HALL ENTRY

Tyler Dellow has two brilliant posts up re: Taylor Hall and the mysterious and ever-increasing drumbeat on his season. The boxcars remain sublime, but the underlying issue with shots for while on ice is real and troubling.

 TAYLOR HALL PLAYER CARD

hall esThe boxcars are terrific, shots on goal in the same range as one year ago and his shooting percentage up .5%. Nothing major there, just a gifted young player moving the needle and emerging as one of the best players in the game. The concern—the Charlie Watts drumbeat in the driving rhythm section—is that Corsi for % at 5×5. It’s gone from north of 50% in his first two seasons to 43.6% this year. Hammer on a drum.

Tyler Dellow’s ability to parse the numbers is well established, but I think we’re seeing him identify a problem and offer a solution in real time. Dallas Eakins needs to read this NOW. I’m not going to quote the entire article, because this is something we all need to read and understand, but wanted to tease the two part series with this:

  • Dellow: What does this all tell us? I see three issues. First, the Oilers have become much more likely to dump pucks into the offensive zone with Hall on the ice at 5v5. Second, (in this sample at least but keep in mind that it was picked because it was pretty representative of the whole), they suck at retrieving those pucks. Third, they’ve gotten worse at generating shot attempts when they do carry the puck into the offensive zone.

If you’ve followed Eric Tulsky’s work on zone entries, you know there’s a very strong connection between carrying the puck in with possession and scoring goals. Tyler’s comment here is very disturbing, because it speaks to a change in strategy (by Hall or by the coaching staff, also addressed in the articles) and that strategy is killing an otherwise brilliant season by 4.

Dellow’s articles are here and here.

changeI think we’re at a point now where the information provided by Dellow, Tulsky and a few others are beyond the things uncovered by Bill James (things like k/w ratio predicting future success, platoon advantage being very pronounced, a 21-year old rookie being more likely to have an impact career than a 24-year old rookie with the same numbers) in his abstracts.

The articles by Dellow are providing real research and analysis (and a solution) about a real time problem. In a world where ‘advanced stats’ is thrown around way too often, Dellow’s two-part item on Hall identifies the problem and offers a solution. We should read it, digest it and then count the days until the Oilers change their zone entry credo.

written by

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

140 Responses to "HALL ENTRY"

  1. borisnikov says:

    How someone has not yet taken a flyer on Dellow and his abilities yet is beyond me. Big, big, brain.

  2. Clay says:

    Ok, so the difference in Hall’s Corsi For % is HUGE, and it begs the question. If a player this good can have his CF% affected that much by just a change in team strategy, then are stats like CF%, etc, really just a measure of team strategy, and not the player himself?

    I mean, if I’m playing armchair GM (like everyone else here), and there’s trade rumours floating around about the Oilers trading for player X, and I go and look up player X’s advanced stats, how can I trust that those stats reflect his ability as a player, and not the team’s penchant for dump ‘n chase?

    You, know, unless I’m prepared to do Dellow-type research each time, which would really defeat the purpose of having statistics…

  3. steveb12344 says:

    borisnikov:
    How someone has not yet taken a flyer on Dellow and his abilities yet is beyond me. Big, big, brain.

    Yeah, he’s got an incredible grasp on the grade 9 math skills it takes to figure this stuff out.

    What if it’s just a matter of his other 4 linemates on the ice dragging him down. For example if you put Sydney Crosby out there with 4 ECHL’ers his Corsi would be very bad. Would that mean that Crosby sucks? No, it means there are 5 guys out there playing as a team affecting each others individual possession numbers.

    IMO that is likely the problem with Hall. The guy is playing as well, or better than he ever has.

  4. Ducey says:

    I am not convinced Corsi is a true measure of possession with these Oilers. They spend lots of time with the puck, often cycling it down low but unlike a team like BOS, it doesn’t result in a pass to the slot and a shot. The team is well noted for trying to pass the puck into the net.

    They need to shoot more and I’d like to see the Corgi’s go forth and multiply, but as I understand it the idea is to use Corsi as a representation of possession.

    Maybe this is a stupid question, but why doesn’t someone fire up the Betamax and just time the amount of offensive zone possession?

  5. Ben says:

    It never fucking ceases to amaze me.

    With the immense stakes involved in professional sport, the staggering dollars invested, the collective efforts of highly experienced people….that such critically sound, relevant analytical writing is being produced FOR FREE by a CURIOUS FAN.

    Whatever it is that Tyler’s auditioning for here, I sure hope he gets it.

  6. Surly says:

    Sort out the D and it will help sort out these zone entry problems. How I dearly miss those Pronger passes up the ice. One the tape, caught in stride, allowed our forwards to keep defenders on their heels opposed to their toes.

    With Rogers taking over NHL coverage next year, I’d love to see them incorporate an analyst like Dellow (either in print on sportsnet.ca or on the broadcast if they can find someone who can handle live TV). Alas, we’ll likely be served up with the usual talking heads.

  7. Lowetide says:

    Ben:
    It never fucking ceases to amaze me.

    With the immense stakes involved in professional sport, the staggering dollars invested, the collective efforts of highly experienced people….that such critically sound, relevant analytical writing is being produced FOR FREE by a CURIOUS FAN.

    Whatever it is that Tyler’s auditioning for here, I sure hope he gets it.

    It is incredible. Dellow should be locked in a room and forced to produce this all day. For pay of course. :-)

  8. mumbai max says:

    OK, I just read both articles, and it is impressive. One thing seems to be missing, but I am sure I am just failing to see it. So, dumping in the puck results in a 50% loss of result compared to carrying it in. BUT, isn’t it true true that it is MUCH easier to dump it in than pass or carry it in? Wouldn’t the increased chance of getting into the zone at all with a dump in, not even out the lower result of the dumping. In other words with a dump in, you are IN, with a carry in, much of the time there will be a turnover at the blue line. In other words you have more quantity of less quality, instead of the other way around. I am sure I have missed why this is not true. Seems too simple to be true :-)
    (Hunkering down, preparing for barrage)

  9. Lowetide says:

    mumbai max:
    OK, I just read both articles, and it is impressive. One thing seems to be missing, but I am sure I am just failing to see it. So, dumping in the puck results in a 50% loss of result compared to carrying it in. BUT, isn’t it true true that it is MUCH easier to dump it in than pass or carry it in? Wouldn’t the increased chance of getting into the zone at all with a dump in, not even out the lower result of the dumping. In other words with a dump in, you are IN, with a carry in, much of the time there will be a turnover at the blue line. In other words you have more quantity of less quality, instead of the other way around. I am sure I have missed why this is not true. Seems too simple to be true :-)
    (Hunkering down, preparing for barrage)

    Not at all. There are no doubt reasons the Oilers are dumping it in more, one of them being the defense has been watching tape and don’t want Hall on an uncontested entry. Lots of ways to counter this:

    1. pass back and re-set sortie
    2. shoot it in
    3. have a set play where player with the puck passes to a fast moving linemate for forced entry (resulting in successful entry or penalty)
    4. one on one deke to beat the defender

    There are options.

  10. sliderule says:

    Ducey:
    I am not convinced Corsi is a true measure of possession with these Oilers.They spend lots of time with the puck, often cycling it down low but unlike a team like BOS, it doesn’t result in a pass to the slot and a shot. The team is well noted for trying to pass the puck into the net.

    They need to shoot more and I’d like to see the Corgi’s go forth and multiply, but as I understand it the idea is to use Corsi as a representation of possession.

    Maybe this is a stupid question, but why doesn’t someone fire up the Betamax and just time the amount of offensive zone possession?

    You make two very important points that the oilers seem to cycle the puck particularly Nuges line with no forward ever getting into a shooting position.They maintain the puck but never get a shot.

    The possession time through all three zones should be easy to measure.I would think that would be an interesting comparison when used with Corsi.

  11. steveb12344 says:

    I’m sure that the systems have factored into it, but I don’t know why most of you just completely discount the effect of linemates.

    Corsi is a team stat that is determined by the play of 5 players working together. If Hall was out there with say, Crosby, and Giroux. With Weber and Suter on D. His Corsi would be through the roof, even if he was just playing ok.

    Obviously if he was playing against those guys with crap linemates it would be the opposite, even if he was playing lights out.

    So I say if all his other stats look good, than maybe his bad corsi is more of an indictment of his linemates (including D) than of Hall himself.

  12. borisnikov says:

    Ben:
    It never fucking ceases to amaze me.

    With the immense stakes involved in professional sport, the staggering dollars invested, the collective efforts of highly experienced people….that such critically sound, relevant analytical writing is being produced FOR FREE by a CURIOUS FAN.

    Whatever it is that Tyler’s auditioning for here, I sure hope he gets it.

    That’s what I meant… My commenting depth needs work clearly:)

  13. justDOit says:

    Then there’s this from the latest ’30 Thoughts’ article:

    9. One Jet said the biggest structural difference made by the new coach is on the breakout. He explained the defencemen felt they didn’t have enough options to make plays because the forwards were taking off. The blue-liners are now instructed to wait or delay if they can instead of just getting rid of it, and the forwards are asked to take a more inside position than an outside one.

  14. borisnikov says:

    steveb12344:

    If Hall was out there with say, Crosby, and Giroux.With Weber and Suter on D.His Corsi would be through the roof…

    But he’s not out there with those guys…

    The point of this analysis, I think, is to find what works best for any situation given. How do you maximize Hall’s abilities to make the team around him better? You analyze the current state to see what can be improved upon.

  15. sliderule says:

    Hall has been called out by team Canada coaches for turning puck over .

    Whether on his own or by Eakins orders he is trying to avoid turning it over by dumping it in.

    The teams that are successful at the dump in like Coyotes and Canucks have plays that are successfully used in recovering the puck.These teams have the dump in as the first and nearly only option.The oilers use the dump in as a bail out option so forwards are often standing still and can’t recover the puck.

    I would say if Hall has the puck against three or four defenders the play should be dump it or pass it .The oiler forwards would know what he is going to do.If it’s Nuge he should carry it in as he has the moves that Hall doesn’t have.

    Over to you Eakins.lol

  16. Romulus Apotheosis says:

    It’s that MacT item in Staples interview:

    Staples: When you look at the results last year in terms of their two-way play they seemed to be doing a little bit better last season, all of those guys, compared to how they’re doing this season.

    MacTavish: Based on the stats?

    Staples: Yes, based on the stats. Now I know there is an injury issue with Hall, he had a major injury and didn’t have one last year, and Hopkins is coming off shoulder surgery, and it took Hall about two months to get going last year, but they all of them seemed to have dropped a little bit and Gagner the same thing. Again, he’s had an injury. What do you think of that assessment that they’ve dropped a bit?

    MacTavish: I don’t think it’s accurate in what I’m seeing. When you talk to people that know hockey very well they will tell you that they are seeing structural changes in Taylor Hall. His game is changing.

    Staples: What do you mean by that?

    MacTavish: Let’s look at the goal he scored last game (vs. Tampa Bay) in four on four. He hammered the puck down the right side and laid it in deep. That’s a big change.”

    Staples: Dumping it in, you mean?

    MacTavish: Yes. Just living to fight another day. He hunted down the forecheck, got it over to Sam, Sam back to Ference, through a simple play. Where before he would try and beat those two guys and take it to the net, turn it over and it would be going to the other way. So that is a big structural change in the game, the way that he’s thinking. When you watch good teams — and these are the things we are trying to impart on these players — you watch good teams. We went to New York and watched Anaheim play. Arguably the best line in the game is Getzlaf, Perry and at that time it was Penner, they’ve made a change since then. And I watched them play throughout the game and Perry had two chances to score the whole game, but he made the right play with the puck all the time. And that’s what we’re trying to impart on our players. Let’s make better decisions with the puck. Like, our guys are inexperienced on how to win. And so they’re making a lot of poor tactical decisions with the puck. Especially earlier in the year, it was unbelievable the amount of turnovers. Most of our losses were self-driven. And, you know, it’s getting them to change that mentality. I’m seeing that. The other guy I’m really encouraged by is Hopkins. The last little while the speed in his game has picked up, which it had to. He’s getting stronger. He’s now a real solid penalty killer for us. So I see those changes in gamesmanship changing for the better. Our structure as a team is much better.

    http://blogs.edmontonjournal.com/2014/01/11/the-craig-mactavish-interview-part-1-im-not-arguing-that-were-close-were-not/

    MacT and Eakins bought into the Ruff, Spector et al. nonsense that Hall needed some kind of Old Manning, earn his keep line of goods.

    They can only see the turnovers.

    Now… this runs entirely counter to Eakins language about not using dump ins and very recently he confessed to Stauffer that they believed they had fucked Hall up worrying about dump ins/carry ins and told him to relax and play his game.

    I think like with the Summer language about skill and then the post-Gagner jaw waiver pick ups and trades suggest this is a hockey mind divided.

    We are seeing an Old Time Hockey Guy and a New Order Hockey Guy in the same person… humans are walking contradictions

  17. Lowetide says:

    Rom: Truth. I think the Gagner jaw also contributed to the in-season additions.

  18. Romulus Apotheosis says:

    Eakins on Stauffer’s show talking Hall
    http://www.630ched.com/oilers-now/
    jan 27th 2nd half hour

    (quoted in Tyler’s piece)

  19. LMHF#1 says:

    So, they have Hall dumping it in to avoid turnovers even though most of their dump-ins ARE turnovers…

    And he’s forcing more bad passes due to more chaos around the puck…

    If you ever wonder why I don’t agree with you on Eakins LT, take Hall and Yakupov’s seasons. They show me a coach that, despite being a strong communicator, is not actually that smart.

