The Who by Numbers

In a year when there is little chance of a second season, it becomes important to keep your eye on positive trends. When it comes to the Edmonton Oilers, one of the major positives this season is the quality of EV scoring among the kids.

The All-Star break is a good time to review the first half (actually more like 5/8′s) of the season and a look at EV scoring totals per 60 minutes reveals that some of the youngsters are holding their own in this very important category.

Here’s the list (courtesy Desjardins):

  1. Hemsky 2.79
  2. Horcoff 2.64
  3. Nilsson 2.35
  4. Gagner 1.93
  5. Cogliano 1.88
  6. Brodziak 1.59
  7. Penner 1.53
  8. Sanderson 1.42
  9. Moreau 1.35
  10. Torres 1.22
  11. Reasoner 1.18
  12. Thoresen 1.04
  13. Pisani .98
  14. Stoll .84
  15. Stortini .70

Hemsky and Horcoff have expected numbers, but that Nilsson total is terrific based on his slower start and inconsistency. It’s a very positive sign. Gagner’s number is also impressive considering his age and Cogliano begins a nice group of three role players (Cogs, Penner, Brodziak) who appear to be earning their keep. After that we get the guys who are done (Sanderson), have been hurt (Moreau, Pisani, Stoll, Torres), the enforcer (Stortini) and our utility infielders (Reasoner and Thoresen).

Again using Desjardins, let’s look at which players are bleeding goals at EVs. We should expect the kids to struggle in this category and as well we need to give a break to the forwards who are doing the “heavy lifting” against the other team’s best players. This category is Goals against per 60 minutes at even strength:

  1. Reasoner 3.44
  2. Brodziak 3.43
  3. Cogliano 3.29
  4. Gagner 3.22
  5. Sanderson 3.04
  6. Stoll 3.04
  7. Hemsky 2.88
  8. Thoresen 2.76
  9. Horcoff 2.56
  10. Pisani 2.35
  11. Nilsson 2.35
  12. Stortini 2.32
  13. Torres 2.31
  14. Penner 2.3
  15. Moreau 2.02

We have to be careful with these, understanding that injuries and callups mean that some of these guys may not have enough at-bats for us to consider these real numbers. Having said that, it’s about what we might expect, with strong veterans like Moreau, Horcoff and Pisani well down the list and the Gagner’s of the world right at the top. We can see clearly here why the coach likes Stortini so much and begin to see the problems MacT has in line shuffling just by the number of “plus 3′s” on this list. MacT has two full forward lines bleeding out shortly after the anthem. For comparison, the Leafs are close to offing themselves with three regular forwards (Tlusty, Bell and Steen) below the Mendoza line. That’s one full line (and as much as 15 minutes) MORE THAN TORONTO per game the coach has to juggle.

Finally we come to toughness of opposition. I’m a little hesitant to get into this discussion because the information doesn’t really reflect what we’re seeing, or at least I’m not able to explain the data in a rational way. It looks like the coach has decided to “share the wealth” in terms of matching up against the other guy’s best players so we have a lot of forwards bunched closely together.

Here, let’s look at Edmonton and then Calgary for some perspective.

  1. Thoresen .10
  2. Reasoner .08
  3. Stoll .06
  4. Hemsky .05
  5. Horcoff .04
  6. Torres .03
  7. Brodziak .02
  8. Stortini .02
  9. Pisani EVEN
  10. Penner EVEN
  11. Nilsson -.01
  12. Gagner -.01
  13. Sanderson -.02
  14. Cogliano -.03
  15. Moreau -.04

Okay, bear with me here. First of all the Thoresen numbers are over fewer games so we need to pay less heed to it. Next we have a large group between .08 and Even (there are 9 forwards, or three lines in that range) which “bunches up” the tough minutes, and in fact there have been times recently when players not among that 9 (Gagner as an example) seem to have been matched up against the other team’s top players. So what I’m saying is the Oilers really are sharing those tough minutes or possibly MacT isn’t as concerned about line matching in the “season of learning” we are witnessing.

Using Calgary’s forwards as an example, their top 3 forwards (Tanguay, .13; Conroy, .09; Iginla, .05) are well clear of their 10-12F’s (Primeau, -.10; Goddard, -.17; Nilsson, -.18) and they have 2 lines above EVEN and the rest of their forwards below sea level.

What does this all tell us?

  1. MacT’s watching the games, just like you and me.
  2. There’s probably not much he can do most nights.
  3. Horcoff and Hemsky are underpaid.
  4. Gagner giveth and he taketh away and he is not alone.
  5. This team will be better next year.
  6. If Craig MacTavish had as much input into the roster as we were told he did in the summer, I imagine his input will be a little different this summer.
  7. You cannot win if you don’t have enough established NHL players. You just can’t.
  8. Developing too many kids at one time is as bad an idea as conventional wisdom suggested, but it’s always good for NHL teams to have a recent example.
  9. The NHL now has a recent example.

written by

The author didn‘t add any Information to his profile yet.
Related Posts

77 Responses to "The Who by Numbers"

  1. dstaples says:

    Interesting numbers, but I’m starting to question, more and more, how useful a statistic plus/minus is.

    Yes, it tells us something, but what exactly?

    On what percentage of goals against where a player is awarded a minus does he actually make an error that contributes to the goal against?

    Are some players on the ice for a higher percentage of goals against where they make no mistake than other players are? For instance, Gagner has a plus/minus about the same as Reasoner, but I’d suggest he is far more valuable at even strength than Reasoner, as he is not much worse than Reasoner on defence, but he helps create far more goals than Reasoner.

    I’m not saying that plus/minus has no use, but until we know the answers to such questions, I’d suggest we temper player evaluations based on plus/minus.

    For instance, I notice that young Mr. Doughty has a plus/minus of minus 11 for his junior team, but he is still touted as the second overall pick, way higher than players with much better plus/minus rankings.

  2. mc79hockey says:

    I’m interested by Penner. Staples and the HF jackals have, in my mind, put forward a pretty low threshold for him to pass for Lowe to win the deal. His ES numbers are significantly worse than last year. He’s seen an ever so slight uptick on the PP. He’s played tougher minutes this year, I would think but I’m not so sure that he’s met the standard that makes the money or the picks worthwhile AND, more importantly, I’m not so sure that the much ballyhooed unstoppable rise, unfettered by age and historic development patterns, that was foretold by the people who liked this move are taking place.

    In short, I’m not yet convinced that this was anything but the kind of move that loser teams make – I don’t think that he’s yet good enough to be a front line player on a top end team and, for the money’s he’s making, he has to be if Lowe’s “plan” is going to pan out.

  3. godot10 says:

    There isn’t a Penner-like player in the top ten in the draft. And Penner is in the present, and not three years from now.

    This is a defenseman draft high, and the Oilers have Chorney, Petry, Wild, and Peckham in the pipeline already, and Smid and a whole bunch of other young D in the lineup already.

    The odds of Anaheim getting 1st overall and Stamkos is pretty low with LA, and the weaker Eastern Conference, with Toronto likely becoming a major seller.

  4. Dennis says:

    Not a surprise that Sanderson’s done and it probably shouldn’t be a surprise that Reasoner’s on the same boat.

    Ty, I’m looking at Penner’s Desjardins numbers and where have his stats declined?

    GD: if you have a GM that can make a deal, you always draft BPA and go from there. So that’s not an excuse for throwing away a pick just because it’s a D heavy lottery and you happen to have a lot of those already in your system. That being said, I liked the Penner move and I still do. He’s big and skilled and gives us something we’ve never been able to develop ourselves. I know that sounds old-timey and doesn’t jive with my usual views on players but when I see Penner holding off two guys at a guy and setting up plays, I’m not worried about how it will turn out.

