Tuesday, September 16, 2008


Lies and the Coaches Who Tell Them

Preamble: this is Part II of what will hopefully be a comprehensive series; Part I with more preamble is here. I'd still like your questions, suggestions, and clarifications in the comments on the content; however, I hope it's clear enough that this is intended to be part of a progression. So Part I is devoted to the conclusion that straight offensive production is best identified by separating EV and PP, and expressing them as rates; factors that affect these rates, coach's usage of the player, etc. are to be addressed later. Today's edition is light on math/"interpretation", but appropriate (I think) as part of the series nonetheless, since the whole business is in large part devoted to distinguishing between myth and reality. Onward.


Part II.

If you've ever been in a hockey pool, you have probably considered drafting someone earlier than you might otherwise because you heard he was tabbed for Colorado's #1 PP unit, or that he's going to start the season as the centre on Jarome Iginla's line.

Likewise, all hockey fans know about the checking line, or perhaps, the shadow. The reputation of Steve Kasper was sterling during a stretch in the 80s for his success at shutting down Gretzky. More recently, we saw Jason Spezza and Dany Heatley stifled in the '07 Finals by Pahlsson, Rob Niedermayer, and Travis Moen.

Maybe those aren't the greatest examples, and we don't want to overstate anything, but I mention them to lead into a simple concept that needs to be accepted by those of us who wish to understand NHL hockey better: when it comes to evaluating a player on a game-by-game, month-by-month basis, you need to know who else is on the ice, because it matters. And NO ONE pays more heed to this than coaches -- they just don't like to talk about it much in a broad or comprehensive sense.

This is a bit of a tangent, but... it's important to understand that NHL coaches aren't simply hockey brains, devoted to strategy and teaching and motivation -- they're people, and successful ones at that. And successful people have the (wise) tendency to focus on the things they can control. Furthermore, they are, outwardly, inclined to stress the positive (the good things that can happen) and disinclined to dwell on the negative (the bad things that can happen, or all the things that have to go right in order to achieve success, or the inherent limitations of what they are working with).

Given the second, you can't expect a coach to be too frank about the relative strengths of his players, because that same frankness will betray the relative weaknesses, at least as he sees them. And given the first, it's completely understandable why they talk a lot about their players' effort (controllable on a day-to-day basis), preparation (ditto), and decision-making (ditto, to a certain extent), and very little about their relative abilities.

Fortunately, this is 2008. We have an accessible, easy-to-use "information superhighway"TM, so if we are interested in how a coach thinks each player is best used, we don't have to ask him -- we can just look it up.

The average fan might have the impression that the strategies of various NHL coaches span the spectrum from "likes to roll his lines" to "likes to match his lines"; this isn't backwards exactly, but it's wrong. The spectrum is more like from "tries to get the matchups he wants as much as possible" to "is ultra-aggressive about getting the matchups he wants nearly every time", where "getting the matchups he wants" also includes "avoiding the matchups he doesn't want". There are NO coaches who manage their own bench without regard to how the other guy is managing his.

There is a simply awesome web effort that is devoted to illustrating how coaches run their benches and how they use individual players: timeonice.com. It reminds me of the old joke about the watch: no, it doesn't tell time, you have to look at it. Which is to say, it doesn't calculate or interpret anything for you, but it shows you everything.

What Vic Ferrari has done is simplicity itself -- a little script that reads NHL.com's play-by-play reports (example) and produces two things:
  1. Matrices of EV ice time that show who played (1)with who and (2)against who (3)how much, all sortable
  2. Shift charts that graph who was on the ice and when powerplays and goals occurred, complete with draggable vertical red line to take the stress off your eyes. And as an astounding bonus, you can move the rows (i.e. the shift chart for a particular player) up or down to easily check how any player was used with or against any other player
The disadvantage limitation of Timeonice is that because it's all about the details, you can't just click around for a couple of minutes and get a broad perspective on which NHL players are seeing the toughest and weakest competition. But if you're trying to get a bead on a particular coach or player, it's invaluable.

I was just playing around with Game 20770: a Flames home game against the Sharks. Home games vs. the Sharks are especially useful for spot-checking, I think, because they have a clearly identifiable #1 forward (as opposed to, say, the Avalanche, where different coaches may have different opinions about who their most dangerous player/line is).

Looking at the H2H TOI, it's clear that Keenan wanted Tanguay/Conroy/Nolan, as well as Regehr/Sarich, out against Jumbo Joe (click on Thornton's name in Table 2 to sort). And clicking around on Table 3 (which Flames played with each other), you can see that Tanguay, Iginla, and Lombardi were never on the ice together (even two of them), so you can use those 3 guys as representatives of the three main lines that Keenan was using.