    His comments about watching all of Yakupov’s goals from last season etc and then not figuring out how to deploy him effectively was a big tell.

  20. Lowetide says:

    LMHF#1:
    So, they have Hall dumping it in to avoid turnovers even though most of their dump-ins ARE turnovers…

    And he’s forcing more bad passes due to more chaos around the puck…

    If you ever wonder why I don’t agree with you on Eakins LT, take Hall and Yakupov’s seasons. They show me a coach that, despite being a strong communicator, is not actually that smart.

    His comments about watching all of Yakupov’s goals from last season etc and then not figuring out how to deploy him effectively was a big tell.

    I think you can be smart and inexperienced too, and that’s the case here. And some of the smartest people I know are also some of the most stubborn. :-)

  21. LMHF#1 says:

    Lowetide: I think you can be smart and inexperienced too, and that’s the case here. And some of the smartest people I know are also some of the most stubborn.

    Let’s hope so!

    Viewing it in the context of one year is always difficult when Game Six was so long ago.

  22. oilersfan says:

    As for Hall’s drop in Corsi, this is my theory.

    I think that Krueger made a system based on the Oilers’ skill. Unfortunately, the only line that could implement the system was Hall-RNH-Eberle, which is why their corsis were positive most of the time even against good teams like St. Louis, Vancouver, etc. But when they weren’t on the ice, the other players didn’t have the skill to implement the system and those players had worse corsis.

    So Eakins comes in and implements a more traditional system, but one that doesn’t take into account the unique skill set of the top line, thus reducing their corsis as their puck retrieval skills are not as good as their puck possession skills.

    WHat do you guys think of that? I am pretty sure I am correct.

    On another note, what do you advanced stats guys use to offset the zonestarts? I know Hendricks and Gordon have a brutal corsi and corsi rel but it isn’t fair to analyze apples to grapefruits when those two guys start every shift in their own end and RNH/Hall start them all in the ozone.

  23. russ99 says:

    Fantastic article by Dellow, everyone should take the time to read it.

    Finally, there’s proof – we’re dumping more, and not getting the puck back after the dump.

    Considering Hall’s play last year (also proved by MC79 Hockey’s numbers) you just can’t say that’s because of the players doing their own thing.

    Eakins can say that he doesn’t want a dump and chase offense, but the proof is on the ice and in the numbers.

  24. LMHF#1 says:

    russ99:

    Eakins can say that he doesn’t want a dump and chase offense, but the proof is on the ice and in the numbers.

    He can say it isn’t dump and chase because he calls it “chip it past the defender” or something like that. Doesn’t change what’s going on out there.

  25. book¡je says:

    As per Dellow producing the ‘best stuff’, I would say its the best stuff that we, as fans, get to see.

    The Oilers have an analytics team as well as some consultants on contract (Darkhorse Anylytics) or at least they did last year. For all we know, Eakin’s admission that they had messed up Hall’s game could come directly from the analytics guys doing a similar analysis to Dellow.

    Also, while I expect that management never visits this blog (perhaps Eakins at times as he says he reads everything about his team online) I do expect that the analytics geeks are regulars here and other sites where there is some/lots of talk about numbers, so if they haven’t picked up on what Dellow has, they will soon do so and hopefully carry it to management the next time they get their 15 minutes with them.

    I think the Oilers management are at the stage where they are consulting their analytics guys on a lot of decisions and ignoring them on half of those decisions.

    I should add that I agree that Dellow is putting out some great stuff.

  26. G Money says:

    The idea of a team carrying the puck in vs dumping it in and treating those as some black-white either-or is a false dichotomy.

    EVERY team has zone entries that are a blend of dump and chase, chipping past the defender (which, with all due respect is NOT the same as dump and chase), and carrying it in.

    These are situational decisions. If a team is challenging four across at the blue line, you either dump it in, chip it in, or you circle back around and try again.

    Some harsh realities for us to deal with:

    - Teams know that the Oilers are too small to win puck battles. They stand our forwards up (especially Hall) every chance they get, and bet that they’ll be able to win the puck back on a chip/dump. And for the bigger teams, they’re almost always right.

    - Dellow’s work shows that Taylor Hall now dumps the puck in much more often, rather than trying to force a carry-in resulting in a turnover and break the other way. This is a good thing, not a bad thing. The learning process means developing situational awareness about when to carry it in and when to dump/chip it in. Overdoing it is part of the learning process.

    - Circling back and trying it again only works if the other team is not disciplined enough to maintain their positioning on the second go round. This is one reason why the Oilers have so much trouble against defensively disciplined teams aka “good” teams.

    - Against the big defensively disciplined teams, the Oilers can neither dump it in nor carry it in, and they get blown out of the water Corsi-wise. No surprises there. This won’t change until the learning process (when to dump vs when to carry) is farther along, and also until the Oilers have some bigger players in the Top 6 capable of winning puck battles on dump ins.

  27. Woodguy says:

    steveb12344,

    I’m sure that the systems have factored into it, but I don’t know why most of you just completely discount the effect of linemates.

    Because he’s had the same line mates year over year for the most part:

    4 most common Forwards for Hall

    13/14 Hall total 718 5v5 min
    With:
    RNH 513min
    Eberle 383min
    Gagner 177min
    Hemsky 126min

    12/13 Hall total 651min
    With:
    Eberle 432min
    RNH 389min
    Hemsky 128min
    Gagner 114min

    The difference there is pretty much nothing.

    Its the fact that his results with the same players are off that allows for this type of analysis.

  28. Romulus Apotheosis says:

    LMHF#1: If you ever wonder why I don’t agree with you on Eakins LT, take Hall and Yakupov’s seasons. They show me a coach that, despite being a strong communicator, is not actually that smart.

    I don’t think this is the case.

    The thing about his communication is that it really isn’t how he says it but what he says. He says all the right stuff about B2Bs, hits, puck possession, carry ins, etc.

    That’s the smart part.

    Now, as Tyler suggests, it’s possible Eakins doesn’t believe in what he’s saying (that would be cynicism, not lack of smarts), or that he he hasn’t been able to translate his philosophy into action yet (or maybe never).

    Personally, I think it’s the latter combined with what I’ve outlined above… A guy like Eakins is of two minds. He knows puck movement is key, he knows fighting and hits are largely irrelevant… but that old school part of him still thinks “heaviness” is really important.

  29. Woodguy says:

    Clay: If a player this good can have his CF% affected that much by just a change in team strategy, then are stats like CF%, etc, really just a measure of team strategy, and not the player himself?

    This is a great point and a reason why its informative to look at a player’s results relative to those on the same team.

    Then of course you have to take into account zonestars, quality of opp, TOI etc.

  30. G Money says:

    Woodguy: This is a great point and a reason why its informative to look at a player’s results relative to those on the same team.

    Yes, but to some extent this *adds to* rather than subtracts from the mystery that is Taylor Hall this year.

    It’s not just that his Corsi has fallen – for the first time that I can recall, his CorsiRel is actually negative. The team’s shot differential is better with Hall off the ice than on the ice, and given that so much of the Oiler offense keys off Hall, this is a brain bender, even in light of Dellow’s work.

  31. LMHF#1 says:

    Romulus Apotheosis: I don’t think this is the case.

    The thing about his communication is that it really isn’t how he says it but what he says. He says all the right stuff about B2Bs, hits, puck possession, carry ins, etc.

    That’s the smart part.

    Now, as Tyler suggests, it’s possible Eakins doesn’t believe in what he’s saying (that would be cynicism, not lack of smarts), or that he he hasn’t been able to translate his philosophy into action yet (or maybe never).

    Personally, I think it’s the latter combined with what I’ve outlined above… A guy like Eakins is of two minds. He knows puck movement is key, he knows fighting and hits are largely irrelevant… but that old school part of him still thinks “heaviness” is really important.

    Let’s take the Yakupov example because I’m more interested in it than Hall. Hall will be fine.

    If you can watch as much tape as Eakins says he did – study it – and then not be able to extract the bits you need to help your player be put in a position where he can be effective I seriously question your smarts when it comes to hockey. I guess it could be stubbornness as LT mentioned, but then why study? I don’t think the man knows what works out there and he’s yet to convince me otherwise.

    Hope that changes.

  32. Romulus Apotheosis says:

    G Money: The idea of a team carrying the puck in vs dumping it in and treating those as some black-white either-or is a false dichotomy.

    I don’t think this is responding to the literature on the subject. Every piece I’ve read on it, including Tyler’s recent work, notes that sometimes the carry in simply isn’t there and you have to dump in.

    This is about what is most effective in the aggregate and what a team should strive for in general.

    G Money: – Dellow’s work shows that Taylor Hall now dumps the puck in much more often, rather than trying to force a carry-in resulting in a turnover and break the other way. This is a good thing, not a bad thing. The learning process means developing situational awareness about when to carry it in and when to dump/chip it in. Overdoing it is part of the learning process.

    For this to be true, you’d have to show that Hall’s game needed work, ie., that he was getting poor results in the past.

    But, he wasn’t.

    We are trying to correct a phantom here. And, in doing so we’ve attained much worse results. Not good.

  33. Romulus Apotheosis says:

    LMHF#1: Let’s take the Yakupov example because I’m more interested in it than Hall. Hall will be fine.

    If you can watch as much tape as Eakins says he did – study it – and then not be able to extract the bits you need to help your player be put in a position where he can be effective I seriously question your smarts when it comes to hockey. I guess it could be stubbornness as LT mentioned, but then why study? I don’t think the man knows what works out there and he’s yet to convince me otherwise.

    Hope that changes.

    Make your case then.

    What has he said about Yak? What has he done with Yak that shows a disconnect between speech and action? What is it that he should be seeing in Yak’s game that he’s missing? And what information should he have taken from this video, etc. that he isn’t leading to his decisions?

  34. mc79hockey says:

    If you can watch as much tape as Eakins says he did – study it – and then not be able to extract the bits you need to help your player be put in a position where he can be effective I seriously question your smarts when it comes to hockey. I guess it could be stubbornness as LT mentioned, but then why study? I don’t think the man knows what works out there and he’s yet to convince me otherwise.

    I’m going to go on a bit of a commenting jag here but I’ll answer this one first. I think Yak’s a bit of a different leopard than Hall. The issues are different. Even last year, I felt that there were times at 5v5 that Yak looked lost. To me, he’s struggled with making the adjustment to playing 5v5 in the NHL in terms of finding spots on the ice where he can succeed and how he needs to play to succeed (forecheck, move his feet etc.). I think he’s making strides there and he’ll be better for his time with Eakins this year.

  35. mc79hockey says:

    Ok, so the difference in Hall’s Corsi For % is HUGE, and it begs the question. If a player this good can have his CF% affected that much by just a change in team strategy, then are stats like CF%, etc, really just a measure of team strategy, and not the player himself?

    I mean, if I’m playing armchair GM (like everyone else here), and there’s trade rumours floating around about the Oilers trading for player X, and I go and look up player X’s advanced stats, how can I trust that those stats reflect his ability as a player, and not the team’s penchant for dump ‘n chase?

    Most of the time, I think that the stats are reasonably good representations. That said, coaching decisions can affect things. LA and NJ inflate Corsi%, IMO. I think Therrien and Carlyle are cratering people. I’ve been pretty consistent all along that the data is a piece of information, not the entire puzzle. Hall’s an example of that, I think. If the data disconnects with what you think you know, you start asking more questions.

  36. Woodguy says:

    G Money,

    These are situational decisions.

    Its dictated by the coach more than you think.

    Perron said that Hitchock told them to dump and chase on 2 on 2.

  37. mc79hockey says:

    Yeah, he’s got an incredible grasp on the grade 9 math skills it takes to figure this stuff out.

    What if it’s just a matter of his other 4 linemates on the ice dragging him down. For example if you put Sydney Crosby out there with 4 ECHL’ers his Corsi would be very bad. Would that mean that Crosby sucks? No, it means there are 5 guys out there playing as a team affecting each others individual possession numbers.

    IMO that is likely the problem with Hall. The guy is playing as well, or better than he ever has.

    The math isn’t the magic. It’s the ability to apply logic. Let’s take your criticism.

    In 2012-13, Hall played about 60% of his TOI with RNH. They had a 54.6% Corsi% together. This year, Hall has played about 66% of hsi TOI with RNH. They have a 44.2% Corsi% together. The defence is pretty similar this year to last year. You don’t need grade nine math to realize how silly your critique is given these facts.

  38. Romulus Apotheosis says:

    Woodguy:
    G Money,

    These are situational decisions.

    Its dictated by the coach more than you think.

    Perron said that Hitchock told them to dump and chase on 2 on 2.

    on/off topic, this stuff from Nelson in Gregor’s article yesterday is incredible to read… I have no idea what Ewanyk was coached to do prior to Nelson, but it sure sounds like he was clueless about his positioning as a center… If I’m Laxdal reading this… I feel great shame.

    http://oilersnation.com/2014/2/10/update-from-the-farm

    Gregor: Can you elaborate a little bit on the ‘not understanding the system’. Is it hockey sense; is it just the pace of the game where they are not reacting properly, what is it? Not with Ewanyk specifically, but young players in general, what is the hardest thing for them to grasp?

    Nelson: Ah… you know what; we play two man pressure here. A lot of the players we get always play a one man pressure system if it’s a 1-2-2 or a trap, but we’re playing two man pressure and when we play it well it’s very effective. The whole thing is that it’s trust within your teammates and I think for young players coming in, if they’re not used to that sort of system they have to fail, and then once they play it well they see that good things happen, and then they have confidence in it.