  5. Lowetide says:

    David Staples: I understand your point but the “error” stat you’ve done a great job of promoting has its flaws too. For instance, WHO decides on the errors? Is this something that we can absolutely be fair about, or is it a matter of discretion? I think you’ll find most stats are somewhat flawed and that if you dig a little deeper you find the answer. I once read an article suggesting Johnny Bench wasn’t a good catcher (which is just about the dumbest thing ever put to print) because he wasn’t throwing out runners anymore.

    When the facts were that by about 1972 nobody ran on him anymore because they were tired of giving up outs.

    Plus minus is confusing because of context. We need to ask ‘who is this fellow playing against and how good is his team?’

    As an example, Lowell MacDonald once went +37 on a team that scored 8 fewer goals than they allowed. He did it in a season in which he scored only 6 PP goals, was on the ice for 16 more and was on the PK for 12 goals against. THAT was a season. My guess is that he played on a ridiculous top line (probably Apps (+25) and McDonough (+20), this is 72-73) and they beat the living daylights out of soft opposition while the second line (the Jean Pronovost line) did the heavy lifting.

    But I have no proof, not like today. So the +/- stat IS a strong one BECAUSE it takes the human item out of it and NOW we have context. We know if the batter is facing Tom Seaver and we also know if he’s facing Gary Gentry.

    What I’ve posted here clearly shows that while Gagner is a pretty good offensive player at age 18 in a league that overwhelms him, he is in a league that overwhelms him. As we go farther along with him we’re going to have to ask the Comrie question (can he be an ‘outscorer’) and the early results are that he has some work to do in that area. Having said that, pretty much all 18 year olds get owned in this league that’s why there aren’t many.

    How did Crosby do? I don’t remember.

    As for your ‘what if he’s on the ice when a goal is scored but it isn’t his fault?’ question my answer would be that it evens out over a season.

    Kind of like range factor.

    Tyler: Yeah, agreed. I see what people see in Penner but we paid Steve Shutt money on Yvon Lambert.

    godot10: People never get this and it kills me. You and I are the management team of the Oilers and we’re talking about signing a guy to an offer sheet. It costs 4M and 3 picks and I want to and you say “go ahead but it better work and work well.”

    But you’re against it.

    Now we sign the guy and he’s delivering exactly what Penner’s doing and I come to you and say:

    1. Well he’s certainly performing well for a 4M free agent.
    2. We’re sure not going to get a Dustin Penner in this draft

    do you buy it?

    WHY? WHY? WHY? You could have spent the money on another free agent OR kept Ryan Smyth PLUS you’d have the picks and so the hell what we don’t get a big LW we get a mondo defender who can have wingers picking Sherwood toothpicks out of their gums for the next 15 years!!!

    It’s about assets. We gave up 4M a year for a bunch of years plus a chance at an impact scorer/D in exchange for a guy who may be able to help us win a Stanley when we get enough impact scorer/D’s who can make a difference.

    I like Penner, believe me. I’m hoping like hell he can deliver at a level that makes his contract worth it.

    However, arguing there isn’t a Dustin Penner type in the top of the draft is like arguing there weren’t any UFA’s out there like Penner who would sign with us for 4M over 5 years?

    All true, but what does that have to do with anything we’re talking about?

    We gave up BOTH!!!!!!! 4M times 5 and a chance at an impact player.

    That’s the equation. We get something and we subtracted something and then we subtracted another thing.

    Two things. Two.

  6. Black Dog says:

    I like Penner – Lowe rolled the dice on the kid and I think he’s going to pay off.

    Of course the fact is that Lowe could have had Penner and Smyth and he could not have signed Souray if we want to throw another scenario out there. If you’re going to spend 25M plus on a guy over 30 I’d rather it be on a guy who is a proven player (maybe not “elite” but pretty damn close) then a guy who has had one good year on the PP. Period.

    Of course now we’ll get all of the “they were right to let Smyth go” arguments, which ignore the simple fact that Lowe himself admitted last summer that he blew it with Smyth.

    As for the argument that the Oilers are loaded with defencemen so who cares if they don’t have a top pick – well, they are loaded with young defencemen and some of them look like they might be pretty good and if Scott Niedermeyer or Scott Stevens were available in the draft I still think that it probably wouldn’t be a bad idea to, you know, USE THAT PICK!

    By that logic if the Oilers have a high pick next year (very likely by the way) and the best available players are all Dmen then they should trade the pick or take a flyer on a goalie ranked in the second round?

  7. dstaples says:

    OK, Tyler, what would your threshold be for Lowe to pass this test, say if the judging were to be done on July 1, 2008?

    Again, my suggested test is that Penner equals his offensive output from last season, that he’s a plus player in plus/minus, and that the Oilers draft no lower than 7th pick overall (I believe).

    This test is no good for settling question if people don’t agree it’s a sound test.

    So let’s come up with a sound test, one that we can agree on, to settle whether or not Lowe made the right call her.

  8. rickibear says:

    Lt: So what you are saying is we should let Pitkanen go for assets and try and sign Redden.

  9. Lowetide says:

    rikibear: Absolutely fine by me. The Oilers started the new era by trading for Pronger instead of signing Niedermayer (who was apparently at least interested) and they had the depth to do it but think about how good this team might have been with Pronger AND Brewer.

    So you trade Pitkanen to Los Angeles for one of their good young kids and then sign someone who’ll come here. I don’t know if Redden is still good, seems to be taking a spiral downward over the last 2 seasons.

    But the idea I’m fine with.

    David Staples: I don’t think the pick matters now, certainly not to Penner (it’s still on Lowe’s resume, though). My own opinion is that Penner has already shown enough for us to welcome him into the fold all the while knowing that either Kevin Lowe really wanted him so badly he overpaid with money and picks or that the Oilers are such a bad UFA option that no one would take our money.

    I don’t think we can win by sending our picks away though. Their value isn’t what it was in 1990 but there are impact players at the top this season and the Oilers could have gotten one.

  10. Black Dog says:

    What he should be doing is finding some dummy to dump Souray off on.

    Matheson talks about trading Pitkanen but he also makes the salient point that he played seventeen of the last twenty two minutes in the Florida game or something absurd like that. He gets hurt a little too much for my liking but if they let him go because they don’t want to pay him that would make two pretty good players gone over money in a year, while an inferior guy got a huge paycheque.

  11. dstaples says:

    Who assigns “errors” and how are the assigned? These are the big questions, for sure, in trying to come up with an enhanced plus/minus stat, where only players who are most responsible for goals against are awarded a “minus.”

    If we can’t come up with objective standards for assigning errors, ones that will be replicated by a number of statisticians, then it’s not a very useful stat. It would be too subjective.

    That said, I’ve been working with the so-called “jackals” at Hockey’s Future, and have been encouraged somewhat on a few fronts.

    First, there seems to be fairly widespread agreement on what a goal-contributing error is, so objective standards could be written out to guide any stats person assigning errors.

    Second, on games we’ve looked at collectively as a group, there’s been much agreement on who should actually get an error (though certainly not completely agreement).

    Anyway, the more I have watched games and assigned errors, the more I have come to think that this likely can work and be a useful metric. Of course, it doesn’t matter what I think. If others don’t accept the “error” as a useful stat, it has no validity.

    BTW, Gagner has made fewer “errors” than Reasoner. Horc is by far the best Oilers forward when it comes to play error-free hockey on defence, with Penner second.

  12. Lowetide says:

    David Staples: Interesting. My next question then would be how much of Gagner’s NOT being in the play has to do with his inexperience? Not trying to shoot down your project but kids are often so badly out of the play they aren’t even in the vicinity.

    Does the error account for this kind of thing? Are there ‘areas of responsibility’ assigned?

  13. dstaples says:

    Lowetide: Errors are definitely assigned to those players who are out of position.

    In fact, one of the most commonly assigned “errors” is for this kind of mistake, as bad positional play factors into the majority of goals against the team.