So then you go to the shift chart, move some rows around, and put Iginla, Tanguay, Lombardi, and Thornton next to each other. What do you see?
Anyway, I'm not trying to bore anyone with specific examples they don't care about; that was intended simply as a walk-through example.

Broad Strokes: What can/do coaches do to manage matchups?

  1. Duuhh: for faceoffs, the home coach can wait to see who the road coach is sending on before making his own decision
  2. In terms of avoiding a particular matchup, both coaches have the same tactic available to them: if they don't want Line X out against Line Y, then they send out Line X when Line Y is finishing (or just finished) a shift
  3. On-the-fly: every hockey fan is familiar with the "we're-tired dump-in", where one D-man goes back to pick up the puck and everyone else on the ice changes up. Tactically (for the coaches), this might as well be a faceoff. Most of us are less aware that coaches routinely change lines, D pairings, or even individual players on-the-fly in an attempt to gain a perceived advantage, even if it's only a momentary one:
Just generally, as Vic is fond of pointing out, it's not just the opposition that coaches concern themselves with, it's where the puck is, or where it's going. Tom Renney might rather eat his tie than put Jagr out for a D-zone faceoff against Crosby, but if the faceoff is next to Fleury, it's a possibility.

And one more note before this post really gets too damn long: on average, an NHL coach (i.e. a guy at the absolute peak of his profession) will make better decisions and have better judgment about how a particular player on a particular roster should be deployed. But even (especially?) when he's right, that doesn't necessarily mean that that's what that player is.


Post-amble: yeah, I know this is probably not enough Stats and too much Primer for a Stats Primer, but I think at least some of this stuff needs to be put together. I wanted to append a bit of a Guide to NHL Coaches -- who matches the hardest, who cares more about D matches than FWD, who's the most active re: pulling their scoring stars on and off to torque their O-zone time -- but I just don't know enough about enough guys to make it useful. Vic doesn't take requests :) but if anyone at all wants to drop some knowledge in the comments, that would be awesome.


Awesome! There is no doubt that Vic F has done a hell of a lot of good work to put his website together.

Matt - I know you have a strong opinion on Tanguay's value to the Flames last season. With your last sentence which you added a few bolded words before the post-amble, were you talking about Tanguay or were you satisfied with his role on the Flames?

What do you think of Sutter repeatedly saying through las season that 'we need(ed) more goals out of him'?

Naaah, I wasn't intending to refer to Tanguay or anyone specific with that comment.

Some players are used against soft comp, it doesn't necessarily mean they can't do it against the tough comp. Some guys score no matter who they're with, some don't. etc.

I'm not sure what Sutter thought Tanguay/Conroy/Nolan were doing last year. By his words and deeds, you'd have to guess that he thought none of them (except maybe Nolan, but it'd be close) earned their money.

If my office blocked timeonice.com, my Oct-June productivity would skyrocket. I like to sort all the players by line/pairing.

There are times when a coach is simply maximizing the icetime of good players, but it could be interpreted as sheltering. The Tanguay/Iginla example doesn't necessarily imply that Keenan wants Iginla on vs [anyone but Thornton]. It says he prefers Tanguay vs. Thornton, and after that he might just be maximizing Iginla's icetime.

I think coaches' choices are somewhat more limited than is implied in some of these matchup discussions. Players fatigue. If you rest forwards for at least two shifts for every shift played, a coach only has two lines to choose from when a shift ends - one of those typically being the 4th. After the key matchups, things usually seem to settle into place with the remaining best players available at the "top of the batting order."

Home or road, I find the most reliable way to pinpoint the most respected players for the evening is to identify who the shutdown D are on against. With D getting more icetime than forwards, it's impossible to get your star forward on vs. the 2nd/3rd pairing D with any regularity.

...but if anyone at all wants to drop some knowledge in the comments

Eastern bias here: In the games I scrutinized Ted Nolan appeared to balance out his lines, and thus care less about matchups than most. He had one rule and it was "Get Bergeron off the ice whenever an opposing forward with more than ten career goals jumps on." On the other side of the spectrum, Claude Julien might have been the most rigid.

Carbonneau prefers to use a RH centre against RH centres.

Awesome stuff Matt.

So much of the arguments in the Oilogosphere lately have been about stats, but I find a lot of the coaching theory being talked about here is just as dismissed as things like ESP/60 and whatnot.

This is the sorta stuff i've really only ever read on a few blogs, it's great you're actually bringing it out into the open and showing what it all means.

"Get Bergeron off the ice whenever an opposing forward with more than ten career goals jumps on."

That's not a Ted Nolan rule: that's common sense.

He had one rule and it was "Get Bergeron off the ice whenever an opposing forward with more than ten career goals jumps on." -- man that made me laugh.

Some good stuff there Jeff, thanks. I appreciate you underlining that there are "limitations", but I most certainly was not trying to imply the opposite. And as a straight indicator of who a coach fears most, yeah -- just look at who he uses his best defensive D-man against.

