The trials and tribulations of Glen Gulutzan

[belsarius]

Why is the focus so much on the offense? The Flames are actually higher in GF/game this year than last. It's the GA that's taken a hit. Shouldn't the focus be on why has the team been worse defensively with a better goalie and adding a top 4 defensive d-man.

[chedder]

Came across this si article from when Glen was hired.

http://amp.si.com/nhl/2016/06/17/cal...ew-head-coach?

From the article. Interesting:

One of the knocks against him in Dallas was his in-game decision making. He struggled to read and react as the game unfolded in front of him and that eventually cost him the room. Has he addressed that glaring issue sufficiently during his years as an assistant? And has he learned to build a bridge with veterans? His success in Calgary will rely heavily on the buy-in of players like Mark Giordano, Matt Stajan and Michael Frolik. Can he get them onside?

[Bingo]

Originally Posted by DeluxeMoustache

Depends.

I would rather see a cross crease tap in than a forward stuffing the puck in to a well positioned goalie pad. Shot location does not work as a proxy for quality, especially without further contextual information.

You saw the Rangers game and in your own write up acknowledged that the Rangers had more scoring chances, despite the Flames racking up a large number of shots.

The Rangers also had the very same stats to support that they had the scoring chances.

I watch all the games.

Most of the time, say 90% of the time I'm on PVR because of all my kid's hockey. Because of that I can't look at the stats during the game and have to wait until after or give away the out come (we sell most of our tickets)

This does one interesting thing for me ... it has me all eye test, followed by a look at the stats after the game. As a result I form an eye opinion as I watch and can't have the stats between periods to sway what I think I'm seeing.

The Dallas game had Calgary dominate in the last 60% of the game and generate a lot of chances. Stats supported it.

The Flames had the share of the stats in the Colordao game, but my eye test said they weren't all that good in the first, terrible in the second, and then pushed hard and ran up stats in the third.

The Rangers game met the eye test perfectly. They outplayed the Rangers badly in the first period, but panicked when the Rangers made it 2-1 and gave up too many chances. Stats said the exact same thing.

So I don't just stare at a website on game nights guys, I'm very much an eye test person. The stats are good to verify what you're seeing, and this year I see a Flame's team that outplays the opposition more often than not but just can't finish.

That's what I see.

I'm not just leaning on stats to make a case.

Just my opinion though.

[Bingo]

Originally Posted by EldrickOnIce

^ I still find it incredibly baffling that people don't see it exactly as simply as this.

If you out shoot and out chance the opposition, particularly with chances from the prime scoring area, you are more likely to be successful.

Does this guarantee more wins? Of course not. But you are playing 'the right way'.
And if you don't have success generating these type of numbers the Flames consistently have, it basically means 1 of 2 things.
1. You are snake bit. Unlucky. Results will even out over time.
2. You simply lack talent. You generate the chances but lack the skill to convert them.

I understand that's a simplistic response, and lots of factors can contribute to both 'luck' and 'skill'.

Perfect ... and exactly how I see it.

I don't profess there is any magic statistic that correlates perfectly. But if you get the puck in high danger areas more often than you give up you're on the right track.

[Bingo]

Originally Posted by Red

You are correct if we assume that all chances are equal. But as we know, if you are in traffic or your backhand etc, you are less likely to get your best shot off.

Eye test tells me that the flames dont get many point blank, rebound etc, really prime chances. Other teams somehow make it look easier.

I won't say you're wrong.

You could be right.

But the second you step out of the objective and into the subjective, for me the analysis dies.

An independent source counts shot attempts in high danger areas and provides data that web sites correlate into splits by each hockey team. The Flames are ranked 4th in this category five on five.

It's certainly possible that they are less dangerous in theirs and more dangerous in what they give up, but that's a subjective view that strays from simple stats.

I'm very hesitant to do that for or against the team.

Staples in Edmonton does his own scoring chances for the Oilers. They are pretty much always higher for Edmonton and lower for the opposition. It's a joke.

[Bingo]

Originally Posted by EldrickOnIce

Since the wheels fell off (aka the Smith injury), the opposition has not been winning because they have generated more high quality chances. Flames goaltending has not been good enough since then. This is somewhat a subjective opinion, but save% numbers support that.