    That’s what we saw from Travis, I think that it was just a situation that if he was our high F3, his reads just weren’t consistently correct at the start of the year. Now he’s getting a lot better; he understands that if the puck is chipped by our pinching defenceman that he has to be our safety valve. Where at the start of the year he would just kind of watch the pinching D and then all of the sudden the guy would come off of his back and that would lead to an odd man rush against.

    So it’s just reading situations in the D zone down low, because he is a centreman. He’ll be our F1 down low. You need to be working with the defence and everybody has to work together in that area because when you’re playing against creative players you have different reads you have to make. Also, off of the rush if he’s back checking hard he has to talk and communicate with our defencemen.

    So all of these little things, it’s just the reads and once again it’s constant teaching from both me, Gerry [Fleming] and Rocky [Thompson] and it’s one of those things that you have to be patient with and just teach through video, and also we have to teach during the game. After a shift you have to point out things and we have to be careful that we are not over coaching because no player wants to hear what they did wrong every shift that they are out there. We have to have good communication with the player and Travis is a guy that, he may have to fail 25 times before he gets it, but now he’s starting to get it and he’s playing some pretty good hockey for us.

  39. mc79hockey says:

    Maybe this is a stupid question, but why doesn’t someone fire up the Betamax and just time the amount of offensive zone possession?

    I did it, although I haven’t posted it yet. It’s not very different.

  40. mc79hockey says:

    Wouldn’t the increased chance of getting into the zone at all with a dump in, not even out the lower result of the dumping. In other words with a dump in, you are IN, with a carry in, much of the time there will be a turnover at the blue line. In other words you have more quantity of less quality, instead of the other way around.

    That’s a good question. Eric Tulsky touches on it here. Implicitly, he’s saying that teams are way too conservative: http://nhlnumbers.com/2013/2/20/zone-entries-and-the-sloan-sports-analytics-conference

  41. book¡je says:

    mc79hockey:
    Yeah, he’s got an incredible grasp on the grade 9 math skills it takes to figure this stuff out.


    What if it’s just a matter of his other 4 linemates on the ice dragging him down. For example if you put Sydney Crosby out there with 4 ECHL’ers his Corsi would be very bad. Would that mean that Crosby sucks? No, it means there are 5 guys out there playing as a team affecting each others individual possession numbers.

    IMO that is likely the problem with Hall. The guy is playing as well, or better than he ever has.

    The math isn’t the magic.It’s the ability to apply logic.Let’s take your criticism.

    In 2012-13, Hall played about 60% of his TOI with RNH.They had a 54.6% Corsi% together.This year, Hall has played about 66% of hsi TOI with RNH.They have a 44.2% Corsi% together.The defence is pretty similar this year to last year.You don’t need grade nine math to realize how silly your critique is given these facts.

    mc79hockey:
    Yeah, he’s got an incredible grasp on the grade 9 math skills it takes to figure this stuff out.


    What if it’s just a matter of his other 4 linemates on the ice dragging him down. For example if you put Sydney Crosby out there with 4 ECHL’ers his Corsi would be very bad. Would that mean that Crosby sucks? No, it means there are 5 guys out there playing as a team affecting each others individual possession numbers.

    IMO that is likely the problem with Hall. The guy is playing as well, or better than he ever has.

    The math isn’t the magic.It’s the ability to apply logic.Let’s take your criticism.

    In 2012-13, Hall played about 60% of his TOI with RNH.They had a 54.6% Corsi% together.This year, Hall has played about 66% of hsi TOI with RNH.They have a 44.2% Corsi% together.The defence is pretty similar this year to last year.You don’t need grade nine math to realize how silly your critique is given these facts.

    STEVEB12344, in case you had trouble understanding this, I ran it through Google translate for you and it roughly translates to “Bitch Slap” or something of the sort.

  42. G Money says:

    Romulus Apotheosis: I don’t think this is responding to the literature on the subject. Every piece I’ve read on it, including Tyler’s recent work, notes that sometimes the carry in simply isn’t there and you have to dump in.

    This is about what is most effective in the aggregate and what a team should strive for in general.

    I think you’re agreeing with me because what you’re saying (that it’s not an either-or, that teams must do both) is the same thing I said, but it’s not couched that way, so now I’m not sure!

    Romulus Apotheosis: For this to be true, you’d have to show that Hall’s game needed work, ie., that he was getting poor results in the past.

    But, he wasn’t.

    We are trying to correct a phantom here. And, in doing so we’ve attained much worse results. Not good.

    Well, no, we definitely don’t agree on this one.

    Hall’s game didn’t have to be “poor” before. The question is – could it improve?

    We know Hall’s an offensive dynamo, the question at hand is whether or not he’s as effective at driving overall results as he could be. Very reasonable argument to be made that he’s not (yet).

    Certainly you can argue whether more dump-ins, or better defensive coverage achieves that end. Maybe it doesn’t.

    But suggesting that Hall was good enough last year and so any attempt to get him to play a more complete game isn’t needed doesn’t strike me as reasonable.

  43. Romulus Apotheosis says:

    mc79hockey:
    Maybe this is a stupid question, but why doesn’t someone fire up the Betamax and just time the amount of offensive zone possession?

    I did it, although I haven’t posted it yet.It’s not very different.

    also done here… tracks with corsi/fenwick

    http://www.pensionplanpuppets.com/2013/9/16/4727746/leafs-attack-time-at-the-halfway-mark

  44. G Money says:

    Woodguy:
    G Money,

    These are situational decisions.

    Its dictated by the coach more than you think.

    Perron said that Hitchock told them to dump and chase on 2 on 2.

    Sure. In your example above, if it’s 2 on 2, you dump and chase, and e.g. if it’s 3 on 2 you try to carry it in. Conversely in, say, Babcock’s system, you only dump if it’s four across at the blue line, but try to carry it in otherwise.

    Awareness of the tactical requirements of the system – and the configuration of your players and the other teams players on the ice at any given moment – is a fundamental aspect of what I mean by situational awareness.

  45. Romulus Apotheosis says:

    G Money: I think you’re agreeing with me because what you’re saying (that it’s not an either-or, that teams must do both) is the same thing I said, but it’s not couched that way, so now I’m not sure!

    Well, if it’s not meant as a criticism of published work… it’s just a banal truism. fair enough.

    G Money: Well, no, we definitely don’t agree on this one.
    Hall’s game didn’t have to be “poor” before. The question is – could it improve?
    We know Hall’s an offensive dynamo, the question at hand is whether or not he’s as effective at driving overall results as he could be. Very reasonable argument to be made that he’s not (yet).
    Certainly you can argue whether more dump-ins, or better defensive coverage achieves that end. Maybe it doesn’t.
    But suggesting that Hall was good enough last year and so any attempt to get him to play a more complete game isn’t needed doesn’t strike me as reasonable.

    But, we aren’t discussing an abstraction like: is there room for improvement in any person in any regard.

    We are discussing a specific critique levelled routinely at Hall: that he’s a turnover machine at carrying the puck into the zone (Hemsky gets this a lot too).

    And, not a criticism isolated to internet hucksters and Mark Spector. Ruff believes. MacT believes it. Eakins seems to believe it to some degree.

    The problem is… other than the “by eye” of the turnover, there isn’t a lot of evidence Hall has a unique problem with zone entries.

    Willis’ look at zone entries didn’t show much to complain about:

    http://blogs.edmontonjournal.com/2013/05/24/the-edmonton-oilers-only-have-one-puck-possession-line/

    http://blogs.edmontonjournal.com/2013/02/28/edmonton-oilers-zone-entries-part-2-which-players-drive-the-bus/

    So, in case 1 we have Hall doing very well with zone entries by number and getting a lot of criticism by eye.

    In case 2, this year, we have Hall doing poorly with zone entries by number and the coach saying it’s probably because they’ve over-emphasized things.

  46. G Money says:

    Romulus Apotheosis: But, we aren’t discussing an abstraction like: is there room for improvement in any person in any regard.

    We are discussing a specific critique levelled routinely at Hall: that he’s a turnover machine at carrying the puck into the zone (Hemsky gets this a lot too).

    I’m responding more to the general remarks you made: … you’d have to show that Hall’s game needed work, ie., that he was getting poor results in the past. and We are trying to correct a phantom here.

    Hall doesn’t have to be a turnover machine to work on improving his situational awareness of whether to carry or whether to dump. (He also unquestionably needs to improve his defensive coverage, still).

    The fact that Hall’s numbers have declined this year isn’t necessarily indicative of a problem.

    It’s like a golf swing. You can be pretty good, but still have issues. While you work on those individual issues, you have to concentrate like hell and focus on the one or two flawed aspects of your swing, and while you’re doing that, your swing will actually get worse. That’s not an occasional thing, it’s a sure thing. It takes time for the newly improved mechanics to take hold and become natural. You only see the benefits after the hard work allows the new movement to become natural instead of forced.

    There are a number of things I don’t like about Eakins, but I’m not in the camp of burning him at the stake.

    I believe a lot of what we’re seeing with Hall (and a number of the other young players, including Gagner who is washing out on the whole “new swing” thing it seems) this year is in fact the decline that comes from trying to change old bad habits into new good ones.

    Time will tell I guess.

  47. sliderule says:

    Are there any statistics on goals against or dangerous shots after a offensive blue line turnover?

    Just by eye it seems it results in a more dangerous chance than a turnover in the offensive zone past the dots or even a turnover in your own corner.

  48. LMHF#1 says:

    Romulus Apotheosis: Make your case then.

    What has he said about Yak? What has he done with Yak that shows a disconnect between speech and action? What is it that he should be seeing in Yak’s game that he’s missing? And what information should he have taken from this video, etc. that he isn’t leading to his decisions?

    Sorry Rom, I thought this was all common knowledge from interviews Eakins did and subsequent commentary.

    Let’s just pick one spot so that it is more manageable. One of the things that Eakins said was (paraphrasing) that he looked at all the goals Yakupov scored (and other tape from last year) including positioning on the ice etc. From that, it should be pretty easy to extract where he might be most dangerous on the ice and work that into how you use him. It is abundantly clear on the powerplay for instance, that this has not been the case for the majority of the time this season. Is there a person out there who doesn’t have a fairly good idea of where the man needs to stand and set up for the shot? Why doesn’t Eakins see it?

    We all saw significant progress in Yakupov’s game under 1/2 a season with Krueger, correct? I see no reason that the step backward this year should have been so pronounced. The conclusion I draw is that despite his review of video, Eakins could not properly assess his player. The tools and enthusiasm are still the same from Yakupov. He hasn’t been injured. He hasn’t been adjusting to a new position or anything of the sort. Maybe it is that Eakins has his mold and that’s just the way it is going to be, but I’d hope he knows better than that.

    Here’s his 2013 goals btw: http://www.youtube.com/watch?v=hQF2ahK6laM Tell me you don’t see a hockey player who looks like he has a pretty specific and simple job to do in the offensive zone.

  49. LMHF#1 says:

    mc79hockey:
    If you can watch as much tape as Eakins says he did – study it – and then not be able to extract the bits you need to help your player be put in a position where he can be effective I seriously question your smarts when it comes to hockey. I guess it could be stubbornness as LT mentioned, but then why study? I don’t think the man knows what works out there and he’s yet to convince me otherwise.

    I’m going to go on a bit of a commenting jag here but I’ll answer this one first.I think Yak’s a bit of a different leopard than Hall.The issues are different.Even last year, I felt that there were times at 5v5 that Yak looked lost.To me, he’s struggled with making the adjustment to playing 5v5 in the NHL in terms of finding spots on the ice where he can succeed and how he needs to play to succeed (forecheck, move his feet etc.).I think he’s making strides there and he’ll be better for his time with Eakins this year.

    Wouldn’t you say that he was trending better in terms of 5 on 5 play at the end of last year and then looked like he’d walked into another dimension this year though? This isn’t “the sophomore slump”, it is structural. Krueger didn’t take too long to figure out what #64 could do.

  50. Romulus Apotheosis says:

    G Money,

    You encounter a very good dress maker. The results year on year are very good.

    A new boss comes to town and doesn’t like how this dress maker “looks” while making dresses.

    This boss SPECIFICALLY identifies something and asks for it to be changed. and it is

    This boss then SPECIFICALLY identifies that he now likes how this dress maker looks while making dresses.

    ————–
    This is what we are dealing with. It’s not a case of burning Eakins or MacT. It’s not a case of saying Hall is perfect and can’t improve.

    It’s a case of something working well (maybe with room to improve, no one contests that)

    decision makers deciding it isn’t working well and suggesting a change

    that change taking place

    the decision makers saying: “yes, this what we wanted to see”

    AND, the results unambiguously suggesting both that a change has happened and that the results have been negative.

    Finally, a decision maker saying: “maybe we mis-read the situation”

    ———
    it’s all pretty clear cut to me.

  51. stevezie says:

    oilersfan: I think that Krueger made a system based on the Oilers’ skill. Unfortunately, the only line that could implement the system was Hall-RNH-Eberle, which is why their corsis were positive most of the time even against good teams like St. Louis, Vancouver, etc. But when they weren’t on the ice, the other players didn’t have the skill to implement the system and those players had worse corsis.

    Is it a given that hockey teams will use the same system for every line? The challenges are obvious, but are there any teams who say, “Line 1 is playing Kruger’s system, the rest are running Eakins’”?

  52. VOR says:

    I am probably about to be bitch slapped as well but Dellow’s work is actually not set in context and suffers as a result. We need to know what variation in corsi looks likes from season to season in the population at large (all forwards in the NHL). Simply put, we may be falsely attributing random events as if they indicate real differences. I am fairly sure that corsi isn’t stable over a players careeer (considering the average player). Thus, we need to know how much of Hall’s corsi performance might be random variation and how much is down to differences in strategy or for that matter player or team mate effects.