Vic had a comment somewhere recently that I'm pretty sure is true. Roughly: "the best players play against the best players a lot in this league, I think".

This is a brilliant post, Matt.

Coaches are dishonest or at the very least misleading in their dealings with the media, and for a variety of reasons, but they don't lie in how they use their players.

This is a very good primer and fits in well with the series.

You are a born educator, Matt.

One time I saw a documentary about the NFL in Dallas, this was long before the NFL was as big as it is today. Apparently there was a coach (I don't know whether he was a team coach, or a retired guy, maybe even a college coach, I dunno, I know little about football) anyhow he hosted a TV show that explained to the fans what had happened in the game.

It was in grainy black and white, and he had a table with some little wooden men, not chess pieces or checkers, more like those wooden things from the old crokinole game that lives somewhere in everyone's parent's basement. Or like big versions of some of the wooden pieces from really old Monopoly games.

And he had them set up, and looked to be moving them around to show what happened with different plays. Keeping it simple I suspect, but educating the audience nonetheless. And people generally don't like to learn stuff in their free time, so he must have been enthusiastic and engaging.

I doubt that he ever once looked up at the camera and said, "in any case, that should have been obvious unless you're a dumbass". And that's probably key.

In any case, the thrust of the film maker (with bias of his own, no doubt) was that this guy made fans appreciate the game more, made the decisions of the coaching staff seem more sensible, and ultimately increased the popularity of the game in that market.

And there is absolutely no doubt that the media in Pittsburgh and Buffalo are much, much more likely to comment on things that a.) aren't obvious to a casual fan. b.) really matter c.) actually happened in the game.

And you can see it in the commentary of much of the fanbase, and you can see it in the TV ratings.

I certainly don't have the natural patience. But this would seem to be the most effective route to take, should you be hell bent on taking it. Best of luck.

Even with this, you've started at second base methinks. Maybe NHL.com will let you have some game film to work with. Hell, Bettman promised game film access six or seven years ago, then he got sidetracked with the crunch-o-meter project (chips were being planted in all the jerseys and the crunchiness(?) of hits was to be displayed for viewers in real time. Perhaps they had technical difficulties, perhaps they realized how foolish it all was, in any case thank Christmas it was abandoned.

Sisu hockey! I thought you were dead.

Along the same vein ... someone should go around to kinger's place and make sure there isn't a bad smell coming from under the door.

And Carbonneau is into RH centres vs RH centres? Damn. What is it with these coaches who were former faceoff/own zone guys. They are off the hook. MacTavish has obsession with RH centres for RH side draws, and vice versa. Even in the offensive end of the rink. Intuitively, the marginal gain just doesn't seem worth the fuckery.

Is that why we see Smolinski subbed in for Latandresse in that graphic? (it's called G4, plus no Koivu, am I right in assuming round 1, game 4 of the playoffs?)

I mean as you say, it's pretty easy to get the D match you want now, even on the road. Because the home coach doesn't get to call all the shots, of course, he just gets first dibs. And almost always it will be a forward matchup at the top of his list, so he'll automatically give back the D matchup. Plus the vast majority of teams run with three forward lines during any stretch of 5v5 play nowadays, so with those two things considered ... you don't need to be a brain surgeon to get a hard match there.


Unless I have completely misunderstood Matt. That's not his point at all.

Coaches can accentuate the positive, but to point out the negatives of players is risky business, unless you have an ulterior motive, or unless they player has already been traded.

BTW sisu:

Thanks for the info, and I just don't watch enough of the East to have strong opinions, but that's some odd stuff.

Savard was a beast last season, bettering the wildest expectations of even his grandma. Lordy, look at the guys he played with and against, and how often he started in his own end. If we had Keenanesque "scoring chance +/- 45 seconds prior" then I bet he would be the bomb as well. Not Zetterberg good, but way better than Kovalchuk, Lecavalier, even Crosby or Ovechkin.

Still, running your 4th line out 3rd in the order, when two of Boston's three best forwards have yet to play? Damn, you just don't see that here, Julien's western equivalent would have Sedin, Gaborik, Hemsky, Smyth or Iginla on them straight away.

I don't know who benefits from a four line game here, even sporadically. But it's going to screw up the D matches big time.

The of course once the penalties get called it goes another way again.


On the whole Tanguay scenario. The simplist explanation, from the Gregory House school of reasoning:

What we know:
1.) Whenever Tanguay played with Iginla they went Power vs Power, and often started in their own end of the rink or with the puck headed south. More work and less glory.
2.) Whenever Iginla played with Huselius there was no way this was going to happen. He got the Jagr treatment.
3.) Keenan has been a power vs power coach in the past.
4.) Tanguay and Nolan are no longer there as safety blankets. Huselius is no longer there as the guy who needs to be kept away from tough on-ice situations whenever possible.