    Gagner’s most recent “error” came this way. The puck came around the boards, pass from Gilbert, and Gagner was out of position, not there to get it, so he got the error.

    The most common of all errors is for players to be drawn to the puck carrier like wasps to a can of Coke, thus leaving other players open for goals. This is a common problem for Smid and a few of the young forwards.

    Note: Gilbert, who had been strong defensively most of the year, is now on a terrrible streak with five “errors” in the past five games.

    Here’s a link to the most recent round-up of who has committed errors on the Oilers . . .
    http://communities.canada.com/edmontonjournal/blogs/hockey/archive/2008/01/22/df.aspx

  14. Bruce says:

    Interesting discussion. I tend to lump the Smyth and Penner deals together; we traded a big salary first line LW for “three assets”, and then effectively traded three assets for a big salary first line LW. Penner is younger and considerably cheaper ($4.2 MM compared to the $5.5 that apparently Smyth would have accepted). Ryan is arguably the better player today, although maybe not $1.3 MM better, but three or five years from now my money is on Dustin.

    Now for the assets. Oilers got three first-rounders, two of them already in hand (Nilsson and O’Marra) and one in the bush (Plante). Of those it already appears Nilsson will be a useful player; the jury is very much out on the other two. Whereas Oilers gave up a #1, #2 and #3 pick. Sure there are nightmare scenarios where Oilers suck down the stretch and Anaheim wins the lottery and gets a Stamkos or a Schenn, but that’s hardly a given; Oilers currently sit in 23rd and seem to be moving up. The second and third rounders maybe turn out but are less of a gamble. The other thing to consider is not who Anaheim picks but who Oilers would have picked, which based on their track record could be the next Niinimaki, Rita, or Pouliot.

    It’s early days but considering the two transactions together I think Oilers did alright.

  15. Dennis says:

    I can’t believe that Lain and David Staples-Lowe would break up just before Valentine’s Day!!!

    What the world needs now…

    OK, to be serious for a moment, the best point I see here is Lain talking about how it evens out over a season if poor old Gagner gets nailed with a minus while being on the ice for a goal he didn’t have any fault in.

  16. Dennis says:

    Bruce: you know I respect your opinions and your contributions. So, while I supported the Penner signing and was the guy who coined the Bustamakki nickname, which was my comment on Oilers drafting at the time, the Oilers have done a better job drafting so I can’t accept, “well they would’ve fucked the pick if they had it anyway” as an argument.

    The Oilers hauled Cogs out of the 25th spot in ’05 and that looks like it’s gonna work out fine. There’s no question we’ve improved in that regard.

    All that being said, I think Penner’s as good as anything we’d haul out of the draft if we say we finished 10th in our Conf; which would’ve been a good bet if we hadn’t been hit with so many injuries.

  17. dstaples says:

    And I can’t believe Dennis is actually backing Lowe on the Penner signing.

    I battle with LT and side with Dennis, all in the same day.

    Shocking, my friends. Shocking.

  18. dstaples says:

    Can we be sure that plus/minus evens out over a season? How do we know this?

    I’m not at all convinced it’s true. Just as likely it doesn’t even out, and some poor guy has a minus 10, and he’s on the ice for 15 goals against he had nothing to do with, and only 6 goals for that he had nothing to do with.

  19. PunjabiOil says:

    So you trade Pitkanen to Los Angeles for one of their good young kids and then sign someone who’ll come here. I don’t know if Redden is still good, seems to be taking a spiral downward over the last 2 seasons.

    But the idea I’m fine with.

    The irony is that Kevin Lowe traded Pronger for good young assets, and then went on to throw an offer at Chara.

    He then traded Smyth for some good young assets, and then threw offers at the likes of Gomez and Nylander.

    Higher profile Unrestricted Free agents aren’t exactly “free” – they end up costing a premium to the point which you question whether the player will truly deliver value of the cap hit.

    I’d love Frolov on the Oilers, and I’d definitely think about a package of 2 1sts, 2nd, 3rd if Pitkanen gets an offersheet – but the problem becomes that your defence just took a step back. Pitkanen isn’t flawless, but he’s got lots of room to grow.

    This is the list of most notable UFA defencemen this summer:
    ____________
    DEFENSEMEN THAT WILL GARNER THE MOST ATTENTION:
    Aaron Ward, Brian Campbell, Bret Hedican, Mike Commodore, Adam Foote, John-Michael Liles, Mattias Norstrom, Brad Stuart, Rob Blake, Nick Schultz, Paul Mara, Marek Malik, Michal Rozsival, Jason Smith, Mark Eaton, Dan Boyle, Jan Hejda
    _______________

    Campbell, Foote, Smith, Schultz, Boyle, Hejda are the only ones interest me.

    I’m not sure how many of them can replace Pitkanen today, but possess the upside in say 3-5 years from now.

    Nor is there any guarantee you will be able to get them. Or if you do, is the UFA premium worth it?

    It’s a tough decision.

  20. speeds says:

    Bruce:

    I think LT’s point is that the transactions are not related, even if one chooses to look at them that way.

    EDM probably could have instead signed Nagy (again, who knows if he’d have come here, but let’s pretend he would have if offered enough money) for Penner’s money and kept the picks.

  21. PunjabiOil says:

    The other thing to consider is not who Anaheim picks but who Oilers would have picked, which based on their track record could be the next Niinimaki, Rita, or Pouliot.

    I don’t like this argument

  22. PunjabiOil says:

    To add to my post:

    Of course you could always trade Stoll/Torres for a defenceman.

    Frolov plays on the Left wing.

    Frolov-Horcoff-Hemsky.

    Frolov was +12 in 2002-2003. +8 in 2003-2004. +17 in 2005-2006. A beauty of a player.

    Then you have Dustin Penner on the 2nd line – making Torres ripe for trade for a defenceman. You’d probably have to add a guy like Schremp in their, but the prospect of this deal is intriguing.

  23. PunjabiOil says:

    I can’t believe that Lain and David Staples-Lowe would break up just before Valentine’s Day!!!

    I admit I laughed.

    Out loud.

  24. dstaples says:

    I admit I laughed out loud at the “Staples-Lowe” comment. Good one Dennis, you evil bugger you :)

  25. Lowetide says:

    I went out and cleaned off my driveway. I am now unable to laugh. Possibly later.

  26. PunjabiOil says:

    you evil bugger

    Evil bugger? LOL.

    We’re all adults here in the ‘sphere.’ Nobody will be on your case if you go all Joe Pesci on Dennis.

  27. Black Dog says:

    Sore back to go with your broken heart, LT?

  28. Lowetide says:

    We shared a two hour conversation over coffee that included Bill James, the great Oiler teams and how we would run the Oilers.

    AND he bought the coffee!!!

    I need a minute here.

  29. dstaples says:

    My indoctrination at the newspaper has rendered me incapable of writing a curse word in print, even though I’m perfectly at ease quoting others who curse, and also curse plenty myself in person.

    And, frankly, I don’t see Dennis as anything but an evil bugger :). He’s got a heart as big as the Oiler symbol, which he hides beneath his gruff Lowe-lynching exterior.

    Of course, when it comes to swearing, I do have one exception to my rule. So Mr PJO, Mr. Devious Trickster, I do not hesitate at all to label you as a kuthapreet phudi (inside joke)

  30. PunjabiOil says:

    I’m interested by Penner. Staples and the HF jackals have, in my mind, put forward a pretty low threshold for him to pass for Lowe to win the deal. His ES numbers are significantly worse than last year. He’s seen an ever so slight uptick on the PP. He’s played tougher minutes this yea

    I’d be interested in Penner’s numbers if you filter out the first 13 games of the season (5 points). I don’t think it’s a coincidence that his numbers improved after he got in shape. He’s also created lots of room for Hemsky and Horcoff, and has done a good job screening goaltenders. Of course this wouldn’t show up in the EV P/60 numbers, but I’d argue he’s helping the Oilers win hockey games.