    Let me be clear. In order to use corsi in this sort of analysis you would first need to know how much corsi typically fluctuates and also how much of that typical fluctuation is owing to “luck”. It is possible to determine this but the math is a bit above the level of the average ninth grader.

    Additionally, research in basketball has shown that sometimes the best strategy within a game is the wrong strategy over the long run.

    Thus, if the research in basketball applies, dumping it in might not give you a good result on average. However, there will be times where even though you have the possibility of controlled zone entry that dump and chase will give you a better result (in that particular case) than the over all better strategy of carrying the puck. Players need the situational awareness to know that such a situation may be what is unfolding in front of them and when that happens the last thing they need in their head is some sort of “you must do this”. I gather Eakins is saying he may have unintentionally embedded just exactly that sort of absolute in Taylor’s brain.

    Statisticians think in terms that are often expressed as simplified statements…controlled zone entry is better than dump and chase. Players and coaches have to deal with the nuance, the variation, and the situation which statistical analysis need not. This is why many “old timey” guys think advanced stats are a waste of time. I don’t agree but understand the criticism.

    Finally, where is the proof of the initial conditions? For example, where is the proof in this case that corsi is a valid tool to compare individual players season over season.

  53. stevezie says:

    Romulus Apotheosis,

    Yes, but just for the sake of defending the boss, their idea is that, while there may be an adjustment period, over time the results will be even better than they were in the first place. These bosses have been in the dress-making business for a long time, they’ve seen these changes work before.

    Off the top of my head… if Lindross had learned to move away from the techniques that brought him so much early success he might have struggled intially as he stopped steam-rollering defenceman, but he could have had more long-term success.

    Maybe. That’s an off the cuff devil’s advocation.

  54. thejonrmcleod says:

    mc79hockey,

    Has anyone ever done a study of the correlation between goals, shots on goal, shot attempts, zone time, and scoring changes? I know some of that has been done, but all together?

  55. G Money says:

    Romulus Apotheosis: Finally, a decision maker saying: “maybe we mis-read the situation”

    ———
    it’s all pretty clear cut to me.

    We differ on our interpretation. I read Eakins’ comments as “we overdid it”, not “we mis-read it”. This is a fundamentally different conclusion. If I’m misremembering Eakins’ discussion on the subject, please feel free to correct.

    Assuming that my memory is at least vaguely on track … In your exceedingly odd dressmaking analogy, the dressmakers boss has identified that the seams are simply not good enough, and has hammered away on the importance of getting the seams right. The seams did get better over time. Unfortunately, as a result, the overall quality of the dress has suffered. “Ooops, overdid the message about the seams.” So now it’s time to go back to paying attention to the overall dress – only now one hopes that the dressmaker is creating dresses with better seams than before.

    The dress fabric seems pretty clear cut to me…

  56. wood99 says:

    I understand the use of Corsi, but analytics sometimes are over done. I don’t need them to tell me Sam Gagner is an over rated player by many who read this excellent blo.Sometimes even in our new way of getting info on a player is over thought. Why have scouts if the new way of determining is analytics. I don’t think corsi can tell hockey sense, character and compete.It sometimes just requires old fashion hockey smarts and knowledge of basic hockey scouting by “supposedly”smart hockey men………I think this is where we have failed mmiserably the last 8 long years.

  57. stevezie says:

    mc79hockey,

    From your twitter: “I lent my copy of Fooled by Randomness to a girl I split up with, which irritates me to no end. There’s a great passage I’d like to cite”

    You can spend days watching film, compiling statistics, and then write a few thousand words, but you can’t stop by the library?

  58. Romulus Apotheosis says:

    LMHF#1: From that, it should be pretty easy to extract where he might be most dangerous on the ice and work that into how you use him. It is abundantly clear on the powerplay for instance, that this has not been the case for the majority of the time this season.

    I thought we were talking 5×5?

    If you mean the 5×4, I completely agree with you. Both MacT and Eakins have been loath to switch RNH out from the shooters spot that belongs to Yak (by rights). Staples talked to MacT about this extensively.

    And, recently

    Romulus’ Apotheosis ‏@RomulusNotNuma Feb 5
    Eakins on @OilersNow notes that problem with the PP is they have too many passers, yet he continues to underuse Yakupov. Hmmmm.

    Note that right around then he starting put Yak in the shooters spot on the PP.

    ————-
    back to 5×5.

    he has one of the most extreme zone starts in the league:

    http://www.behindthenet.ca/nhl_statistics.php?ds=30&s=63&f1=2013_s&f2=5v5&f4=C+LW+RW&f7=40-&c=0+1+3+5+4+6+7+8+13+14+29+30+32+33+34+45+46+63+67

    He’s exactly playing the same amount of time as last year (11.96 TOI/60 this year; 11.91 last year)

    Most common linemates this year?

    Gagner
    Perron
    RNH
    Eberle
    Hemsky

    Most common last year?

    Gagner
    Magnus
    Hemsky
    Horcoff
    Jones

    So, he’s getting considerably better linemates this year. WAY better.

    And, his qualcomp (via extraskater’s measures) is virtually identical to last year:

    http://www.extraskater.com/player/578/nail-yakupov

    ………

    So, what’s the argument here? Make your case. Convince me.

  59. mc79hockey says:

    We need to know what variation in corsi looks likes from season to season in the population at large (all forwards in the NHL).

    Why don’t you go and check how often elite players come with 43.5% Corsi%? I’ll wait.

  60. VOR says:

    mc79,

    If that is your best answer to valid criticism you really aren’t able to support your argument with facts. So right back at you, define elite player. Then go look at 100 of them over the last 7 years (all years they played). What is the variation?

  61. DeadmanWaking says:

    G Money:
    This is a good thing, not a bad thing. The learning process means developing situational awareness about when to carry it in and when to dump/chip it in. Overdoing it is part of the learning process.

    The opportunity to pursue helpful overcorrection is thin on the ground once the team is in legitimate contention. If now now, when?

    Performance is doing things that work.

    Growth is doing things that don’t work–yet.

    Frustration is doing things that always used to work, after your adversaries have all learned your tendencies and how to read your tells.

    This kind of analysis is useless unless it factors in shifting the burden. If playing behind Krueger’s do-it-all Hall is forcing the defensemen to master special tactics to accommodate the unique chaos Hall brings, that comes at the expense of habit-forming consistency when they are back out there in other contexts.

    The thing I liked most about Krueger is that he often spoke to the impossibility of a player having his head stuffed full of every possible thing the coach can explain and still playing well.

    Krueger worked by slowing increasing the heat and cooking in small batches. Eakins works by cranking the heat up to deep fat fry then dialing it back once the smoke is so thick the next armload of fresh ingredients starts to miss the pan.

    I think MacT came in with the attitude that you have to break some eggs to make an omelette. Mission accomplished. The question now is whether Eakins can put Hallsie back together again. I think he can, and that Hall will be better off for having explored some different looks.

    Make Your Own Products Obsolete

    The fear of ‘cannibalisation’ has prevented many a promising idea. And yet it seems clear that if you do not cannibalise your own product line with better, cheaper, faster, more effective or more appealing products then your competitors surely will.

    That’s the voice of Always Be Changing.

    Dave Moss: What’s your name?

    Blake: Fuck you. That’s my name. You know why, mister? You drove a Hyundai to get here. I drove an eighty-thousand dollar BMW. THAT’S my name. And your name is you’re wanting. You can’t play in the man’s game, you can’t close them – go home and tell your wife your troubles. Because only one thing counts in this life: Get them to sign on the line which is dotted. You hear me you fucking faggots? A-B-C. A-Always, B-Be, C-Closing. Always be closing. ALWAYS BE CLOSING. A-I-D-A. Attention, Interest, Decision, Action.

    And that’s the voice of Make Them Your Boxcars.

    It’s a fine art to rebalance a jet engine mid-flight when the blue paint is a Matryoshka orrery.

  62. G Money says:

    VOR: Finally, where is the proof of the initial conditions? For example, where is the proof in this case that corsi is a valid tool to compare individual players season over season.

    I pretty much agree with all of the rest of your post, and this afternoon if I have time (may or may not be a busy afternoon) I might even take a stab at calculating Corsi variation, just for gits and shiggles.

    That said, I don’t think you need “proof” that Corsi is a valid comparison tool. All you have to accept is that Corsi is at least a somewhat valid measure of the effectiveness of a forward (and personally, I think it is).

    This would allow me to conclude that on the basis of the fancy stats, Hall was more effective last year than this year.

    My eye concurs.

  63. Romulus Apotheosis says:

    G Money: We differ on our interpretation.I read Eakins’ comments as “we overdid it”, not “we mis-read it”. This is a fundamentally different conclusion.If I’m misremembering Eakins’ discussion on the subject, please feel free to correct.

    Assuming that my memory is at least vaguely on track … In your exceedingly odd dressmaking analogy, the dressmakers boss has identified that the seams are simply not good enough, and has hammered away on the importance of getting the seams right.The seams did get better over time.Unfortunately, as a result, the overall quality of the dress has suffered.“Ooops, overdid the message about the seams.”So now it’s time to go back to paying attention to the overall dress – only now one hopes that the dressmaker is creating dresses with better seams than before.

    The dress fabric seems pretty clear cut to me…

    Yep. We’re at a real differend here.

    Oddly it’s over a kind-degree (or quality-quantity if you prefer) which often lead to confusion and difficulty locking down positions.

    Not sure why my analogy is odd. But that doesn’t matter.

    What’s at issue here IMO is that “we over did it” and “we mis-read it” are the same thing. They’ve taken something like

    “Hall could probably improve defensively, in situational awareness and his read on zone entries”

    which, hey… that’s every player if you want to get right down to it… because, we accept that humans are never finished projects in relation to their flourishing.

    And… with that simple sentiment, they’ve tried to dramatically re-structure his game in one particular regard.

    Or, they’ve tried taken their quantitative tweaking so far it has resulted in a qualitative change.

  64. G Money says:

    DeadmanWaking: It’s a fine art to rebalance a jet engine mid-flight when the blue paint is a Matryoshka orrery.

    I’m not sure I get the latter half of your analogy, but the first half is a beautiful description of the act of coaching young Taylor Hall! (and Yak I would add)

  65. G Money says:

    Romulus Apotheosis: Not sure why my analogy is odd. But that doesn’t matter.

    Dressmaking is not widely used material (see what I did there) for analogies in hockey. Odd good, not odd bad!

    Romulus Apotheosis: Or, they’ve tried taken their quantitative tweaking so far it has resulted in a qualitative change.

    Very well could be.

  66. Lowetide says:

    VOR:
    mc79,

    If that is your best answer to valid criticism you really aren’t able to support your argument with facts. So right back at you, define elite player. Then go look at 100 of them over the last 7 years (all years they played). What is the variation?

    This doesn’t really help the conversation. a 7-year look would no doubt identify all of the elite players who were hurt and playing during specific times. We can know intuitively that someone who pushes the river has pushed the river.

  67. WeirsBeard says:

    Tyler,
    Is there a desirable ‘dump range percentage’? Carry ins generate better shots/chances, but aren’t always advisable/possible.

    I am interpreting from all your work that Hall’s problem is that they’ve over corrected, but am guessing that they wanted him to increase his usage of dumpins on certain situations, right?

    Another thought I have is if the Oil are dumping the puck into the corner, they aren’t going to win many battles, given the size and dirtiness they generally give up. Do they have to focus on giving better dumpins to open ice, so they can use their speed still, and not have to grind as much?

  68. Romulus Apotheosis says:

    G Money: Dressmaking is not widely used material (see what I did there) for analogies in hockey.Odd good, not odd bad!

    It just popped in there… not sure why. I don’t know any dress makers.

    G Money: Very well could be.

    the most important thing IMO is Eakins seems to have recognized something is off…

    A good test will be how Hall finishes the year. If he can correct this (and maybe, as you say, even improve Hall’s d-awareness along the way), he’ll basically win over a lot of people.

  69. steveb12344 says:

    Woodguy:
    steveb12344,

    I’m sure that the systems have factored into it, but I don’t know why most of you just completely discount the effect of linemates.

    Because he’s had the same line mates year over year for the most part:

    4 most common Forwards for Hall

    13/14 Hall total 718 5v5 min
    With:
    RNH 513min
    Eberle 383min
    Gagner 177min
    Hemsky 126min

    12/13 Hall total 651min
    With:
    Eberle 432min
    RNH 389min
    Hemsky 128min
    Gagner 114min

    The difference there is pretty much nothing.

    Its the fact that his results with the same players are off that allows for this type of analysis.

    If you are referring to 14, and 93. I’d say the trio has been split up quite a bit this year as Eakins has been running the blender to full effect. Also the D-men have been for sure different this year.

    Even if you say those are his regular linemates (disregarding D of course) Then according to L.T , as recent as yesterday I believe. 14, and 93 have both been a little behind expectations in their boxcars, while Hall seems to be doing well, near the top of pts/60 League-wide.

    So with his (this year I’d say arguably) most common linemates having off-years so far (once again, traditional boxcars) Coupled with the fact that the Oiler D-men have been either less experienced, less consistant, or both, than the last 2 years. I’d say that Hall’s individual Corsi numbers have likely been impacted far more by those around him, than by Hall himself.

    More simply put: If Hall’s boxcars are good, but his other 4 linemates are struggling, then I don’t think it’s a stretch to think that Hall’s possession numbers would suffer, as a result. It is afterall a team game. The players aren’t doing anything alone out there, and if they are. They will probably get benched for it (Yak)

  70. steveb12344 says:

    mc79hockey:
    Yeah, he’s got an incredible grasp on the grade 9 math skills it takes to figure this stuff out.