I think with Smith in goal, the Flames get at least 4 points in the last 3, instead of zero - and we aren't having this discussion right now.

Since Smith has gone down the Flames have averaged 8.4 five on five chances against in ten games.

Previous ten game segment was 8.6

One before that was 8.2

[mrdonkey]

Originally Posted by Bingo

I won't say you're wrong.

You could be right.

But the second you step out of the objective and into the subjective, for me the analysis dies.

There is always subjectivity involved in analysis. Even very complex algorithms and machine learning techniques require human input to make sense of the data. So I definitely disagree with this, if only on the basis that any 'purely objective' analysis will be rudimentary and not very interesting.

An independent source counts shot attempts in high danger areas and provides data that web sites correlate into splits by each hockey team. The Flames are ranked 4th in this category five on five.

It's certainly possible that they are less dangerous in theirs and more dangerous in what they give up, but that's a subjective view that strays from simple stats.

I'm very hesitant to do that for or against the team.

This would also not be a hard thing to quantify if one were so inclined. How many of those 'high danger' shot attempts from the Flames get blocked? How many go wide? How many other players occupied that high danger zone at the time the shot was made? This is all data one could look at to put an objective lens on a subjective claim.

Staples in Edmonton does his own scoring chances for the Oilers. They are pretty much always higher for Edmonton and lower for the opposition. It's a joke.

Interpreting data to test a hypothesis and reach a conclusion, versus fudging the numbers to try and fit your pre-conceived conclusion, are both matters of subjectivity but I hope you can appreciate how they are two very different things. That difference (among other things) is what makes Staples a hack.

[Bingo]

Originally Posted by mrdonkey

There is always subjectivity involved in analysis. Even very complex algorithms and machine learning techniques require human input to make sense of the data. So I definitely disagree with this, if only on the basis that any 'purely objective' analysis will be rudimentary and not very interesting.



This would also not be a hard thing to quantify if one were so inclined. How many of those 'high danger' shot attempts from the Flames get blocked? How many go wide? How many other players occupied that high danger zone at the time the shot was made? This is all data one could look at to put an objective lens on a subjective hypothesis.



Interpreting data to test a hypothesis and reach a conclusion, versus fudging the numbers to try and fit your pre-conceived conclusion, are both matters of subjectivity but I hope you can appreciate how they are two very different things. That difference (among other things) is what makes Staples a hack.

Right now they have counters that aren't employed by the team, are fans of the team, or media covering the team.

That's a good place to start in finding objectivity.

And I'm all for a new wave of counts that break scoring chances into degrees of danger for sure.

I'm not avoiding data that disproves what I think. There just isn't any.

[mrdonkey]

Originally Posted by Bingo

Right now they have counters that aren't employed by the team, are fans of the team, or media covering the team.

That's a good place to start in finding objectivity.

And I'm all for a new wave of counts that break scoring chances into degrees of danger for sure.

I'm not avoiding data that disproves what I think. There just isn't any.

Fair enough, I can appreciate that if the data don't exist then we can't say one way or the other. But we can certainly do our best with what we have to try and make sense of why the underlying numbers don't agree with the outcomes.

I'm not necessarily against employing a little subjectivity to try and prove a point (this is after all a sporting activity, not science), but I am against reductionist thinking as a means to cut subjectivity out of the picture. I think simplifying the problem down to 'the Flames are playing the right way but are not getting the bounces' falls dangerously close to this line of reasoning.

I suspect there is a lot more that goes into playing 'the right way' than good possession numbers and high danger chances. And at the very least, this kind of hockey is not entertaining to watch so it's debatable whether 'the right way' is really the 'right' way as a medium for what is at its core a spectacle for fans. Subjectively, I would rather drink paint than sit through an entire year of Glen Gulutzan hockey, even if improvements are made in the lineup that translate to slightly more wins.

[Red]

Originally Posted by Bingo

I won't say you're wrong.

You could be right.

But the second you step out of the objective and into the subjective, for me the analysis dies.

An independent source counts shot attempts in high danger areas and provides data that web sites correlate into splits by each hockey team. The Flames are ranked 4th in this category five on five.

It's certainly possible that they are less dangerous in theirs and more dangerous in what they give up, but that's a subjective view that strays from simple stats.

I'm very hesitant to do that for or against the team.