  31. dstaples says:

    And you, Lowetide, you even linked once to my blog post on the error statistic. . .

    And now it has come to this . . .

    All out warfare!

    The gloves are off, my friend.

    P.S. Since I’ve been posting regularly at HF boards about the “error,” I am now well-used to all manner of attacks on this project, both level-headed and ridiculous. Hey, if this “error” stat works, it works. If not, oh well, I can always keeping trying to decipher all the complex stats and flow charts you guys feast upon.

  32. Lowetide says:

    David Staples: You’ve already posted my photo there’s nothing more you can do to me. :-)

  33. dstaples says:

    No, there is something more . . . :)

    I can post it again.

    And again.

    And again . . .

    BOO HA! HA! HA! HA!

  34. Bruce says:

    I wrote:

    The other thing to consider is not who Anaheim picks but who Oilers would have picked, which based on their track record could be the next Niinimaki, Rita, or Pouliot.

    Key word: “could”. I agree Dennis, Oilers track record is better of late, and that comment probably came across as overly negative. This Oiler fan has learned to temper all the hoopla about draft picks with years of bitter experience. My primary point was when all the dust settles and these deals are judged, what Anaheim gets will be different than what Edmonton gave up, just cuz we would have made different picks. The Lowe-lynchers (geez, David, I would have thought that of all words would be verboten among you media types just now) will be able to go through the ’08 Entry Draft years after the fact and pick the best guys who were availabe in each of those rounds and say that Penner wasn’t worth that price. By the same token one could judge the Smyth deal by saying we couldawouldashoulda had Cherepanov with the Plante pick. But at least in that case we can judge our side of the trade based on the guy we actually picked. In the Penner compensation we’ll never know.

    I think LT’s point is that the transactions are not related, even if one chooses to look at them that way.

    Well I think K-Lowe did look at them that way, Speeds. As a result of the Smyth deal he had depth in the system therefore assets to move, cap room to spend, and a big gaping hole on the first line to fill. I seem to recall a couple of quotes from him at the time of the offer sheet to that effect. And if that was part of his thought processes the deals were related.

  35. Andy Grabia says:

    Lowetide: Errors are definitely assigned to those players who are out of position.

    Excuse me for being blunt, but how do you know he’s out of position? Do you know where every single play is supposed to be in every single situation on the ice? Because if not, I’ll just take the +/- stat. I’m well aware of its problems, but the error idea is way more subjective, and therefore way worse.

    In fact, one of the most commonly assigned “errors” is for this kind of mistake, as bad positional play factors into the majority of goals against the team.

    This makes it sound like some independent, unbiased third party, or a robot, is assigning these things. You are doing it, David. It’s guessing.

    Can we be sure that plus/minus evens out over a season? How do we know this? I’m not at all convinced it’s true.

    You’ve run the numbers, then?

  36. speeds says:

    Because if not, I’ll just take the +/- stat. I’m well aware of its problems, but the error idea is way more subjective, and therefore way worse.

    Interested to hear this expanded upon, if you don’t mind.

  37. speeds says:

    bruce:

    It could be that was Lowe’s attitude, who knows, but if so it doesn’t seem all that smart to me.

    Because you’re rich in assets and cap room you should squander them?

  38. Steve says:

    Because if not, I’ll just take the +/- stat. I’m well aware of its problems, but the error idea is way more subjective, and therefore way worse.

    Interested to hear this expanded upon, if you don’t mind.

    Well, I’m not Andy, but I share his opinion on this, so I’ll offer my expansion on it:

    1. The league has yet to find anything approaching a consistent standard in counting hits.
    2. An error is far more subjective than a hit (how do you know exactly where a player was supposed to be? Do you automatically penalize an unsuccessful pinch by a defenseman, even if said pinch had a reasonable chance of leading to a goal?), so it’s a safe bet that this wouldn’t be counted with any degree of consistency.
    3. +/- is at least objective, meaning that once you have the numbers you can stop arguing about them and start trying to find explanations for them.
    4. There exist measures like quality of opposition that greatly assist in this process of finding explanations.
    5. A player getting fifteen minutes a game for eighty games is not going to have his +/- much affected by the vagaries of chance. It’s the law of averages. Of course, it could be affected extensively by quality of opposition (for which we have a measure). It could also be affected extensively by linemates (in the event that he’s had the same ones all year), which is where reading the stat intelligently comes in.

    +/- is always going to be of more use to a thinking hockey fan than “errors” will be.

  39. Bruce says:

    Interesting numbers, but I’m starting to question, more and more, how useful a statistic plus/minus is. Yes, it tells us something, but what exactly?

    Good question, David. I have always put a lot of stock in +/- as opposed to pure counting numbers, but like any stat it needs to be put into context. Quality of linemates, teammates, opposition, and the relative strengths of the goaltenders all have a bearing.

    e.g. Nicklas Lidstrom leads the league at the break with an astonishing +40, but he plays a large chunk of his time with two-way superstars Henrik Zetterberg and Pavel Datsyuk, and when he does make a mistake he’s got Dominik Hasek or Chris Osgood behind him. Whereas Brad Richards — ironically, the only player in hockey with a larger cap hit than Lidstrom — has compiled a league-worst -25 playing with Jan Hlavac and Jason Ward and Gumper the Goalie. Is Lidstrom 65 goals better than Richards? Probably not. Is there enough information there to suggest that Brad shouldn’t make more money than Nick? Indubitably.

    Another consideration is special teams, where only SHG count in the +/- column. A player who plays much more on one special team than the other will find his +/- skewed somewhat as a result. Many of those same players will be further skewed by empty netters; e.g. John Madden or Chris Phillips will be on the ice for a lot more SH/ENGF than SH/ENGA, whereas Marc Savard and Daniel Briere will suffer the reverse. (This does NOT explain the gap between Lidstrom and Richards as both play lots in all situations.)

    Is plus-minus perfect? Of course not. I think of Marc Pouliot’s dumb-ass line change which wound up not as a minus for him but for Dustin Penner who was on the ice but still in the bench area when Filpulla scored. I generally agree with LT that these sorts of things balance out over the course of the season, just like cheap assists do or bad reffing calls or other breaks of the game. You can always find a single example of any stat that is egregiously wrong but that will disappear into the noise over time.

    I do like your Errors stat, David, it’s imperfect but then again, so are all the others. Potentially it’s another column of information that can be judged in context of the others. In baseball Errors are probably less important than Range Factor, but that doesn’t make them UNimportant. The two taken togteher can tell you a lot about the player: does he reach a lot of balls, and does he handle them cleanly when he gets there? In hockey it’s not so clearcut, but I’d like to think a combo of +/- and errors will tell you more than just +/- alone. Context is still key; a great goalie can make errors and minuses disappear completely, in the manner that a vacuum cleaner at first base can nullify a lot of wild throws. The raw numbers certainly help us to judge, but in a team sport statistics are more guide than gospel.

  40. speeds says:

    I’m wondering how this might compare with scoring chances vs. shots on goal?

    Most seem to think they can give you a better indication of the game than shots, yet scoring chances are subjective while SOG are objective (at least, they are ideally, no doubt they count some shots going wide as SOG while not counting some that are actually on goal).

  41. Bruce says:

    Because you’re rich in assets and cap room you should squander them?

    Well no, Speeds, I don’t think you should squander them, I think you should use them. “Squander” is in the eye of the beholder, but it’s not a word I would use w.r.t. Dustin Penner, one of the more promising young power forwards in the game.

    Oilers are flush in prospects, er, “players in the system” and very definitely shy of NHL-quality players on the big team. Given those circumstances, which were exacerbated by the Smyth trade, putting a Dustin Penner in the line-up right away might make more sense than spending 2 or 3 or 5 years developing 1-3 more prospects that may or may not even have room to grow. A GM cannot (or should not) make each of these decisions in a vacuum, because each move changes the mix.