    What if it’s just a matter of his other 4 linemates on the ice dragging him down. For example if you put Sydney Crosby out there with 4 ECHL’ers his Corsi would be very bad. Would that mean that Crosby sucks? No, it means there are 5 guys out there playing as a team affecting each others individual possession numbers.

    IMO that is likely the problem with Hall. The guy is playing as well, or better than he ever has.

    The math isn’t the magic.It’s the ability to apply logic.Let’s take your criticism.

    In 2012-13, Hall played about 60% of his TOI with RNH.They had a 54.6% Corsi% together.This year, Hall has played about 66% of hsi TOI with RNH.They have a 44.2% Corsi% together.The defence is pretty similar this year to last year.You don’t need grade nine math to realize how silly your critique is given these facts.

    You guys get so bloody defensive when someone questions your analysis. All I’m saying is if the other 4 guys that go over the boards with Hall are not playing up to expectations. (ie, RNH being behind his R.E. this year, and the other 4 spots being mostly a revolving door of inexperience, and inconsistency. 14 has spent a lot of time away from Hall, and he has been behind expectations as well.

    I’m not saying your work is wrong, just that this is imo a very important item relating to this subject that is being overlooked in the exhausting pursuit of the explaination to why Hall’s Corsi numbers don’t match his boxcar production.

  71. jake70 says:

    Woodguy:
    G Money,

    These are situational decisions.

    Its dictated by the coach more than you think.

    Perron said that Hitchock told them to dump and chase on 2 on 2.

    I think he said 3 on 2….was during phone interview with Gregor on day of trade to Edmonton. Not that it matters that much, but does drive home the point a little more.

  72. Bag of Pucks says:

    Looking forward to the next part of Dellow’s analysis wherein he concludes that upon touching the puck, the best option for Hall is to immediately shoot it at the opposition net. That will get those pesky Corgi numbers moving in the right direction again.

    Btw, Dubnyk’s sv pct was down markedly this year. Running some data segmentation in my utility muffin research kitchen (method after the jump), I discovered Dubie also stopped less shots with his face this season. Ipso facto, Dubynk needs to get his face in front of more rubber from hereon. Shame Eakins couldn’t see that before it was too late (wail, moan, gnash of teeth).

  73. VOR says:

    G Money,

    I am a purist. I like my stats to actually be grounded in reality. So while I agree wholeheartedly that within a season corsi correlates with individual outcomes for forwards I don’t believe anyone has demonstrated it works from one year to the next. Since that is the very argument that underpins the Taylor Hall’s CF% is bad this year it is important to be able to put his CF% or corsi performance in context. It is also a crap load of work. I was merely pointing you are all once again assuming something that is unproven to be true.

    In general one of the more devastating criticisms I could offer of “advanced stats” is that we rush over fundamental questions and say as you just did that since it cooresponds to what our eye sees (or rather what we want to believe it sees) that we don’t need to do the hard scut work that true staitistical analysis of complex systems requires.

    We ado need to know how much corsi fluctuates and what causes the fluctuation from season to season. Consider Eric Stall, a good player on a team with mixed fortunes. We know his corsi on has moved quite a bit over the years 7.3, 0.29, 14.75, 5.52, 0.19, 4.89, .72. Presumably his CF% has also. Anze Kopitar (who pushes huge corsi) has also seen his corsi on jump by as much as 12.10 in a single year and over ten in another…and he was pushing a corsi on of something like 8.3 before it jumped ten points. Or Pavel Datysuk who has a difference in corsi on of 14.56 from his best to worst seasons. So elite players experience huge corsi on variation. Rick Nash goes from -4.80 to +12.56 as he is traded to the Rangers. I assume he didn’t just suddenly figure out how to outplay.

    How much of these players variation in corsi is luck, teammates, situational, or even statistical collection driven (is a chance always a chance, are their rink effects, game effects, defintional issues) is a very pertitent and valid question and lies at the heart of using corsi to comment on individual player performance which is what Tyler is doing above. Think of all the possible questions and the insights we might gather:

    -how important are team effects
    -how much of corsi is luck
    -is a plunging corsi more telling than when it rockets upwards
    -how many years of good corsi do you need to establish you are a true outplayer
    -how much should corsi drop (should it at all) as you go from playing softs to playing toughs
    -as we see here how much of corsi is driven by coaching strategy or perhaps how much impact can coaching have on corsi

    These sorts of questions never get addressed when we make huge assumptions that are convenient and don’t test them against reality with appropriate statistical tools.

  74. VOR says:

    LT,

    All due respect but Tyler was being a twit and I gave it back to him. I can’t think it is very helpful for any of us to believe that a 43% CF% means the same thing on the Oilers as it would mean on say LA and I doubt Tyler is dumb enough to believe that. He was responding to valid criticism with either very sloppy thinking or deliberate snarkiness. He could have challenged me with specific parameters. Truthfully, your commenter above said it for all of us who haven’t swallowed gallons of the koolaid…boy you guys really can’t handle criticism. Intellectually bullying is not ever going to advance the conversation here but you allow it all the time.

  75. steveb12344 says:

    VOR
    says:

    February 11, 2014 at 2:17 pm

    G Money,

    I am a purist. I like my stats to actually be grounded in reality. So while I agree wholeheartedly that within a season corsi correlates with individual outcomes for forwards I don’t believe anyone has demonstrated it works from one year to the next. Since that is the very argument that underpins the Taylor Hall’s CF% is bad this year it is important to be able to put his CF% or corsi performance in context. It is also a crap load of work. I was merely pointing you are all once again assuming something that is unproven to be true.

    In general one of the more devastating criticisms I could offer of “advanced stats” is that we rush over fundamental questions and say as you just did that since it cooresponds to what our eye sees (or rather what we want to believe it sees) that we don’t need to do the hard scut work that true staitistical analysis of complex systems requires.

    We ado need to know how much corsi fluctuates and what causes the fluctuation from season to season. Consider Eric Stall, a good player on a team with mixed fortunes. We know his corsi on has moved quite a bit over the years 7.3, 0.29, 14.75, 5.52, 0.19, 4.89, .72. Presumably his CF% has also. Anze Kopitar (who pushes huge corsi) has also seen his corsi on jump by as much as 12.10 in a single year and over ten in another…and he was pushing a corsi on of something like 8.3 before it jumped ten points. Or Pavel Datysuk who has a difference in corsi on of 14.56 from his best to worst seasons. So elite players experience huge corsi on variation. Rick Nash goes from -4.80 to +12.56 as he is traded to the Rangers. I assume he didn’t just suddenly figure out how to outplay.

    How much of these players variation in corsi is luck, teammates, situational, or even statistical collection driven (is a chance always a chance, are their rink effects, game effects, defintional issues) is a very pertitent and valid question and lies at the heart of using corsi to comment on individual player performance which is what Tyler is doing above. Think of all the possible questions and the insights we might gather:

    -how important are team effects
    -how much of corsi is luck
    -is a plunging corsi more telling than when it rockets upwards
    -how many years of good corsi do you need to establish you are a true outplayer
    -how much should corsi drop (should it at all) as you go from playing softs to playing toughs
    -as we see here how much of corsi is driven by coaching strategy or perhaps how much impact can coaching have on corsi

    These sorts of questions never get addressed when we make huge assumptions that are convenient and don’t test them against reality with appropriate statistical tools.

    Very well said. I agree with this entirely, and wish I had even a fraction of your ability to articulate. :)

  76. LMHF#1 says:

    Romulus Apotheosis: I thought we were talking 5×5?

    If you mean the 5×4, I completely agree with you. Both MacT and Eakins have been loath to switch RNH out from the shooters spot that belongs to Yak (by rights). Staples talked to MacT about this extensively.

    ————-
    back to 5×5.

    So, what’s the argument here? Make your case. Convince me.

    I was never limiting the argument to 5-on-5. Think you must have cross-wired it with the Hall discussion.

    My arguments wrt Yakupov don’t have to deal with linemates or zone starts, but what he is being told to do on the ice. I’m watching his positioning, his skating lines and what happens when he has the puck. These things don’t jump around and change for a player (especially a guy like Yakupov) unless there’s something he’s being told to do. The “agreement” that he and Eakins came to has something to do with this I imagine.

    I know some don’t think analysis of this side of things is relevant, but it is still how I break down the game. The stats are great context on top of that.

  77. VOR says:

    Lt,

    By the way, I wasn’t holding myself above that standard. I have frequently resorted to intellectual bullying. I just have come to realize that it isn’t useful, entertaining perhaps, but not useful. Bitch slapping just shows the slapper lacks true confidence in their ideas and capacities.

  78. VOR says:

    LT,

    So you are saying that intuition trumps statistics unless the statistics tell you what you want to believe?

  79. Bag of Pucks says:

    In all seriousness, Dellow’s current series on Hall is the very definition of starting with the conclusion and trying to find the data to support it – as opposed to letting the analysis reveal the valid hypotheses as they emerge. Hence language like “I’ve got a theory…etc.” He’ll argue this isn’t the case, but c’mon, you don’t start crunching the numbers without some endgame in mind. The very fact he focused on zone entries as the fulcrum for the decline in Hall’s play reveals the agenda.

    For me, Dellow is good when he’s using advanced stats to play amateur capologist and questioning contract value (e.g. Khabibulin, Eberle’s shooting %). When he’s trying to ascertain optimum coaching tactics without controlled data sets or accounting for massive variability amongst comparative sets, not so much.

  80. godot10 says:

    Surly:
    Sort out the D and it will help sort out these zone entry problems.How I dearly miss those Pronger passes up the ice.One the tape, caught in stride, allowed our forwards to keep defenders on their heels opposed to their toes.

    The D was worse the last two seasons than this year. The problem is a coach trying to play his system, instead of playing a system suited to his players strengths.

    Over time, a GM and coach can acquire players suited to play their preferred system, but forcing players to play a system that doesn’t play to the players strengths is moronic.

    A GM can use analytics in planning for trades and for next season’s acquisitions, but a coaching staff has to move faster than to wait 40 games, and lose a season before making corrections. You have to recognize your players strengths and weaknesses faster than that.

    That the coach wasn’t molding his system to his players strengths was pretty obvious from early on. The typical cocksure AHL know-it-all-I-am-smarter-than-everybody-else blunder.

  81. book¡je says:

    steveb12344: Yeah, he’s got an incredible grasp on the grade 9 math skills it takes to figure this stuff out.

    I think this was probably the thing that people got defensive about, not the questioning of the analysis.

  82. Bag of Pucks says:

    Godot10, apologies if I missed it, but what system of play are you advocating the Oilers play to optimize their current roster makeup?

  83. borisnikov says:

    2 cents
    When criticism starts with “Yeah, he’s got an incredible grasp on the grade 9 math skills it takes to figure this stuff out…” I don’t think the party criticized has any responsibility to not fire back with snark.

    Look, we’re all guilty of it from time to time but lets call a spade a spade and say there are more constructive ways to go about criticizing work. It seems fairly clear to me, and I am no intellect, that the groundwork behind Dellow’s assessment(s) has/have been pretty effectively laid by many, many individuals. Corsi and its applications are not new any more. Can we move past “What does it prove?” and just enjoy the development of the numbers and the analysis? If you don’t “believe” in it fine, but that doesn’t make it any less applicable to the game, especially in the depths of how Dellow presents it.

    I’d have to agree that this comment section can be a little intellectually intimidating at times.

  84. gvblackhawk says:

    VOR:
    Lt,

    By the way, I wasn’t holding myself above that standard. I have frequently resorted to intellectual bullying. I just have come to realize that it isn’t useful, entertaining perhaps, but not useful. Bitch slapping just shows the slapper lacks true confidence in their ideas and capacities.

    I just left myself a reminder note next to my computer: Never pick a fight with VOR.

  85. Lowetide says:

    VOR:
    LT,

    So you are saying that intuition trumps statistics unless the statistics tell you what you want to believe?

    No. I’m saying that if you see Jonathan Toews has a 43% Corsi next fall, he’s probably injured or hanging out with Hole.

  86. Romulus Apotheosis says:

    LMHF#1: I was never limiting the argument to 5-on-5. Think you must have cross-wired it with the Hall discussion.

    My arguments wrt Yakupov don’t have to deal with linemates or zone starts, but what he is being told to do on the ice. I’m watching his positioning, his skating lines and what happens when he has the puck. These things don’t jump around and change for a player (especially a guy like Yakupov) unless there’s something he’s being told to do. The “agreement” that he and Eakins came to has something to do with this I imagine.

    I know some don’t think analysis of this side of things is relevant, but it is still how I break down the game. The stats are great context on top of that.

    Well, I presumed 5×5 because that is were the bread is buttered.

    The thing about Yak is that as best I can recall he looked just a lost for long stretches last year too, he just happened to be luckier (much) last year:

    http://www.extraskater.com/player/578/nail-yakupov

    If we want to dig into things like his positioning, etc. year over year… we are going to have to actually get some data to test.

    There is no way we are going to be able to rely upon our flawed memories as to whether Yak is reading plays correctly, positioning himself well, etc. year over year.

    So, whether it’s something like Staples’ mistakes on scoring chances data, or some kind of video work, doesn’t matter so much… what matters here is that until we have something we can all look at, we aren’t going to get anywhere with this approach.

  87. Derek says:

    Bag of Pucks:
    Looking forward to the next part of Dellow’s analysis wherein he concludes that upon touching the puck, the best option for Hall is to immediately shoot it at the opposition net. That will get those pesky Corgi numbers moving in the right direction again.

