied at the data collection level. Dennis doesn't either, nor do the shot and Corsi counters, etc. That is a filter that needs to be applied as a next-order analysis. In an individual game it can be done using Vic's head-to-head stuff and knowledge of who's who on the other team, and over time Gabe's QualComp calculations can be folded in. (We are just getting to the point in the season where those have a sufficiently large sample to be starting to become meaningful in my experience, and they certainly get better as the season wears on.)
This comment has been removed by the author.
ReplyDelete@LT: re Dave Keon being "born to intercept", one thing that he was absoluely exceptional at was how he would angle off the puck carrier in such a way that both his stick blade and both skates would be virtually perpendicular to the expected path of the puck. He'd always seem to have one skate just in front of the other and the quick stick in front of both and just take away acres of target area. Some modern players do this to a degree, but Keon was the best I've ever seen at it.
ReplyDelete(Of course I too was a committed Leaf fan back in the day, and "Little Davey Keon" could do no wrong. Nowadays I just laugh when the Leafs get beat 5-0!)
Counting the errors on all scoring chances helps the sample size problem.
ReplyDeleteI suggest that data has some value and the errors-leading-to-goals has virtually no data.
It doesn't solve the subjectivity problem. Scoring chance data has a tolerable amount of subjectivity. (some counters disagree about some chances) What counts as an error on a given chance will be much more subjective. (I've seen 100 debates this year alone about whether Gilbert made an error or whether he was forced into a spot by a bad play from his teammates.) thus, errors on chances will have a compounded subjectivity.
Nor does it tell us anything about other teams. (we do have multiple teams' data for scoring chances and it fits with what we know already.) The best way we have of checking the stats, e.g. Ovechkin, is to see if they show the players we know to be the best to be better than say Brule. To make error data work you need more of it.
Finally, as LT says, if we already have the chance data, what do the errors on chances add, especially given how subjective errors can be. The error is supposed to tell us which player is responsible for a chance against. But why bother. If a particular player is always on for more chances against than his teamates, no matter who he plays with and even though he's played with a lot of players, shouldn't we conclude that he is causaaly responsible for bringing about more chances against than his teammates?
Doesn't the chance data -if we have enough of it- already tell us who is committing the most errors? Isn't that more objective than Staples' judgments about who caused the errors?
---
and I disagree, amiably, with Bruce about the analog of Sh%, but it doesn't matter. So I'll give in.
Sorry for the typos.
If a particular player is always on for more chances against than his teamates, no matter who he plays with and even though he's played with a lot of players, shouldn't we conclude that he is causaaly responsible for bringing about more chances against than his teammates?
ReplyDeleteIf a particular player is always on for more goals against than his teamates, no matter who he plays with and even though he's played with a lot of players, shouldn't we conclude that he is causally responsible for bringing about more goals against than his teammates?
People have long argued stridently about plus/minus not being fair to the players, and scoring chance ON plus/minus will have some of the same "false positives" as goals ON plus/minus.
If Ales Hemsky and Ladi Smid play on a 5-man unit for a game and the Oilers generate 8 chances for while they're on the ice, do you really think that Smid is equally responsible for those 8 chances as Hemsky? Me neither. That's what individual scoring chances attempt to measure, both ways. Yes it's subjective, but do you really think an experienced observer analysing the play is going to get it wrong more often than he gets it right? The data ain't never going to be perfect, but the analysis of who is really involved adds a level of sophistication to the data.
Hey Bruce,
ReplyDeleteAlways fun to chat with you.
The problem with traditional +/- is primarily sample size and secondarily a lack of context, e.g. qual comp., qual team, and zonestart.
Chances against (and shots) have a large sample size, solving the first problem, and they are easily correlated with qual comp. and qual team. So the analogy you're trying to draw doesn't hold.
For example, if at the end of the season Smid has played tougher opponents, with worse teammates and the same zonestart than X, and X has been on for more chances against, it's inductively valid to conclude that X caused more chances against than Smid. (Inductive validity never yields certainty, of course, just a high degree of confidence.)
Second, let me put my point about subjectivity another way. We've followed Dennis in counting chances and it's turned out there's a high degree of intersubjective agreement about who had how many chances.
Is that true with Staples-errors? Until there at least a few games where the errors are counted by multiple people and there's a high degree of agreement, Staples-errors are pie in the sky.
And even if they're more objective than I think, as I said in my first point, they'll be less usefull than a large sample of chances against correlated with qual team, qual comp. or in a WOWY analysis.