  42. Bruce says:

    I’m wondering how this might compare with scoring chances vs. shots on goal?

    It’s a very good comparison, Speeds. This number cruncher would like to see both stats made available. Assuming there is some reasonable definition that is applied across the board, a subjective stat is surely better than no stat at all.

    If standards are wildly different from one arena to the next (as is the case for RTSS) perhaps more can be gleaned by isolating a player’s away games and factoring out the home scorer (similar to “park effects”).

  43. Sean says:

    I’m not a huge fan of assigning errors because its way too subjective as others have said. At what point is the error assigned? Can you assign partial errors? Errors must be evaluated given context and it becomes too hard to do and agree on. Hockey is a team game and ultimately +/- works ok for me.

    As for Penner. I think its fair to package Smyth and Penner when analyzing if its a good move. I still say given our inability to attract UFA’s and surplus of picks we already had it was a good gamble.

  44. speeds says:

    I can certainly understand EDM deciding to bid on RFA’s after being shut down in the UFA bidding, I guess I still just find it strange that they decided to commit the money and the picks to Penner when they could have signed a UFA and kept the picks, but I guess that rests on one’s evaluation of Penner. But you’re right – once they shut were shut out of the UFA market, going that way is by no means crazy – in fact, I’ve encouraged it in the past.

  45. dstaples says:

    Andy wrote: “Excuse me for being blunt, but how do you know he’s out of position? Do you know where every single play is supposed to be in every single situation on the ice?”

    First off, this work on the “error” stat is new, it’s in its infancy, it may not go anywhere. But I’m reasonably sure there’s something to it, that it does have some validity, and I’m not at all convinced by critiques that hold the old plus/minus is superior.

    As for the Gagner error, when the puck comes around the boards, it’s the forward’s job to be there, so he can get it out or freeze it or whatever. Don’t you think so? I mean, who else should have been there but Gagner?

    I don’t see hockey as an overly complex game, at least not in the defensive zone.

    Now, it’s true, that as a strategy a team can be passive or aggressive, but it is always man-on-man, or, at least it is until there’s some kind of defensive breakdown, when suddenly it can be two-on-one or three-on-one.

    With the “error,” we’re looking for that moment of the defensive breakdown, which from watching and playing hockey most fans consistently identify. Think about past blog posts people have made, and how they point out that this player or that player made an error that led to a goal.

    All of us are doing this all the time anyway. I’m seeing if we can come up with a structured way of doing it, and I’m keeping track of the errors I’ve seen.

    On this play, where Gagner was assigned the error, there was a discussion on the HF Boards, for instance, and there was no disagreement at all that Gagner made the primary error.

    That said, for this to work, there has to be an agreed upon criteria of what constitutes an error. There isn’t one now. So not everyone is going to agree on when an error occurs. I’m working on a list and will publish it soon enough.

    But yes, Andy, I do think a hockey fan who has watched hundreds and hundreds of games can watch a defensive play and identify where the breakdowns occur, though they may have to watch the videotape 15 or 20 times to be sure.

    It’s true — assigning errors is a subjective judgment, it’s not mechanical (though, if you watch goals against closely, it certainly starts to feel mechanical), and there will indeed be disagreements at times on what constitutes an error. But with the regular plus/minus we know with certainty that “minus” and “plus” points are consistently assigned in error all the time, to players who had nothing to do with the play. This rating system is mechanical and objective, but problematic.

    Can we improve upon it?

  46. dstaples says:

    Steve wrote: “A player getting fifteen minutes a game for eighty games is not going to have his +/- much affected by the vagaries of chance. It’s the law of averages.”

    How we do know that the law of averages works to eliminate error from plus/minus over a season?

    I’ve been looking at it a few games, and I can certainly tell you this isn’t the case over a few games. So what is the magic number for the law of averages to take effect?

    I’ve also watched every goal against the Oilers at even strength 15-20 times in order to assign errors, and I don’t think Nilsson and Gagner are any different when it comes to defence — they are both less than average for a forward — but one (Gagner) has been on for a lot more goals against then the other, and has a far worse plus/minus, but I’d argue it’s undeserved.

    Now, you can then look at quality of teammates at even strength and quality of competition, that’s true. Maybe that will tell you Gagner and Nilsson are more even than plus/minus shows, if you simply go by plus/minus. I don’t know. But I do know the error stat is telling me that right now, as both have been assigned the same number of errors.

    Now, it could be, as Andy says, I’m just guessing on assigning errors here, and others would guess in a completely different manner, and nobody can really assign errors in a way that others would agree with, and the game is too complex to do so in any useful way, and we’re stuck with plus/minus, and its various problems.

    But I’d ask people not to make their minds up too quickly on this issue. Sure, if I go too far for your likely in promoting this stat and making claims about its usefulness, that’s worth a blast.

    But could this thing work, perhaps? I think, maybe, it can be useful and, frankly, I’ve been encouraged that some folks agree with me.

  47. Scott says:

    Hey David, do you provide a list of errors/60 for all of the Oiler players somewhere? Maybe we could check to see what kind of correlation there is between e/60 and ev-/60.

    As for the Penner deal, it seems pretty reasonable to me if a team adopts it as a consistent strategy (giving up a 1st,2nd,3rd for a top4 dman or top6 forward). By signing a player to an offer sheet we get the NHL talent now. He is signed for his prime scoring years. Not every first round pick a team makes will work. The team saves money because it doesn’t need to develop as many players. You don’t pay for the lower performing entry level seasons. As long as you don’t consistently suck and finish bottom5 in the league, I think you come out on top making an offer sheet of this ilk every season. I honestly don’t understand Detroit not doing it. Especially if they can do it to a team in their own division. Frankly, they should kill the Nashville dmen with offer sheets if they aren’t already locked up.

  48. dstaples says:

    Scott, yes, there is a link on my blog, the Cult of Hockey. I can’t figure out how to put a link here or I’d provide one. Just go to the Edmonton Journal site, go to blogjam. Or, now that I think of it, there’s a link off of Lowetide’s mainpage here to the Cult of Hockey. The post in question was made last Monday or Tuesday, and it’s called, “More proof that Horcoff is a hockey God,” as this stat has really highlighted his outstanding defensive play.

  49. Scott says:

    David, is there a way to factor all of the sh goals and 4on4 goals out of your error stat so we can measure just 5on5e/60 and 5on5-/60? I think this would probably be more helpful when trying to find a correlation. It seems like the error is a bad stat to compare against +/- because the error doesn’t factor in offensive play at all, while +/- obviously does. The difference between Nilsson and Gagner is somewhat because of the difference in their ev+.

  50. dstaples says:

    Scott asked, “David, is there a way to factor all of the sh goals and 4on4 goals out of your error stat so we can measure just 5on5e/60 and 5on5-/60? I think this would probably be more helpful when trying to find a correlation. It seems like the error is a bad stat to compare against +/- because the error doesn’t factor in offensive play at all, while +/- obviously does. The difference between Nilsson and Gagner is somewhat because of the difference in their ev+.”

    But the “error” does factor into offensive play.

    It is meant to be a shadow stat for the point. Just as we have as many as three points awarded for a goal scored, we have as many as three errors are assigned on a goal against. So when we do plus/minus with this stat, it is even strength errors vs. even strength points (with a few exceptions, as I explain later).

    With this stat, right now, Horcoff is something like +29, Hemsky +22, Gagner/Cogliano/Penner/Gagner all around plus 10-12, and Reasoner and Stoll around the even mark, terrible for a forward, so far as I can see.

    In assigning errors, I have counted four-on-four play, as that is even strength. I’ve also counted shorthanded goals against, as clear errors are made on these goals.