    Btw, Dubnyk’s sv pct was down markedly this year. Running some data segmentation in my utility muffin research kitchen (method after the jump), I discovered Dubie also stopped less shots with his face this season. Ipso facto, Dubynk needs to get his face in front of more rubber from hereon. Shame Eakins couldn’t see that before it was too late (wail, moan, gnash of teeth).

    What is the point of posting this? You might get a thumbs up on ON I suppose. Why can’t we have nice things?

  88. denny33 says:

    Romulus Apotheosis,

    For this to be true, you’d have to show that Hall’s game needed work, ie., that he was getting poor results in the past.
    But, he wasn’t.
    *****************************************************
    He was. Period.

    Insert many, many, many, many coaches comments on Hall’s game

    Insert visuals of actually watching Taylor and the famous 1 on 3 moves….Hello Lars Ellar.

  89. smokinmonkey8 says:

    Is it not possible the Hall is not firing off as many wrist shots from just inside the blue line (scoring chance for?) and being a little more patient and trying to find real scoring chances by taking better shots or holding on for passes that may not be turning into scoring chances?? Which would alter his corsi? I dont fully understand the corsi so I may be way off with this…

  90. mc79hockey says:

    In all seriousness, Dellow’s current series on Hall is the very definition of starting with the conclusion and trying to find the data to support it – as opposed to letting the analysis reveal the valid hypotheses as they emerge. Hence language like “I’ve got a theory…etc.” He’ll argue this isn’t the case, but c’mon, you don’t start crunching the numbers without some endgame in mind. The very fact he focused on zone entries as the fulcrum for the decline in Hall’s play reveals the agenda.

    I’m not really sure how to respond to this. If I’m sitting on the couch in my house and I feel a draft, I check to see if my girlfriend has closed the goddamn door properly because she frequently doesn’t. I’ve learned from experience where to look when my feet are cold.

    There’s a growing body of research that points to zone entries as being of critical importance in shot rates. I’ve learned from those and they gave me an idea where to look.

    In general, conclusions that are arrived at without a hypothesis are the ones that people should be leerier of, IMO. There’s a great piece that I read about this somewhere, wish I could remember where. When people start with a data set and just look around in it to come up with things, they’re more prone to coming to some bizarre conclusion. Having a hypothesis that you’re explicitly checking leads to better work, IMO.

  91. Romulus Apotheosis says:

    denny33:
    Romulus Apotheosis,

    For this to be true, you’d have to show that Hall’s game needed work, ie., that he was getting poor results in the past.
    But, he wasn’t.
    *****************************************************
    He was. Period.

    Insert many, many, many, many coaches comments on Hall’s game

    Insert visuals of actually watching Taylor and the famous 1 on 3 moves….Hello Lars Ellar.

    Show your work!

  92. denny33 says:

    G Money,

    Excellent Post.

    Clear and concise…

  93. russ99 says:

    So the new arena is opening in Fall 2016. I wonder what our roster is gonna look like then?

    As of Capgeek, only Hall, Eberle, RNH, Hendricks (yikes) and the last year of Ference’s contract (double-yikes) will still be on the books.

    We probably can assume Nurse and this year’s top 3 pick (if we keep it) will probably be there too.

    Any guesses on the rest of the keepers?

    In our current group and prospects, I’d go with Yak, Chase, Marincin, Klefbom – and Gordon if he takes less in summer 2016 as a UFA to come back and center the 4th line.

    So that’s:

    Hall — RNH — Yak
    X — X — Ebs
    X — X — Chase
    Hendricks – Gordon? – X?

    Ekblad — Nurse
    X — Marincin
    Ference — Klefbom

    X — X

    Hopefully we can fill in some of the X’s this deadline/summer.

  94. Woodguy says:

    steveb12344: If you are referring to 14, and 93.I’d say the trio has been split up quite a bit this year as Eakins has been running the blender to full effect.Also the D-men have been for sure different this year.

    Even if you say those are his regular linemates (disregarding D of course)Then according to L.T , as recent as yesterday I believe.14, and 93 have both been a little behind expectations in their boxcars, while Hall seems to be doing well, near the top of pts/60 League-wide.

    So with his (this year I’d say arguably) most common linemateshaving off-years so far (once again, traditional boxcars)Coupled with the fact that the Oiler D-men have been either less experienced, less consistant, or both, than the last 2 years.I’d say that Hall’s individual Corsi numbers have likely been impacted far more by those around him, than by Hall himself.

    More simply put:If Hall’s boxcars are good, but his other 4 linemates are struggling, then I don’t think it’s a stretch to think that Hall’s possession numbers would suffer, as a result.It is afterall a team game.The players aren’t doing anything alone out there, and if they are.They will probably get benched for it (Yak)

    Two things:

    1) Those TOI numbers are not an opinion. The % of his time with his most common F line mates is essentially unchanged.

    2) I don’;t think you read Tyler;s posts. Your questions are answered in his posts.

    You’re trying to waive away Hall’s drop in CF% with a wave of the hand and blaming his linemates, when Tyler went into much, much deeper detail on where exactly the issues are.

  95. godot10 says:

    sliderule:
    Hall has been called out by team Canada coaches for turning puck over .

    Hall turning the puck over is a problem for Team Canada. Hall turning the puck over trying to create offense on a near NHL cellar dweller is not an urgent problem that has to be solved immediately. It is a problem that can be solved with a lot of video coaching over time on situattonal awareness and decision making.

    You want him to play instinctively (like Patrick Kane) but just make better decisions, but it is not something that has to be fixed overnight. The current coach is asking him to playing non-instinctively, which is a really dangerous thing to do, and that bad decision is now screaming out in the analytics. Eakins doesn’t want to do the hard work of coaching. (He probably wouldn’t have time to train for triathelons then. He goofed off all summer rather than study Oiler video.) He wants the simplicity of a rule, dump the puck in, rather than hard video work of teaching a player situational awareness and decision making.

    Did you listen to Gregor’s interview with Nelson about all the video work they did with Marincin, and now Klefbom in OKC, about situational decision making. That is coaching.

  96. wood99 says:

    I get the feeling if you don’t use corsi ,believe in Corsi your deemed that you know nothing about hockey.it bothers me that this new tool is now the end all, be all when evaluating players. It is one resource among many.the game is always evolving and so is the way we determine the value of a player. Analytics is the in thing right now.what will it be in ten years?

  97. russ99 says:

    godot10,

    Nice post.

    Also, if you ingrain into the head of a player that if he does things a certain way it will hurt the team, how is he going to get those needed reps in that situation to be able to make the right play instinctively.

    We’re still talking about a 22 year old kid, despite that his top-15 numbers make us think otherwise.

  98. Lowetide says:

    wood99:
    I get the feeling if you don’t use corsi ,believe in Corsi your deemed that you know nothing about hockey.it bothers me that this new tool is now the end all, be all when evaluating players. It is one resource among many.the game is always evolving and so is the way we determine the value of a player. Analytics is the in thing right now.what will it be in ten years?

    Corsi is a helpful and useful tool, but no one on this blog (or any other) would suggest it’s the be all. It’s like figuring out a puzzle, you need all the pieces. Believe me, Corsi is a newer piece and really adds to the view!

    But you still need the other things, boxcars, time on ice, observation.

  99. VOR says:

    LT,

    My response would be that you have just exhibted the exact same flawed thinking as Tyler. You’ve assumed what is not proven. Which sort of makes my point. You have no valid basis for saying CF% of 43 has statisitical significance for any player. The fact you are so convinced it must does not make it more or less true. It is a testable hypothesis. How much variation is their in underlying corsi over a career and what portion of that variation appears to be down to luck? That we are using CF% to dissect the play of Taylor Hall without having tested the assumption is at the core of my criticism. You saying if Jonathan Toews had similar numbers we could assume… simply increases my concerns, since without the fundamental research we couldn’t assume anything.

    I am actually with MC79 on his point about starting off to test an idea you believe is true. You have to start with a point of view or you will drown in a muddy, stinking pile of crap. Data smog, fog, whatever you want to call it will render you blind and foolish. I’d like to think if any of us found the data saying, “you are a pudding headed fool” we’d abandon that original point of view. Sadly, I doubt my own abilities in this regard and assume most people are similarly challenged. When we like an idea we really, really like it, and you will have to pry it from our cold dead hands.

    If more hockey sabremetrics were published in reputable journals we would see some auto-correction. Because, unlike all the other professional sports hockey culture has ended up with non-professionals trying to figure out which sabremetrics are and aren’t valid it is left up to the debate of lively minds in forums such as this to provide some checks and balances. Which is what I have been trying to do here today.

    The problem with contrarian views is they aren’t well accepted here and then once the proponent of that view leaves we miss him terribly (hi, DSF). So yeah, I get some of you think I am being a pain in the ass but I will continue to point that all hockey sabermetrics that are in the public domain rest on a foundation of sand (and lots of the sand is missing). It behooves those who say corsi (or fenwick or whatever) allows for this analysis to be done or this point to be made to prove that the foundational assumptions of their point of view are sound.

  100. denny33 says:

    godot10,

    Hall turning the puck over is a problem for Team Canada. Hall turning the puck over trying to create offense on a near NHL cellar dweller is not an urgent problem that has to be solved immediately. It is a problem that can be solved with a lot of video coaching over time on situattonal awareness and decision making.
    *************************************************************************************
    Turning the puck over is a problem for Team Canada, the Edmonton Oilers and my beer league team.

    A problem is a problem.

    In Defence of Eakins – lots of people have tried coach Taylor Hall the last few years….some kids are not as receptive as RNH to instruction.

    Against Eakins – I truly take him at his word when he said he did not watch one clip of the Oilers.
    People can tell.

  101. book¡je says:

    Lowetide: Corsi is a helpful and useful tool, but no one on this blog (or any other) would suggest it’s the be all. It’s like figuring out a puzzle, you need all the pieces. Believe me, Corsi is a newer piece and really adds to the view!

    But you still need the other things, boxcars, time on ice, observation.

    …onions, shoeboxes, butter…

  102. wood99 says:

    I agree it is a useful tool.The issue I have is to read this article and its comments is that there is so much weight put into corsi. I still believe in watching the game and formulating an opinion on what I see rather than make an opinion based on numbers. As we all know numbers get be very deceiving.

  103. Bag of Pucks says:

    mc79hockey:
    In all seriousness, Dellow’s current series on Hall is the very definition of starting with the conclusion and trying to find the data to support it – as opposed to letting the analysis reveal the valid hypotheses as they emerge. Hence language like “I’ve got a theory…etc.” He’ll argue this isn’t the case, but c’mon, you don’t start crunching the numbers without some endgame in mind. The very fact he focused on zone entries as the fulcrum for the decline in Hall’s play reveals the agenda.

    I’m not really sure how to respond to this.If I’m sitting on the couch in my house and I feel a draft, I check to see if my girlfriend has closed the goddamn door properly because she frequently doesn’t.I’ve learned from experience where to look when my feet are cold.

    There’s a growing body of research that points to zone entries as being of critical importance in shot rates.I’ve learned from those and they gave me an idea where to look.

    In general, conclusions that are arrived at without a hypothesis are the ones that people should be leerier of, IMO.There’s a great piece that I read about this somewhere, wish I could remember where.When people start with a data set and just look around in it to come up with things, they’re more prone to coming to some bizarre conclusion.Having a hypothesis that you’re explicitly checking leads to better work, IMO.

    Much as I appreciate the wisdom of your cold feet, I think valid analysis has to start first with consideration of the validity of all the potential datasets & variables that can and do impact this output, before diving headfirst into the data du jour to arrive at postulated conclusions.

    To be fair, I think this wont to leap to conclusions plagues a lot of the advanced stats rhetoric, and you’re far from the only one guilty of it.

    If the analysis ENDED with, ‘this is interesting, but there’s a ton of variables or other valid metrics I could have looked at and thus I’m hesitant to draw ANY conclusions from it,” it would probably pass muster as something more closely resembling scientific hypotheses. But instead, you editorialize and try to nebulously connect it to some tactical philosophy that Eakins’ has not expressed verbatim. So, you’re parsing data to suit your hypothesis and then further parsing isolated quotations to reinforce the tenuous conclusions your data reveals. Talk about a house of cards.

    It’s a convincing argument you make, but in many ways it looks to me like ‘lawyering’ masquerading as impartial and unbiased science.

  104. Captain Smarmy says:

    mc79hockey:
    I

    There’s a growing body of research that points to zone entries as being of critical importance in shot rates.I’ve learned from those and they gave me an idea where to look.

    In general, conclusions that are arrived at without a hypothesis are the ones that people should be leerier of, IMO.There’s a great piece that I read about this somewhere, wish I could remember where.When people start with a data set and just look around in it to come up with things, they’re more prone to coming to some bizarre conclusion.Having a hypothesis that you’re explicitly checking leads to better work, IMO.

    Oh sure dump the Grade 9 math and run to the Grade 9 science. ;)

  105. Romulus Apotheosis says:

    book¡je: …onions, shoeboxes, butter…

    butter good point.

  106. steveb12344 says:

    book¡je: I think this was probably the thing that people got defensive about, not the questioning of the analysis.

    I’m just saying that this isn’t exactly rocket-science. It wasn’t meant as a shot at Tyler himself. He’s clearly a very bright guy, and does some very good work.

    The thing I find disconcerting is the arrogance sometimes displayed by him and others when you question the relevance of these things and they automatically assume that you just don’t understand.

    You havn’t exactly reinvented the wheel here. Corsi is just a traditional stat (Shots on goal) with a twist. You also count the shots that were deemed to have been an attempt at a SOG, despite the fact that those “attempts” have 0% chance to actually be a goal.