Staples in Edmonton does his own scoring chances for the Oilers. They are pretty much always higher for Edmonton and lower for the opposition. It's a joke.

If I understand this correctly, these sources aren't necessarily watching the actual games, they use NHL.com's heat maps to form their conclusions.
If that's true, how reliable can they be?

I'd love to see a few NHL coaches sit down and break down all scoring chances in a given game as recorded by those formulas. I am willing to bet that these numbers would be under a lot of scrutiny.

[Strange Brew]

Originally Posted by EldrickOnIce

^ I still find it incredibly baffling that people don't see it exactly as simply as this.

If you out shoot and out chance the opposition, particularly with chances from the prime scoring area, you are more likely to be successful.

Does this guarantee more wins? Of course not. But you are playing 'the right way'.
And if you don't have success generating these type of numbers the Flames consistently have, it basically means 1 of 2 things.
1. You are snake bit. Unlucky. Results will even out over time.
2. You simply lack talent. You generate the chances but lack the skill to convert them.

I understand that's a simplistic response, and lots of factors can contribute to both 'luck' and 'skill'.

This is so accurate it's almost indisputable. But in addition to the 2 factors you described, I think equally important is;

1) Goaltending. Frankly it's 50% of the equation
2) Timing. I guess this is the same as score effected, but to me its an undeniable fact that winning the scoring chance battle when you're losing by 3 goals is simply not the same as winning the battle in a tie game in the third period. Teams simply play differently in those circumstances.
3) Obviously doesn't incorporate special teams

I agree these stats are useful. But to say Flames play well when you don't factor in goaltending, special teams and clutch play isn't saying much.

That's like saying the Jays are a good baseball team when you don't look at pitching, defense and hitting with RISP.

Am I off base here?

[blender]

More good stuff, guys.
One thing that springs to mind is the old hockey adage, "good teams find ways to win."
Hockey is full of cliches that are the antithesis of advanced stats but describe the game well despite their inaccuracy.
I don't think you are off base, Strange Brew, because what you are describing is the chaos of hockey. The good (more accurately, better) teams won't win every game, but when the chips are down, they find a way to win more than an even share.

[GullFoss]

Originally Posted by Regorium

For the discussion, a high danger shot attempt is any shot attempt (shot on goal, missed net, blocked shot) within the following area:

This is exactly why the flames lose while having more shot attempts in a high danger area.

Take seven examples of opportunities of shots in high danger areas:

1) three-on-one
2) breakaway
3) two-on-one
4) Three-on-two
5) four-on-three
6) shots of the cycle
7) rebounds off the cycle.
8) rebounds off an odd-man rush

You can't possibly believe that all of these opportunities are equal scoring chances. But saying "the flames had 15 high danger scoring opportunities vs 9 for the opposition" assumes exactly that. It ignores that all 15 of the flames chances game might come on 3-on-5 cycles (the flames dont activate their defensemen) where the shots we're blocked or the angle was narrowed for the goalie to make an easy save. It ignores that the opposition might have had 9 breakaways.

If the flames get 15 high danger scoring chances, each with a 10% chance of scoring. And the opposition gets 9 scoring chances, each with a 50% chance of scoring. What's the end result in the game vs what do the advanced stats say should happen?

This has been my argument all year long in the game threads. The flames often have more opportunities to score, but they're inferior opportunities. Why? Its the system. The D-to-D passes and east-to-west hockey reduce the liklihood of generating scoring opportunities on the transition. And its the scoring chances that come from the transition that hasve the higher probability chances - the 2-on-1s the 3-on-1s the breakaways.

[Bingo]

Originally Posted by Red

If I understand this correctly, these sources aren't necessarily watching the actual games, they use NHL.com's heat maps to form their conclusions.
If that's true, how reliable can they be?

I'd love to see a few NHL coaches sit down and break down all scoring chances in a given game as recorded by those formulas. I am willing to bet that these numbers would be under a lot of scrutiny.

No they watch.

The data they compile become the heat maps

[Red]

Originally Posted by Bingo

No they watch.

The data they compile become the heat maps

Oh ok, I thought you said yesterday that the stats websites you follow use nhl.com shot trackers and label them accordingly by where the shots came from.

My bad.