    I have not counted errors made while the Oilers kill penalties, as the powerplay is a man advantage, and it’s evident to me it would be unfair to count errors made here as the same as errors made at even strength.

    Some critics, though, believe it would be wise to count errors made on powerplay goals, and this is up in the air somewhat, though I still think it best not to count them.

    I also have not counted errors on even strength empty-net goals. Again, it seems to me that errors made here aren’t the same as other errors. Players are supposed to take great risks when the goalie is pulled, and any small error can easily lead to a goal against. So, for now, I haven’t included empty net goals in the error totals.

  51. Dennis says:

    LT: I’m glad you didn’t go overboard complaining about snow. Try putting up with 40 odd cm with winds over 100 KM and then get back to me. I repeat, don’t grumble about 5-10cm;)

    Staples: I don’t hate everything Lowe’s ever done, I just hate the obvious stupid stuff he’s done. Things like Pronger to Ana for magic beans and letting 94 walk over 100K. You know, things like that:)

    Because my mood’s been brightened by yet another killer ep of The Wire, here’s my crack at some of the dialogue in that two hour conversation between LT and Staples.

    LT: Oh how I wish the Oilers would just let Lowe do everything!!!

    Staples: Oh my stars, can’t you imagine the things he’d accomplish with full control!!!!

    LT: Oh laws yes, I certainly do. He’s having a tough go of it this year but he’s no idiot, he’ll figure it out.

    Staples – extending mug to tap with Lain-: TO VISH!!

    LT nods vigorously and enamel meets.

    And scene;)

  52. rickibear says:

    David: Thank heavens for people like you who seek to make NHL statistics real.

    Lt: How much do MLB score keepers get. A core of properly trained score keepers assigning: errors, Shots (NBA style), Blocks, Etc. would be great.

    Sean: The CBC assigning scoring chances is subjective but that stat usually says more about the game than shots.

    Having played contact sports at an international level. I can say the body breaks down as one ages. When it does it goes fast.

    No doughting smyths skill.The smyth decision to me was a decision of money vs the managements belief of smyths body breaking down.

    It looks like they may be correct.

    If management is balancing Pitkanen’s toughness, and ability to have longevity, in there valuing of him. I do not have a problem with moving him for assets.

    The same could be said of other players on the team. Though the guys whom trained hard for the first time this year will see no consistent benefit until the second year of advancing core strength.

  53. Lowetide says:

    Hayzoos, were you there?

    Classic stuff Dennis.

  54. Lowetide says:

    rickibear: Back before baseball died it would often be a local scribe (since Staples has been here all day I’m using the Thesaurus now) and it probably was a small amount of money.

    Since the game passed from the scene I doubt you could find someone still alive to give an accurate account although I understand they have found the exact spot where Wrigley used to be.

  55. mc79hockey says:

    The smyth decision to me was a decision of money vs the managements belief of smyths body breaking down.

    I’d have a much easier time believing that if they didn’t sign Sheldon Souray three months later.

    David – yours is a good question, worthy of a post on my site. I seem to be out of bandwidth, but I’ll get to it. Suffice it to say, I don’t think that the question will be answered one way or another this year as to whether this was a win for Lowe, although I think it was a stupid risk regardless.

  56. Scott says:

    David, the error doesn’t factor into offensive play. It just doesn’t measure anything offensive. You can compare it with points to come up with some new kind of stat, but since points aren’t measured in the same way as errors, I don’t think that’s altogether helpful. Sometimes a good screen, or a pick isn’t awarded a point, but a relatively simple second assist is. I think if you started giving anti-errors that lead directly to goals, they wouldn’t always look the same as points.

    As for wanting a situational breakdown, since we have the largest sample size for 5v5 play and the numbers from Desjardins at our disposal on a per hour basis already, it seems easy enough to get a 5v5errors/60 and compare it to 5v5-/60.

    As far as I can see, you may as well keep track of errors in every situation and note which situations they happen in. It’s easier to make an error leading to a goal 4v4 than 5v5 since there are more goals 4v4. I think it would be nice to be able to see the breakdown. Plus, that way you can start keeping track of PK errors and EN errors. Just keep them in a distinct pile.

    I think what you’re doing has a lot of merit, but I think that you really need to admit you’re measuring defensive play and not try to make it look like a +/- stat.

    Finally, is every goal awarded errors? Or are some goals error-free (the puck bouncing off two skates into the net against Florida comes to mind)?

  57. rickibear says:

    MC: We all know the Souray signing was a PR hit for EIG and the Nylander brush off. Souray even said they were not in the picture till the end.

    Scott has hit it on the head. Just like baseball. We could even end up with an adjusted Goaltender GAA and save percentage (like era reduction in baseball) based on position of shots and errors.

    LT: Do not know much baseball history and off the feild logistics. Learned all my ball knowledge from a dad who played catcher as a fifteen year old in the western canadain mens leagues in the early 50′s. He lied about his age and joined the air force when his dad would not let him go on a profesional opportunity with one of the teams.
    Watched lots of games and learned to catch standing at the wall of a shed and he fired them.

    Didn’t put a glove on again till mid 90′s when a freind at school (NAIT)said he used to pitch for alberta baseball. We went to the butterdome he knew the boys. He could fire them. Learned he turned down a full ride to LSU. Left hand pitcher with 5 pitches.(stupid) Rights belonged to Atlanta. Learned more from him as well.

  58. Bruce says:

    I too feel there’s considerable merit in the Error concept, at this early stage it’s newborn and fragile and should be nurtured to some extent. David I salute you for trying something new, and especially for extensive consultation of others in the process.

    I agree with Scott’s point that it is not exactly a mirror for offensive stats, and if it is it’ll never balance out; forwards get most of the points while D-men and goalies take more than their share of errors I would think. Also David goes to more trouble determining errors than scorekeepers do scoring plays. I also agree with Scott that a good screen may be more worthy of a point than, say, a simple pass from the perimeter back to the point, but the scoring system is cut and dried. Whereas if Staples’ method was applied to the offensive end — now there’s a thought — the guy setting the screen would probably get his due.

    Good question about the double-bounce fluke in Florida (I think it hit Moreau’s stick as well), maybe somebody screwed up in letting Jokinen shoot from the slot, but it was one of those acts-of-the-hockey-gods kinda goals.

    David, I’m also interested to know if Pouliot was assigned an error for that egregious line change against Detroit back in late October. To me that was the worst error resulting in a goal-against this entire season (magnified by the situation no doubt). Some blamed Torres, Gagner, Tarnstrom or even Roloson, so you had lots o’ choices, but I wonder if your fledgling system is sophisticated enough to catch the guy who didn’t even get a minus on the play.

    And one more for you, David, Staal’s penalty shot goal. Is that one all on Garon, or did you finger the guy who pulled him down on the original play, or maybe somebody up ice who got burned for the breakaway? (I can’t name names cuz I recorded the third period of that game and just deleted it when I heard the score.)

  59. Lowetide says:

    In the words of Larry Miller:

    But at level three, that devil is a little bit bigger….and he’s buying and tells you to that as long as you get 3 hours sleep, and a change of blood – “Cool.”

  60. Bruce says:

    Thanks David. One outstanding question: the penalty shot?

    Be happy you were able to forget the Filppula goal, I remember it every time I look at the standings. Losses are one thing, but regulation losses something else again, and last-minute regulation losses totally suck. Pouliot wrote his ticket to Springfield with that one.

  61. dstaples says:

    Ah, the penalty shot. I’ts not yet come up, as the only penalty shot goal against was while the Oilers were killing a penalty.

    But I thought about the issue then, and my leaning is there’s discretion enough to go back to the play that caused an even strength penalty shot, and award errors to the players who caused that breakdown.

    What do you think?