    I agree there is some value in it when analyzing the overall team play, and to help suss out which systems and or personnel are working. However I draw the line short of declaring a player is struggling because of Corsi, despite the fact that said player is at or exceeding expectations with his boxcars, and near the top of the league in pts/60. All this while playing on a bottom-feeding team with of his 5-man unit on the ice being either inexperienced or struggling themselves.

  107. Lowetide says:

    VOR:
    LT,

    My response would be that you have just exhibted the exact same flawed thinking as Tyler. You’ve assumed what is not proven. Which sort of makes my point. You have no valid basis for saying CF% of 43 has statisitical significance for any player. The fact you are so convinced it must does not make it more or less true. It is a testable hypothesis. How much variation is their in underlying corsi over a career and what portion of that variation appears to be down to luck? That we are using CF% to dissect the play of Taylor Hall without having tested the assumption is at the core of my criticism. You saying if Jonathan Toews had similar numbers we could assume… simply increases my concerns, since without the fundamental research we couldn’t assume anything.

    I am actually with MC79 on his point about starting off to test an idea you believe is true. You have to start with a point of view or you will drown in a muddy, stinking pile of crap. Data smog, fog, whatever you want to call it will render you blind and foolish. I’d like to think if any of us found the data saying, “you are a pudding headed fool” we’d abandon that original point of view. Sadly, I doubt my own abilities in this regard and assume most people are similarly challenged. When we like an idea we really, really like it, and you will have to pry it from our cold dead hands.

    If more hockey sabremetrics were published in reputable journals we would see some auto-correction. Because, unlike all the other professional sports hockey culture has ended up with non-professionals trying to figure out which sabremetrics are and aren’t valid it is left up to the debate of lively minds in forums such as this to provide some checks and balances. Which is what I have been trying to do here today.

    The problem with contrarian views is they aren’t well accepted here and then once the proponent of that view leaves we miss him terribly (hi, DSF). So yeah, I get some of you think I am being a pain in the ass but I will continue to point that all hockey sabermetrics that are in the public domain rest on a foundation of sand (and lots of the sand is missing). It behooves those who say corsi (or fenwick or whatever) allows for this analysis to be done or this point to be made to prove that the foundational assumptions of their point of view are sound.

    You see Lowetide you’re wrong, and short sighted. Be open minded, like me!

  108. Hammers says:

    godot10: Hall turning the puck over is a problem for Team Canada.Hall turning the puck over trying to create offense on a near NHL cellar dweller is not an urgent problem that has to be solved immediately.It is a problem that can be solved with a lot of video coaching over time on situattonal awareness and decision making.

    You want him to play instinctively (like Patrick Kane) but just make better decisions, but it is not something that has to be fixed overnight. The current coach is asking him to playing non-instinctively, which is a really dangerous thing to do, and that bad decision is now screaming out in the analytics.Eakins doesn’t want to do the hard work of coaching.(He probably wouldn’t have time to train for triathelons then.He goofed off all summer rather than study Oiler video.) He wants the simplicity of a rule, dump the puck in, rather than hard video work of teaching a player situational awareness and decision making.

    Did you listen to Gregor’s interview with Nelson about all the video work they did with Marincin, and now Klefbom in OKC, about situational decision making.That is coaching.

    Tend to agree with your comment about ‘ that is coaching ” but there is no proof that Eakins doesn’t want to coach . If anything his old players from the Marlies give him ” cudos” for his work with them .Who knows for sure maybe there is more time to actually coach in the AHL than the NHL due to most of there games being Friday to Sunday .

  109. Eastern Oil says:

    wood99:
    I agree it is a useful tool.The issue I have is to read this article and its comments is that there is so much weight put into corsi. I still believe in watching the game and formulating an opinion on what I see rather than make an opinion based on numbers. As we all know numbers get be very deceiving.

    I am a supporter of stats like Corsi, Fenwick, etc. because it gives me an ability to either support what my eyes are seeing, or make me question them. Those, on both sides, that say their way is the only way to properly make decisions on the game are just as bad as the other. Advanced stats have their downfalls but it is a young field and needs developing.

    But, with all due respect, I don’t know you from a hole in the ground, why should I put any stake in what you are “seeing” in regards to the Oilers?

  110. G Money says:

    VOR: I am a purist. I like my stats to actually be grounded in reality. So while I agree wholeheartedly that within a season corsi correlates with individual outcomes for forwards I don’t believe anyone has demonstrated it works from one year to the next.

    I can definitely tell that you are a purist, and for that reason (being a mathhead myself) I read your posts with great interest.

    That said, I think that the arena of advanced stats as applied to hockey is in its infancy. On the one hand, that means working to advance the rigour of the science (as you imply) is certainly a good thing. On the other, I won’t throw out sensible and well-reasoned conclusions because the art and science are not yet mature. “Fermi questions” are as important (maybe moreso) to the advancement of science as rigour and purism.

    VOR:
    In general one of the more devastating criticisms I could offer of “advanced stats” is that we rush over fundamental questions and say as you just did that since it cooresponds to what our eye sees (or rather what we want to believe it sees) that we don’t need to do the hard scut work that true staitistical analysis of complex systems requires.

    That’s not really what I’m saying when I speak to what I see.

    Rather, given the early stage of development for the statistics and the fluid nature of hockey (which I believe makes the sport less amenable to the kind of statistics used in baseball), I do not “trust” a statistic unless I can validate it visually.

    The greatest value I find in advanced statistics is when I see something good or bad on the ice and the statistics do not concur. This almost inevitably leads to further investigation, deeper understanding of what is really happening on the ice, and a slight advancement towards “ground truth” (even as that may be nebulous).

    The second greatest value I find in the fancy stats is when they DO concur with what I see on the ice. This leads me to conclude that my visual assessment is more likely to be valid, or at least, less tainted by the myriad of cognitive biases (recency, bandwagon, expectation, confirmation, etc) that affect us all, purist and non-purist alike.

  111. borisnikov says:

    steveb12344,

    … I draw the line short of declaring a player is struggling because of Corsi, despite the fact that said player is at or exceeding expectations with his boxcars, and near the top of the league in pts/60…

    It’s not just Corsi.

    Hall is basically a unicorn. He has an IPP of ~105%. He is scoring at 1.06 PPG but his corsi, fenwick, shot differential and goal differential #s are in the toilet… and nothing is wrong?

  112. godot10 says:

    denny33:
    godot10,

    Hall turning the puck over is a problem for Team Canada. Hall turning the puck over trying to create offense on a near NHL cellar dweller is not an urgent problem that has to be solved immediately. It is a problem that can be solved with a lot of video coaching over time on situattonal awareness and decision making.
    *************************************************************************************
    Turning the puck over is a problem for Team Canada, the Edmonton Oilers and my beer league team.

    A problem is a problem.

    In Defence of Eakins – lots of people have tried coach Taylor Hall the last few years….some kids are not as receptive as RNH to instruction.

    Against Eakins – I truly take him at his word when he said he did not watch one clip of the Oilers.
    People can tell.

    Except the analytics are now showing that Eakins has made Hall and the Oilers worse with his “solution” to Hall’s turnover problem. But Eakins “solution” isn’t a solution. Instead of working with Hall all situational decision-making (i.e. actual coaching and doing his job), Eakins took the easy way out and said play my dump-it-in system where situational decision-making doesn’t matter. Play the system. Not learn how to improve your game by better situational decision-making.

    Note: by worse, the Oilers are now being outshot far worse with Hall on the ice now that he is turning the puck over less because he is playing Eakins system too well. What that shows is that Hall is coachable, but the coach is a bad coach, because he gave the wrong “solution” to the problem.

  113. commonfan14 says:

    So Hall’s shot attempts have declined sharply this year, even as his shots on goal per game has remained essentially unchanged?

    Is that not exceedingly odd?

  114. G Money says:

    Bag of Pucks,

    That’s actually a very generous assessment of how science works. The scientists I know (past self included) tend to be very partial and biased under a veneer of objectivity. It’s a human thing. That’s why peer review is actually a fundamental and critically important part of the modern scientific process.

    In this context, Tyler’s work is classically scientific. He made an observation. He formulated a hypothesis. He gathered data and tested his hypothesis. He drew a conclusion about his hypothesis, and then used it to formulate additional hypotheses.

    As a fellow scientist, you are free to point out the flaws in his reasoning, in his data, in his experimental design, etc. Depending on what you point out, this may invalidate his conclusions, put them in doubt, or merely widen the error bars around what he concludes.

    But to be honest, neither

    … valid analysis has to start first with consideration of the validity of all the potential datasets & variables that can and do impact this output, before diving headfirst into the data du jour to arrive at postulated conclusions.

    nor

    … there’s a ton of variables or other valid metrics I could have looked at and thus I’m hesitant to draw ANY conclusions from it

    actually fit the mold of what I believe to be valid scientific criticism.

    No experiment that I know of has ever looked at “all potential datasets and variables”, because if that was your criteria for validity, no experiment could or would ever have led to valid conclusions.

  115. mc79hockey says:

    VOR -

    The problem that I have with the self-appointed critics of this stuff is kind of two-fold. One, they frequently can’t be bothered to stay on top of the sorts of things that people do. There’s been a lot of writing over the years about Corsi% and how it repeats and you don’t seem to be familiar with it. I don’t think that anyone has an obligation to give a shit about this stuff but if you want to be the critic, there’s sort of an obligation to pay attention to what’s going on.

    What happens is that guys like me write something and we wade through the same nonsense objections from a different crowd of people who aren’t familiar with the background (but are happy to make declarations about foundations of sand) every goddamn time. It is tiresome.

    Second, there seems to be a view the analytics community is just a bunch of people saying “Holy shit, this is brilliant” to each other. I don’t really think that that’s the case. I was as critical of the QoC numbers that Gabe introduced and how people used them as anyone. Tulsky’s not afraid to be critical. The people who are interested in this stuff are generally interested in it because they like to figure out how things work.

    I think what you’re really after here is some sort of explanation as to what sort of change we see from year to year in hockey players anyway. To me, the most sensible way to look at that is by looking at the standard deviation of the change from year X to year Y. There’s a zillion different ways you can define year X and year Y.

    Say we look at players who played at least 1250 minutes in the two years preceding Year Y and at least 500 minutes in year Y. So in Hall’s case, he played 1525.82 minutes in 2011-13 and 781.27 minutes so far this year. Corsi% was 51.4% in years 1/2 and 43.5% in year 3. Difference of 7.9 points. Repeat for every player for whom we have at least 1250 minutes in years 1/2 and 500 minutes in year 3.

    That gives a list of 1703 guys. The standard deviation is about 3.3 points. So the fall in Hall’s Corsi rate is about 2.4 standard deviations. There’s more to this though. The data against which I’m comparing this is polluted to an extent by the fact that teams and situations change over time. The true change in a player’s ability is probably smaller. If I, for example, took every player in the NHL and numbered their shifts 1, 2, 1, 2 for a season, created two subsets and then compared subset 1 to subset 2, I’d expect that the standard deviation would be much smaller.

    Why do I think this? The list of big Corsi moves is dominated by guys who’ve had significant changes in their circumstances.

    I don’t think I’ve captured all the coaching/team changes in there but it’s obviously more than the number of coaching/team changes than the typical player goes through.

    All of which is to say: Hall’s big Corsi% slide is unusual even by looking at how much Corsi% tends to move around. It probably tends to move around less than it appears looking at this way because there are factors that I can’t avoid bringing into it in this analysis.

  116. mc79hockey says:

    My link got eaten in the previous post – you can see it here: http://www.mc79hockey.com/wp-content/uploads/Screen-Shot-2014-02-11-at-6.14.51-PM.png

  117. VOR says:

    LT,

    I get the impression that I have once again managed to offend someone by pointing out that I disagree with them and have good reasons for doing so. My apologies.

  118. PhrankLee says:

    Wow. Quite the Tuesday you guys have had. And it’s only 4:20 there! Guest appearance from Tyler, name dropping galore, successful deployment of the word “snarky” in several places. Awesome. When I read these threads and the articles LT linked to I became convinced we were no longer discussing hockey in this blog today. But awesome arguments. Being stuck away from home for these years I turn to this blog the most because you guys are pretty smart. Insecure and shockingly prone to taking and giving internet offence, but damn smart. This blog is pretty much one of a kind. Thanks, LT. Find 1G

  119. godot10 says:

    Hammers: Tend to agree with your comment about ‘ that is coaching ” but there is no proof that Eakins doesn’t want to coach . If anything his old players from the Marlies give him ” cudos” for his work with them .Who knows for sure maybe there is more time to actually coach in the AHL than the NHL due to most of there games being Friday to Sunday .

    Coaching an elite talent is more challenging than coaching an AHL schmuck. i.e. When you are coaching an elite player, you are coaching an outlier, someone who is probably more than an standard deviation away from the norm.

    And well, Hall has clearly demonstrated that he is coachable, because he has taken to Eakins coaching and become a worse player.

  120. Lowetide says:

    VOR:
    LT,

    I get the impression that I have once again managed to offend someone by pointing out that I disagree with them and have good reasons for doing so. My apologies.

    Sigh. I’ll get the cross and nails, you bring the hammer. Poor Vor.

  121. steveb12344 says:

    borisnikov:
    steveb12344,

    It’s not just Corsi.

    Hall is basically a unicorn. He has an IPP of ~105%. He is scoring at 1.06 PPG but his corsi, fenwick, shot differential and goal differential #s are in the toilet… and nothing is wrong?

    Hall is just one of 6 players on the ice affecting shot, and goal differential #s. There’s definitely a problem, I’m just not sure how much of it is directly attributable to Hall.

    It seems to me that he’s been doing his job. Maybe Corsi applies a disproportionate amount of blame on an individual, while his teammates, especially Defence have been clearly struggling.