[djsFlames]

Originally Posted by chedder

Came across this si article from when Glen was hired.

http://amp.si.com/nhl/2016/06/17/cal...ew-head-coach?

From the article. Interesting:

One of the knocks against him in Dallas was his in-game decision making. He struggled to read and react as the game unfolded in front of him and that eventually cost him the room. Has he addressed that glaring issue sufficiently during his years as an assistant? And has he learned to build a bridge with veterans? His success in Calgary will rely heavily on the buy-in of players like Mark Giordano, Matt Stajan and Michael Frolik. Can he get them onside?

Spot on.

Too bad Treliving and co. didn't pick up on this, or didn't think it would impact results. We've had some so-so coaches before, but this is the first time I've witnessed one that appeared to be this out of touch with the happenings of the game as it pertains to coaching tactics. Sure he's started double shifting better players, and shortening the bench at times of late, but where was that the rest of the year? In many losses we had to watch players that contribute little to no offense skate full shifts late in games we were chasing and in reach of tying. GG appears to have had the same shelf life he did in Dallas. Getting the distinct feeling the last few games of diminishing belief among the players that this system can lead to the desired results.

[DeluxeMoustache]

Originally Posted by Bingo

The Rangers also had the very same stats to support that they had the scoring chances.

I watch all the games.

Most of the time, say 90% of the time I'm on PVR because of all my kid's hockey. Because of that I can't look at the stats during the game and have to wait until after or give away the out come (we sell most of our tickets)

This does one interesting thing for me ... it has me all eye test, followed by a look at the stats after the game. As a result I form an eye opinion as I watch and can't have the stats between periods to sway what I think I'm seeing.

The Dallas game had Calgary dominate in the last 60% of the game and generate a lot of chances. Stats supported it.

The Flames had the share of the stats in the Colordao game, but my eye test said they weren't all that good in the first, terrible in the second, and then pushed hard and ran up stats in the third.

The Rangers game met the eye test perfectly. They outplayed the Rangers badly in the first period, but panicked when the Rangers made it 2-1 and gave up too many chances. Stats said the exact same thing.

So I don't just stare at a website on game nights guys, I'm very much an eye test person. The stats are good to verify what you're seeing, and this year I see a Flame's team that outplays the opposition more often than not but just can't finish.

That's what I see.

I'm not just leaning on stats to make a case.

Just my opinion though.

I respect your perspective and appreciate the time you put in to this.

I too pay attention to advanced stats, and sometimes appreciate them and sometimes the dependence guys (like that Wilson guy who was on another site) have on them, and from which they attempt to draw conclusions (or bring insight) disappoint me.

The best way I can describe it is this - during the Rangers game, the Flames generated many shots, and made Lundquist’s night a busy but easy one (except for the inside blocker save) and never did I feel like they were actually going to win.

That is where advanced stats have fallen down at this altitude. The flames put forth a game which, to horribly paraphrase, was coached full of fire and fury, signifying nothing.

[Toonage]

Originally Posted by chedder

Came across this si article from when Glen was hired.

http://amp.si.com/nhl/2016/06/17/cal...ew-head-coach?

From the article. Interesting:

One of the knocks against him in Dallas was his in-game decision making. He struggled to read and react as the game unfolded in front of him and that eventually cost him the room. Has he addressed that glaring issue sufficiently during his years as an assistant? And has he learned to build a bridge with veterans? His success in Calgary will rely heavily on the buy-in of players like Mark Giordano, Matt Stajan and Michael Frolik. Can he get them onside?

Remarkable bit of proof that Gulutzan just doesn't get it.

Not only has he not improved on his previously perceived inability to read and react as the game progresses but he went too far in the other direction so far as building a bridge with his veterans. Relies on some of them too much you could say.

I would argue that he has in fact lost the room. They may never say it, but their play does.

Again, not trying to pump Flamesnation's tires but the 3 part series they did on why GG should be let go is tremendously informative especially regarding his last year in Dallas and how it compares to this past year in Calgary.

[nfotiu]

Originally Posted by GullFoss

This is exactly why the flames lose while having more shot attempts in a high danger area.