  62. Bruce says:

    I agree that whoever is responsible for the penalty shot should be on the hook for an error. It’s just a particularly focussed example of how the goalie can bail out the culprit by making the stop on the subsequent attempt.

    I had forgotten that the Staal penalty shot was in fact on the PP, so this time it is you who are off the hook. :)

  63. RiversQ says:

    The error doesn’t work in baseball because it unfairly singles out players that actually manage to get themselves involved in the play. The good defenders might make a few more errors because they actually get to the fucking ball.

    I think you can watch a guy like Stortini or Gagner get pinned in his own end all night and not make an egregious error. That doesn’t mean they played well.

    I have yet to see Staples mention anything that counteracts this problem.

    Lastly, apparently he’s claiming that the stat shows that Gagner has been a good player this year.

    I can barely type that.

    That pretty much clinches it for me, because that couldn’t be further from reality.

    (That kid has a future, but he’s not a decent NHLer right now and he proves it every night.)

    I’ll stick with the GF/GA and things like the Corsi number. As long as you consider the context, there’s nothing wrong with those figures.

    David, if you really want to make a difference, wind up those HF monkeys into counting scoring chances +/-. That’s what really matters here. It’s better than goals, shots, or anything else.

  64. Andy Grabia says:

    Now, it could be, as Andy says, I’m just guessing on assigning errors here,

    You are guessing. You can call it an informed guess if you want, but you are still guessing. You are also biased, as you are counting errors for a team you love and cheer for. You are correct in asserting that we as fans discuss who is at fault on any given play all the time. The difference is, none of us are trying to manipulate subjectivity into some objective metric on defensive prowess.

    But I’m reasonably sure there’s something to it, that it does have some validity, and I’m not at all convinced by critiques that hold the old plus/minus is superior.

    Assuming the validity of this, which I don’t beyond a very elementary level, what’s the stunning revelation you’ve made so far? That Shawn Horcoff is a good defensive hockey player? Not exactly shocking, is it? Many of us became convinced of that looking at other metrics besides plus/minus and your error system. So where is the value-added here? Furthermore, do you believe your error metric is better or worse than those other metrics?

    It’s subjective, based on what I would argue are objective criteria.

    Huh? Please explain this one.

  65. dstaples says:

    RIVERSQ: I know you guys are wedded to the position that Gagner is a bad player. That may be, but I’m saying is no worse a defensive player than Nilsson, and that’s what the error stats shows right now, but that’s not what regular plus/minus show.

    Is there some context with plus/minus that does show that, I don’t know?

    Or, perhaps some of you guys think Nilsson is a better defensive player than Gagner. I don’t see it.

    It’s true that a forward can be responsible for his team being hemmed in his own zone and not get an error, just as it’s true a forward can help hem the team in the offensive zone and not get an assist. But goals and assists tell us who is doing most to score goals and I think the error can show us who is doing most to allow them.

    I can see RiversQ your mind is made up on this. But I’d suggest over the next few games you try to assign errors yourself. At my blog, I’m going to be doing this, and I’ll be asking for comments, so I’d invite you to give it a go, see what you think of it then. I’ve done this on the HF Boards, and learned a lot and changed my mind about a few things, including the assignment of a few errors.

    As I said, I’m not perfect at this, so you’re right to take this for what it’s worth right now. But I’ve yet to see a critique that is particularly convincing that this concept isn’t going to be useful.

  66. dstaples says:

    Åndy, I’d also invite you to indulge in a little “guessing” of your own. Try to assign errors for five or six or seven games. Watch the replays closely 15-20 times.

    As for criteria, that sort of thing, I’m in the process of writing up a FAQ on the error, which will deal with this issue. I’m using all the feedback I’ve been getting to address concerns about this stat.

    At this point, the “error” seems a fairly divisive issue, with some people intuitively liking it and accepting it as useful, and others — generally those more wedded to the plus/minus as a powerful indicator of player value — rejecting it as far too subjective.

    It could well be true that on many of the errors I’ve assigned so far that my “guesses” were wrong, and there’s little validity in what I’ve found.

    But I think I have found a few interesting things, and not just that Horcoff is so strong defensively. I’ve also found that despite their relatively strong plus/minus rankings, Robert Nilsson and Joni Pitkanen aren’t great defenders, and equalled at even strength by a number of other Oilers forwards and defenders.

    To me, that’s worth something, but again, as you say, my guesses could be right or they could be wrong.

    We are in agreement on one thing — the concept needs refinement.

    Still, I’m convinced it will work and enthusiastic about it, and that comes from having actually done this work to date, which has led me to believe it has promise.

  67. dstaples says:

    As for my bias, yes, this is an issue.

    I’m biased because I like certain players (Smid), but I’m not a big fan of others (Reasoner). I’m aware of this and have tried to make up for this by bending over backwards (which may mean Smid has been assigned more errors than he deserves).

    I’m also biased in seeing something in this stat in that it was my idea, and we all love our own ideas, do we not Mr. Grabia! Hard to kill our babies, right! So this is a fair comment, but it’s true of every idea that’s ever been hatched, and, in the end, the “error” will rise of fall based on what others think, not what I think. So it’s up to you and Bruce and Lowetide and all the folks at the Oilers and Hockey’s Future boards, not me.

    To deal with my bias, I’m looking at assigning errors next season not just to the Oilers, but to a team and to players I don’t give a fig about, either the Detroit Red Wings or the Toronto Maple Leafs. I am also going to try to recruit a few other folks who are interested in watching a team or two, or the same teams as me, and see if we can come up with the same results.

    That’s my plan, anyway.

    Yes, I am obsessed with this, but then, I’ve always been obsessed with my pet projects.

  68. Dennis says:

    Every now and then I tape a game and go back over it and mark scoring chances for and against and list what lines were on the ice.

    If you do that and the same guy ends up being on a line that’s outchancing or getting outchanced, there’s a good chance you’re on to something.

    Staples, you should try this. And you should also try responding when I kick your ass like I did over on HF when you said you didn’t see a problem with Sanderson:)

    I know LT didn’t crib from me but when he listed GO60, I did the same thing over at HF only I went further to say that of the vets, Sandy’s the second biggest sieve to Reasoner. And while I think Reasoner’s done, at least we can soften his figures with a brush dipped in quality of comp.

  69. Bruce says:

    At this point, the “error” seems a fairly divisive issue, with some people intuitively liking it and accepting it as useful, and others — generally those more wedded to the plus/minus as a powerful indicator of player value — rejecting it as far too subjective.

    I don’t fit your profile David, I’ve been a huge plus/minus buff for years, but I like the potential of your errors too. I think the two stats complement each other precisely because one is objective and the other subjective. They can be combined with each other and with other stats. e.g. Steve Staios has made 17 errors (your stat) and been on the ice for 47 ESGA (Desjardins’). (One is thru 50 games and the other thru 52, so this ain’t perfect, just a f’rinstance of what can be done.) Therefore Steve has been deemed at least partly to blame on 36% of the goals against that occur when he’s on the ice. Gilbert is 16 of 38 (42%) , Smid 15 of 33 (45%), Pitkanen 16 of 27 (59%), Grebeshkov 16 of 24 (67%). So Grebs has by far the worst ratio of goals against scored due to his own errors. So does this mean that other than the 5% of the game he’s a secret agent for the other team (Tom Barrett’s great descrition of Grebs), he’s actually pretty good at keeping us out of trouble?

    How does he compare at errors per 60? Grebs 1.44; Smid 1.14; Pitkanen 1.14; Staios 0.89; Gilbert 0.83; yeah that kind of squares with what I’ve seen this season. Grebs’ GA/60 ranking is actually half a goal or so better than Tarnstrom, Staios, Smid … one tends to remember the goal-bleeding blunders, and he makes more than his share of them, but otherwise maybe he is fairly effective. Again, that kind of squares with what I’ve seen of Grebeshkov, it just adds an extra dimension to the already-telling stat of GA/60. Kind of interesting IMO.