  122. wood99 says:

    Eastern Oil,

    I agree they are all useful tools.To say that what I see is wrong or right.is a bit narrow minded.You read these blogs to get other peoples opinions and thoughts on players , teams etc……You than express your opinions and the debate continues. This is not to say one is right or wrong just different. We allsee the game our own way.that is what makes these debates fun.

  123. icecastles says:

    stevezie: Is it a given that hockey teams will use the same system for every line? The challenges are obvious, but are there any teams who say, “Line 1 is playing Kruger’s system, the rest are running Eakins’”?

    I suspect this was a rhetorical question, but I’ll respond anyway, in the hopes of securing myself a team Zisou toque.

    In his interview with Stauffer when he was first hired, Eakins was asked about what sort of system he would be employing. his response, paraphrasing, was that there would be multiple systems and that he found it baffling that some teams would run only one specific system. He pointed out that football players are required to study and learn numerous systems and numerous plays, and that while the correlation with hockey was far from exact, he saw no reason hockey players shouldn’t be able to do the same.

    My guess is that this also has a lot to do with why the team struggled so mightily early on and looked so confused. For a team composed of a mix of non-NHL players and “natural skill” players, any kind of advanced systems play just left their heads spinning and the learning curve was too steep for them to assimilate so much at once.

    Suspect that Eakins may have also confused the difference between learning different set plays and different systems, and the difficulty in assimilating one over the other into such a fluid and fast-moving game as NHL hockey.

  124. mc79hockey says:

    Consider Eric Stall, a good player on a team with mixed fortunes. We know his corsi on has moved quite a bit over the years 7.3, 0.29, 14.75, 5.52, 0.19, 4.89, .72. Presumably his CF% has also. Anze Kopitar (who pushes huge corsi) has also seen his corsi on jump by as much as 12.10 in a single year and over ten in another…and he was pushing a corsi on of something like 8.3 before it jumped ten points. Or Pavel Datysuk who has a difference in corsi on of 14.56 from his best to worst seasons. So elite players experience huge corsi on variation. Rick Nash goes from -4.80 to +12.56 as he is traded to the Rangers. I assume he didn’t just suddenly figure out how to outplay.

    Good choice not using Corsi%.

    E. Staal – 52.1, 55.9, 49.8, 51.6, 50, 50, 52.1
    Kopitar – 47, 54.2, 54, 53.9, 58.2, 61.3, 60.9
    Datsyuk – 63.3, 60.2, 57.4, 55.8, 59.7, 57.3, 55.3
    Nash – 54.3, 53.6, 48.6, 51.6, 47.1, 55.5, 54.3

    Those numbers are all from 07-08. Now – I’m not sure I see a lot of Hall sized implosions in there. Staal’s been within a pretty narrow range during the period in question, except for 08-09 when Carolina had a really strong team. Kopitar was on a terrible team in 07-08 and was very young. Younger than Hall is now. He’s gone up and up. Datsyuk has certain regressed over time – he was also starting from north of 63%, with one of the greatest defencemen of all time on his side and is now 35. Guys get old. Nash’s numbers seem to be better when he’s got a better team.

    I’m not sure what any of this proves though. There’s nobody good throwing up 43s in there. If Nash hadn’t stopped producing in 2009-10, maybe Hitchcock doesn’t get fired. The question of why Nash dropped five points of Corsi is probably an interesting one on its own – he certainly had more in him, as he’s shown since getting the hell out of CBJ.

    That’s all next level stuff though. None of this supports a case that what’s going on with Hall is a normal thing that just resolves on its own.

  125. stevezie says:

    icecastles,

    No, I was genuinely curious. I forgot about Eakins saying that.

    I think it was interesting that so many people said it was going to take the team awhile to “get” Eakins. It was pretty widely known he was going to be ambitious and he was going to be original, and that things might actually get worse before they got better while be broke them down to build them up.

    You could argue that he got too ambitious, or that he’s just an idiot, I guess, but my personal theory is the goaltending caused losses and losses hurt the buy-in which just retarded the whole process.

    Cap and speedo are in the mail.

  126. whiskeytangoeberle says:

    I just want to explain my position. I’m an objective guy, live and work, stuff like that. Not being too invested in advanced stats but with a good head on my shoulders I have this to add to the conversation. Stats are kinda funny sometimes, Losang Rampa in his novel Wisdom of the Ancients says: It’s difficult to explain to a three-dimensional world that which is an occurrence in a more multi-dimensional world. So can you measure confidence in milliliters but I’m sure you Tyler Dellow*, can measure elephants in the banannas they eat. Or let me put it this way. Would you like to do corsi for the battle of three hundred Spartans verse the whole Persian army. Who ate the puck anyways, not Xerxes or the guy his dad killed to become emperor. To put simply what so many are saying complexly if you consider yourself an intellegent person why have you read this far. Fuck off.

    *The idea not the person, I wish the person only the best. Seriously.

  127. WeirsBeard says:

    Wow are people ever ornery today. Maybe I am suffering from the after effects of a Steve Smith, but I swore the topic of the article was Hall and not Horcoff.

  128. Ribs says:

    I’ve learned from experience where to look when my feet are cold.

    This and a solid grasp of grade nine math?? I now pronounce you King of the Internets!! :)

    Seriously though, great stuff as usual.

    If a player gains the blue line and then chips a puck deep without any intention to play it to a teammate in a better position, it’s effectively the same thing as dumping the puck in. Possession has been abandoned. I’ve got a theory that they’re doing more of it this year, along with something else that I want to look at.

    I’d guess they’re doing this, too. Louie DeBrusk even has a name for it when they do it. I believe he calls it a “Chip and Charge”. The only problem with it is that a lot of the time they either forget the charge part (like it’s a normal dump-in) or are just not physically able to get the puck back from the defenders (either size or the defender’s speed being the issue).

    I think they’re trying to add this type of play to Hall’s game so that he has more options that give him a bit of a safety net and it is his decision making and execution that are not there yet. If he learns to pick his spots better, it will be of benefit to him. Will that happen, though? I hope so?

    Waiting for the lights to turn on – The Life of an Oilers Fan

  129. Bag of Pucks says:

    G Money, I’m not saying the scientist has to look at ALL the data, but I would think they certainly have to look at more than one limited dataset for their conclusions to attain any sort of validity in a peer review process?

    If I’m doing a study on materialistic attitudes of women in western society and my dataset is confined to a 10 block radius around Rodeo Drive, I think any conclusions I derive have to be taken with a grain of salt.

    To answer your point directly, where Tyler’s conclusions fall drastically short for me is when he attempts to correlate the findings with the very limited verbal offered up by Eakins re: Taylor’s play. There simply is not enough context in those limited quotations to arrive at the suppositions that Dellow makes.

    It probably doesn’t look like it in print today but I actually hold Tyler’s critical thinking skills in high regard, which is why I feel the need to challenge him on tenuous links like that put forth in today’s article. Far too many folks who are not as nuanced in their thinking will immediately run these findings up the ‘Eakins is ruining Hall, the team, western civilization as we know it, etc.’ flagpole.

    Folks like Dellow, Parkatti and of course, LT, must be cognizant of Ben Parker’s maxim, “With great power comes great responsibility’ ; )

  130. spoiler says:

    mc79hockey: That’s all next level stuff though. None of this supports a case that what’s going on with Hall is a normal thing that just resolves on its own.

    Which isn’t what VOR is saying. He’s not saying that your study doesn’t support your hypothesis. He’s saying that without the analysis that you have provided in your last two responses to him, we can’t know whether it does or not. And I’m guessing that because you have provided those two responses that you at least tacitly agreed.

    Placing Hall’s drop into the context of a greater than 2 sigma move against the general population means a great deal to me. 2+ Sigmas tells me we are likely not dealing with luck. Personally I feel that data point should have been included in the original study. Your idea of splitting the data too, from what I remember of stats, is a robust test.

    Thank you, Mudcrutch, for doing all the leg work on this issue and then a double scoop of thank you for taking VOR’s questions seriously and responding in kind. Well done.

  131. VOR says:

    mc79,

    You got it exactly. I am not sure you picked the correct statistical approach but we could argue arcana all day and I reserve the right to do so at some future date. With that caveat I’d say that your comment nicely addresses my concern and as you rightly point out it raises some very interesting question so thank you for doing that.

    Having re-read your article I would say that now I am wondering whether some players are more effected by a coaches zone entry strategies than others. You isolate on Hall but we can see there has been some but lesser deterioration in his two linemates games than his.

    Let me try asking you it directly. Assume you are right about the causes of Hall’s problems. Why is he suffering more than his linemates which looking at extraskater certainly seems to be the case. Is he dumping it in more than they are? Doing an even more pitiful job of retrieving it than they are? is the time he is on the ice without them worse than the time he is on the ice with them?

    Can you tell me exactly how each of RNH, Eberle and Hall are doing based on the same table you use in this article. It just shows me Hall on the ice not Hall in isolation and I find it odd that he would be so much worse off than his two most common linemates so it might be helpful to see them in isolation or even the other two guys on ice entry stats as you did for Hall.

    I just offer this up as a thought but one of the weaknesses I can see in your argument (and I am not trying to pick a fight but rather point out things you might one to base future posts on) is that if it is a coaching strategy/dump in problem it should be possible to see whether or not Hall’s CF% (which bounces around from game to game like crazy) is higher in the games where he carries the puck in more). I personally would find that an immensely compelling (not that you are required to care what I think) argument for too much dumping and chasing causing the corsi deterioration.

    Anyway, again thank you for listening and addressing my concerns.

  132. stevezie says:

    steveb12344,

    The fault, dear Brutus, is two fold:
    1) By eye, Hall was leading the pack. That pack included a banged up RNH. So in the very least you have to ask, “Why ain’t he leading it anymore?”
    2) By number Hall has had by far the most dramatic drop-off of that pack this year, so there are not that many reasons to suspect they are dragging him down. Unless you feel Hall is playing well enough to save their numbers, but not well enough to save his own, but this seems magical to me.

    This data set is amazing, and, like you, I haven’t really had time to read it yet. A few people have hinted at a question which focuses less on Taylor Hall and more and hockey and hockey stats in general: How is it possible for Hall’s corsi to get so much worse while his boxcars remain the same? It’s not shooting percentage.

    It’s probably explained in the article, but I’m too busy not understanding Aristotle to peruse that right now. Is anyone interested in committing academic fraud for me?

  133. VOR says:

    MC79,

    Again, not to be argumentative put you just made my point about variation.

    Eric Stall biggest CF% change in 1 year 6.1, Kopitar 7.2, Nash 8.4, Datysuk single season 4.2, career 8.0

    Taylor Hall this year over last 6.8.

    Thank you spoiler.

  134. mc79hockey says:

    Ah VOR, I didn’t really make your point for you but I think I see what’s going on here. Have a good one.

  135. sliderule says:

    I think this has been a great discussion.
    The points were well made though I have to admit that I got a little lost on some of them.
    I would say well done Mc79 ,VOR and others who contributed

  136. whiskeytangoeberle says:

    sliderule,

    Yeah no problem, although i missplelled the word banana, sorry bout that.

  137. teddyturnbuckle says:

    Here is the deal with Hall. The kid plays every shift like his team is down a goal with 30 seconds left in the game. He needs to realize that he doesn’t need to score every time he comes over the boards. Stop taking so many chances and play the high percentage plays. It’s not a failed shift if you dump it into your opponents end and go for a line change.

  138. oliveoilers says:

    Does anyone remember a couple of weeks back when Scot Howson was on the intermission on a Sportsnet Oilers game? He was asked by Spector (I know, I know, but bare with me) about advanced stats. His reply was that the Oilers were very interested, but at the moment they were trying to decide exactly what constitutes a useful stat. He gave the example of blocked shots, and the possible interpretation that they could be a bad thing. This is because to register a blocked shot your team;

    a) Doesn’t have possession
    b) Is in their defensive zone
    c) Being on the defensive back foot enough to have a shot attempted at your net
    d) Law of averages says at least some will get through

    This would lend credence to the Oilers current philosophy of possession hockey (different to the execution, I know). If you have the puck, they cannot shoot. I thought Scot laid it out very business like for the layman (me) without the patronisation that often accompanies what some people, who put some very hard work into and understand intricately, demonstrate from time to time. If the a guy who works in hockey professionally (Howson) can go on T.V. and admit that they’re not quite sure how this stat thing is supposed to work yet, then the non-stattees can be cut some slack.

    At the end of my working day, reading through this thread, it is clear that this a very provocative subject. I ask for understanding on all comments and hope they are treated with the respect they deserve. Understand that Tyler has put a lot of work into this, that it is a passion and he has clearly made a great emotional investment into something he and a growing portion of us believe in.

    And Tyler, if you read this, understand that you don’t yet understand everything and sometimes thing just happen. What the hell? If Hall’s stats aren’t what you suppose they should be, but he’s still scoring at the clip he is, then big deal. All I’m saying is; in this season of discontent, perhaps a more sensitive analysis of one of the bright spots and fan favourites would have been wiser? Perhaps not single out our best player and do an article about why he sucks so much?

    I see LT has a new article up. I hope we can all have a good debate about the pros and contras of Belov without resorting to snide comments and being hit with an intellectual purse. Peace out.

  139. rickithebear says:

    I finally know how to describe corsi.
    It is a moment in time event.
    Like penetration.
    It does not say how you got there!
    It does not say how you finished.
    A lot of you are not concerned about the finish.
    Alot of you do not believe in a sustainable shot rate!
    That says alot!

Leave a Reply

Want to join the discussion?
Feel free to contribute!

Leave a Reply

© Copyright - Lowetide.ca