Reasonable assumption (if we pretend that the media may be misleading us, even a touch, when they tell us that NHL players are the best humans on Earth):

Iginla isn't the noble character that we've all been lead to believe, and this is the season that Keenan tries to break him.

It's a reasonable assumption, Vic, for the reasons you have stated, but I'm still going to guess the opposite (per my old take that Bertuzzi is a taller, less Swedish Huselius): it'll be Langkow that gets the Tanguay "treatment" this season.

I just don't see that (particularly on the road) he'll have Iginla and Langkow skating together. My guess is that Iginla gets the comparatively softer TOI (and the consequent challenge of dragging Bert along at EV), while Langkow and 2 other forwards TBA get the brutal minutes. (Moss? Bourque? Lombardi or Conroy converted to wing? Dunno.)

And unless I'M misunderstanding something, you and Jonathan are both right about Coaches & Lies: there is simply no upside to verbally acknowledging that (say) Player X is being sheltered from the real badasses, even if it couldn't be more obvious once you look at the reality of how he's being used.

Oh, and thanks Vic, but something tells me that when I get to Corsi & possession, my gift for edumacating is going to be less admired.

On the subject of the title of your post, I think that the guy who talks nonsense for no reason at all is Ron Wilson. At least at times.

I listened to quite a few of his podcasts in 05/06, and the guy is just playing a lot of the time. He contradicts himself constantly.

My favourite though, was when he was asked why his team was so good in the second periods (They had outscored the opposition by a wide margin in the second period that year iirc), their best period by far.

He answered, to the best of my memory:

"Coincidence. Look at last year ... it's a credit to our coaching staff that they analyze the game after the first period and are able to make adjustments"

Well the contradiction is obvious. Choose the answer you'd prefer, I suppose, but they can't both possibly have been right.

I did check the previous year's record at the time, NHL.com list the goals out by period. The Sharks had the opposite thing going on in 03/04; great in the first period, but in the second .. not so much.

Good of him to tell us how to check the theory, I suppose. And damn strange as well.

I just don't know how to take the guy, and it's just too much effort to check up on everything he says. Still, an interesting cat.

Yeah, after I'd hit the publish button I realized that "Huselius is no longer there as the guy who needs to be kept away from tough on-ice situations whenever possible." shouldn't have been included.

Goading aside, though, I don't think that it's necessarily safe to bet that Iginla is the most righteous man in the NHL.

Matt said:
something tells me that when I get to Corsi & possession, my gift for edumacating is going to be less admired.

Nothing of the sort, so long as it's rational. Reasonable arguments to the contrary are good things.

Vic - take that "info" with a grain of salt. Those are just some general impressions I was left with after gawking at those charts all last Winter. Sometimes I can make sense of them, usually not. I thought the RH/LH centre depending on the corner of the rink was SOP (it could have been that situation with Smolinski/Latendresse in the example). The last few coaches in MTL have worked that way, and it seems like they're shopping for a RH guy every summer.

Boston's 4th line - all series they spanked the Habs 4th line, which consisted of a rookie agitator and two converted defensemen. In Dandenault's case converted/reverted, poor guy. Maybe by this time (g4, as you guessed) Julien realized Carbo was sticking with the 4vs4 match and figured the margin was better for the Bs than the 2vs2 matchup? Sometimes I wonder about Carbonneau.

jeff j

I'm sure that you're right with most of it. Personally, I can't make much sense of them unless I am familiar with the players and the coaches, so I'm pretty lost on the east.

Islanders and Rangers, I catch a few games of them. And Tampa in the past as well, though not so much last year.

Renney builds the whole game plan around Jagr, has since the lockout. So for all the ice time that he gets, he rarely starts Jaromir at the beginning of a period, he'll wait for the situation to be favourable (direction of play, offensive zone draw, weaker forwards, bottom D pair out) before shifting Jagr over the boards. He STILL points the finger at Renney when he or the team slumps, though. Still, he's better than he used to be, and his shifts finish in the right end of the rink most of the time now, that lockout year spent in Omsk did him a world of good methinks. His shifts used to end on tired legs with him trying to beat one last guy one on one, usually at a blue line.

Nolan preaches hard, short shifts, and keeping all four lines in the game, but he still managed to get winger Trent Hunter out for 109 more defensive zone draws than ones in the good end. And Comrie out for two more o-zone than d-zone ones. That can't be coincidence.

And you can't play that sort of game with a bunch of young players on the roster. They'd end up out there against high end opposition too much and it would end badly a lot. I like Nolan as a coach, but he would be a better fit with a physical team that has a veteran roster, I think.

Post a Comment

<< Home

This page is powered by Blogger. Isn't yours?