    On the other hand maybe all it tells us is that Staples hates Grebeshkov and will blame him for a GA every chance he gets. That’s where a much larger sample size is essential.

    I’m looking at assigning errors next season not just to the Oilers, but to a team and to players I don’t give a fig about, either the Detroit Red Wings or the Toronto Maple Leafs.

    For a (much) larger sample size,I suggest Toronto. Should be good for comic relief too … e.g. Andy Wozniewski(sp.?) has a surprisingly decent +/-, but commits blunders that combine the worst qualities of Smid, Greene and Grebeshkov with a little seasoning of MAB thrown in/up/away for good measure. I just LOVE Andy Wozniewski highlights.

    I am also going to try to recruit a few other folks who are interested in watching a team or two, or the same teams as me, and see if we can come up with the same results.

    If I have the time in the fall, I will volunteer to take on a team. I think there’s some real potential in this idea, and whether or not it flies I think a guy could learn a hell of a lot in the process.

    Come to think of it, I’ll take Toronto. My brother’s a Leafs fan… :D

  70. Slipper says:

    Do these errors account for what happens a minute before a goal, or are they only assigned to events directly preceding a goal against? Is there a maximum number of errors assigned per goal?

    What about failing to bury a chance at one end of the rink, or shooting for the far corner, missing entirely, and having the puck ring around the boards creating an odd man rush?

    Remember Fillpula’s last minute game winner against the Oilers early in the season? Who gets to wear the goat horns on that goal?

    There was Roloson with the shoddy positioning, Tarnstrom blew a wheel, Pouliout made the bad change, and Gagner didn’t forecheck. Does anyone remember who turned the puck over in the Detroit end? Take away any single one of those “errors” and no goal occurs. Take away the very first one, or hell, have the Oilers score, and none of the rest occur at all.

  71. dstaples says:

    Dennis: Yes, I remember when you did a post at MC79 earlier this year where you marked down scoring chances for and against when players were on the ice. That was good stuff.

    As for you kicking my ass over Sanderson, I did miss that, but feel free to refresh my memory. It sounds like you were suggesting Sanderson has faced weak competition, so if he has few errors, that would help explain it.

    That’s a fair enough critique.

    The “error” stat doesn’t factor in quality of competition, but I think it could if there was an “error” stat for every team.

    Right now, to know the true defensive worth of an Oiler, I’d suggest the best way would be to look at the error stat and quality of competition combined.

  72. dstaples says:

    Bruce, that would be great if you would take on a team next fall. Super!

    I like the way you’re combining the error stat with the other stats, and I agree, the error stat must be used in combination with other stats to get a better idea of a player’s value.

    The thing that caught my was the errors vs goal against percentage. Interesting that Greb has been deemed responsible (by me) for such a high percentage of goals he has been on the ice for.

    I can say that at this office, as Tom Barrett has raged against “Garbagekov” I’ve been saying, “Give it time, Tom, maybe he’ll turn it around.” So I don’t believe I have any hate on for the guy.

    Slipper, you asked about the Fillpula goal. We talked about that earlier in this thread. I assigned errors to Pouliot and Tarnstrom, I believe. I haven’t looked at the goal again, but I assign errors on the following basis.

    Hockey is a game of mistake and recovery, mistake and recovery. Finally, there is a mistake and no recovery and a goal is scored.

    In assigning an error, the clock starts ticking with that first mistake or breakdown that leads directly to a goal. Everything is back under control, then someone screws up (bad line change, missed check, bad rebound, missed assignment, bad pass, turnover). After this initial breakdown, the primary error, other screw ups often occur (most often, defenders fail to pick up the attackers in dangerous scoring positions) and secondary errors are assigned as well.

    Finally, errors go to the goalie, in the end, if the shot that comes out of this sequence is save-able, but he blows it, or gives out a big juicy rebound, as poor Roloson has done so many, many times this year. The most common notation in my log bog is “Roloson, bad rebound.”

    As I mentioned to Andy, I’m working on a F.A.Q. on the “error,” as one is badly needed.

    I should also say I watch most goals 15-20 times before assigning errors.

    . . . unless, of course, the play involves that frakkin commie Garbagekov, where if he’s on the ice, I simply immediately assign the bum an error ;).

  73. Dennis says:

    Staples: my conclusion was that while Sanderson was the fifth easiest Oiler to score upon, as it were, three of the guys in front of him hadn’t played 30 games and weren’t over the age of 25. So in terms of vets, Sanderson was the second easiest guy to score upon and the only guy he lead was Reasoner and we can give Marty at least some kind of a break because he was playing tough comp.

    Listen, I like Marty but I think he’s done so I won’t fight you on that one. But when I took a piece out of Sanderson, specifically to point out how the Pitkanen trade wasn’t just for Lupul and how the Oilers had to give up Gator and take on a useless 1.5 mill bill attached to Sanderson in the process, you said you didn’t think he’d played all that badly.

    I, sir, as I am so often wont to do, begged to differ;)

  74. Dennis says:

    One last thing. Not just because I’ve done it because, hey, anyone can watch a game and count scoring chances and list who’s on the ice for/against.

    But, AFAIC, that’s the hands down best way to determine a players worth for that game. It lists who was on the ice for chances and who they were on the ice against and what more do you need?

  75. Bruce says:

    Do these errors account for what happens a minute before a goal, or are they only assigned to events directly preceding a goal against?

    Slipper, I think the idea is that errors are (somewhat) like scoring points, there can be a huge sequence leading up to a goal, but the error-counter is only triggered by the goal being scored and analyzing the events that directly contributed to it. Rather than assigned arbitrarily as in points (the last up-to-three guys that touched the puck), errors are assigned through judgement, which is fairly likely to include the original change of possession (a.k.a. turnover) whether it be a cough job, a failure to make a routine break-out play, or your excellent example of a far-side slapshot that rings the boards and traps the shooter, his two linemates, and Pitkanen.

    Is there a maximum number of errors assigned per goal?

    Three, just like scoring points. I think there is a limit of one per player, unlike baseball where a guy can make, say, a fielding error and then throw the ball away. The key being each error costs his team something, an out or a base, whereas in hockey the end result will always be just one goal against, even if the same guy coughs the puck, lollygags on the backcheck, fails to identify his man and doesn’t clear the rebound, it’s just one error. But Staples might have a different view on that.

    Is it perfect? No. As somebody pointed out the multiple deflection fluke off of Brett McLean’s skate was a harsh one to be finding fault on. That said, McLean was credited with a GOAL on the play, so that mechanism ain’t exactly perfect either. Shit happens.

    The error doesn’t work in baseball because it unfairly singles out players that actually manage to get themselves involved in the play. The good defenders might make a few more errors because they actually get to the fucking ball.

    The error has its weaknesses as a stand-alone stat for that very reason, RQ; however errors and range factor together can paint a reasonable picture of a fielder’s ability. The very good defenders get to more balls and still make most of the plays.

  76. dstaples says:

    Bruce, your description of the error is the same as what I’ve been thinkng.

    As you say you may be able to help with this stat next fall, can you send me your email so we can work out details later on?

    As I said, this stat only works if we can duplicate each other’s work on assigning errors, and that will take some person-power.

    Cheers David. dstaples@thejournal.canwest.com

  77. RiversQ says:

    Dennis said…

    But, AFAIC, that’s the hands down best way to determine a players worth for that game. It lists who was on the ice for chances and who they were on the ice against and what more do you need?

    Shhh, Dennis.

    That’s way too fucking obvious and too easily adapted to objective analysis. Hell, you might actually be able to define scoring chances before you count them.

    That’s no fun.

Leave a Reply

Want to join the discussion?
Feel free to contribute!

Leave a Reply

© Copyright - Lowetide.ca