Take seven examples of opportunities of shots in high danger areas:

1) three-on-one
2) breakaway
3) two-on-one
4) Three-on-two
5) four-on-three
6) shots of the cycle
7) rebounds off the cycle.
8) rebounds off an odd-man rush

You can't possibly believe that all of these opportunities are equal scoring chances. But saying "the flames had 15 high danger scoring opportunities vs 9 for the opposition" assumes exactly that. It ignores that all 15 of the flames chances game might come on 3-on-5 cycles (the flames dont activate their defensemen) where the shots we're blocked or the angle was narrowed for the goalie to make an easy save. It ignores that the opposition might have had 9 breakaways.

If the flames get 15 high danger scoring chances, each with a 10% chance of scoring. And the opposition gets 9 scoring chances, each with a 50% chance of scoring. What's the end result in the game vs what do the advanced stats say should happen?

This has been my argument all year long in the game threads. The flames often have more opportunities to score, but they're inferior opportunities. Why? Its the system. The D-to-D passes and east-to-west hockey reduce the liklihood of generating scoring opportunities on the transition. And its the scoring chances that come from the transition that hasve the higher probability chances - the 2-on-1s the 3-on-1s the breakaways.

I agree 100%, and also think that the system they play at home is giving up a lot of odd man rushes and breakaways. I think they flashed a stat on TV on Friday saying the Rangers were 9 to 1 in odd man rushes and breakaways. If your team is constantly putting pressure on the other team and getting some some solid scoring chances, but give up a breakaway time after time, then that is not a winning formula. Are there stats for breakaways and 2 on 1's for and against? My eye test says we are giving up way more than those than we are getting at hone. If these fancy stats don't take breakaways and 2 on 1s into account, and weight them more heavily than other danger stats, then they are just not telling the whole story.

[Bingo]

Originally Posted by nfotiu

I agree 100%, and also think that the system they play at home is giving up a lot of odd man rushes and breakaways. I think they flashed a stat on TV on Friday saying the Rangers were 9 to 1 in odd man rushes and breakaways. If your team is constantly putting pressure on the other team and getting some some solid scoring chances, but give up a breakaway time after time, then that is not a winning formula. Are there stats for breakaways and 2 on 1's for and against? My eye test says we are giving up way more than those than we are getting at hone. If these fancy stats don't take breakaways and 2 on 1s into account, and weight them more heavily than other danger stats, then they are just not telling the whole story.

There is never going to be an answer to this debate as eye test is individual and the stats don't support inferior or superior chances within the box, at least at this point.

But these charts do help in at least proximity to the net for shot attempts.

Dallas Game

Eye Test - Flames iffy start, finish first strong, get down two in the second and pour it on.

Five on Five

Five on five I think makes your point. The Flames have the edge in scoring chances but they don't really concentrate their shots in the dangerous areas. Five on five chances were 8-4 Flames.
All Situations

Here the Flames dominate, showing the Bishop effect in the game. The Flames added nine more chances without being five on five (shorthanded chances and powerplay), and they were very concentrated in close and dangerous. This is Bishop stealing a game.

Colordao Game

Eye Test - Flames aren't the better team in the first but come out with a lead. Avalanche take it away in the second, Flames down push hard in the third.

Five on Five

Heat map shows a pretty even game, but the Flames broke down and gave up too many huge chances. Heat map shows a lot of the Flames shots came from well out, but they did get some in close chances. Flames gave up only four five on five scoring chances all game, all in the second period.
All Situations

In all situations the Flames had a 14-6 edge in scoring chances, and the map reflects many were in tight. However score effects are the difference as the Flames had a 8-1 edge in scoring chances after the Avalanche fourth goal. They didn't deserve to win this game, but Rittich wasn't great.

Ranger Game

Eye Test - Flames all over the Rangers in the first, but lose their mojo giving up an early 2nd period goal. Give up way too many chances in trying to tie it up. Better third, but lost in the second.

Five on Five

Graph fits the eye test as the Rangers have the better five on five chances. Flames go for it, and give up too many odd man rushes and breakaways, heat map reflects that. Chances are 18-14 Rangers at even strength.
All Situations

Flames are 7-0 in scoring chances in other situations, which catches the heat map up overall making the game look a lot more even than the shot clock would have you believe.

All in all these three games have the eye test match the stats. And by looking at heat maps I think the "Flames get weak chances in the box while giving up strong ones" suggestion is at least somewhat challenged.