Francis: Treliving filled with disappointment, anger after team's collapse

[Red]

Originally Posted by Jay Random

A lot of people on CP – I think you may have been one of them; I would have to go over a lot of posting history to be sure, but if this doesn't apply to you, great – a lot of people watched, analysed what they saw, came to a conclusion some time in October or November, and then only saw the things that supported that conclusion. Which is worse than not watching the games at all.

If you watched the games and didn't keep count of the scoring chances, you only have an impression of what you saw. Impressions are not analysis. That's why people keep statistics in the first place: so we can check our impressions against the data, and not merely go on looking for things that confirm the impressions we have already formed.

Yes, this is very valid. I may have been biased all along thus my take is not accurate. It's fair game to question my abilities to properly analyze it.
But my analysis being ineffective does not change the facts about how stats people collect their data. By watching the games like I do. Those stats are a compilation other peoples impressions. Is it not?

Unless we are agree that by looking at heat maps you can tell a good shot from bad one etc. Because I do not agree with that.

So in the end, there is no scientific way to determine which shot was more dangerous than another. There is no micro chip there, or some other advanced tech or analysis that tells us how shot 1 is better than shot 2. We are all left with our eyeballs.

I watch them play and come to some conclusions. I see the results and they match my conclusion.
Stats do not disprove my conclusion. Stats conclude many things, but the results don't match those conclusions.

One takeaway is that it is very hard to predict hockey. A lot of random events and momentum swings that effect results.

[Calgary4LIfe]

Originally Posted by Cali Panthers Fan

The point is your analysis may be flawed because of the subjectivity of it. There's no empirical data to go along with your interpretation. So yes, this time your interpretation is correct, but it's about 50/50 or maybe 60/40 at best to assume your interpretation will be correct in the future.

I can't tell you how many times I thought I was right about player X or team Y only to be proven wrong later. Objective analysis will often give you perspective on a situation that you may not have been aware of, and it's an objective reality, so you have to include it in your thought process somewhere. How you decide to form a conclusion should have some sort of objective process to it, because we are all prone to bias, and we are especially prone to emotional responses/overreactions when we are talking about sports.

This is where a disconnect is forming with advanced metrics to me. The empirical data SHOULD be showing something different than what we as fans are seeing.

So, do we believe in the empirical data provided and find some relief in that the system implemented is indeed a good system that the Flames should continue with, and that it is predicting a turn-around for next season?

Or perhaps the empirical data is incomplete as we know it, and that the 'eye-test' not aligning with the advanced metrics is providing enough of a discrepancy whereby we should be questioning the data in the first place, and/or looking for differing/additional data?

I understand exactly what you are meaning here - and I actually mostly agree. I just think that the data itself is faulty (or at best, forming an incomplete picture), and the lack of additional data is making the argument difficult to complete.

Too many teams this year seem to be showing good CORSI and so on, but haven't matched up with the expectations. I do agree that the existing advanced metrics are SOMEWHAT predictive - there are definitely a lot of examples where they align, and they shouldn't be completely discredited or irrelevant. I just think they are incomplete and the confidence interval is too low for people to be this fixated on them.

They are additional info at best right now (IMO), until more data sets measuring different events (or measuring them in a more effective manner) start coming out.

For now, we are left with the 'eye test' vs 'advanced metrics', and because they aren't always aligned, one crowd shouts "advanced metrics are garbage" while another crowd shouts out "it is random luck affecting them".

Truth is somewhere in the middle, and I don't think you can necessarily discount either side. This very team's own history over the last 4 or 5 years has shown too much of an inconsistency to have advanced metrics predictive of anything. They are descriptive, and they allow me to better understand a certain system, and it makes me take interest in possible explanations either way (which I actually find very interesting).

I just think that existing data is incomplete and can't be as trusted/accepted as it seems to be, and that the eye test can't be as trusted either as people have biases or just plain see things differently (or interpret them differently).

Is this team's CORSI really reflective of how well they play? Definitely not. Are they only taking long range shots from bad angles to 'pad the stats'? Definitely not. Truth is somewhere in the middle for me.

The only thing I come up with is contrasting and comparing the two systems we have seen this team play in recent memory - Hartley's and Gutluzan's - and see why both seemed to buck the trend that the advanced data was telling us would happen. One relied on a lower volume of shots and trying to get odd-man rushes, the other relies on having a higher number of shots and breaking out as a 5 man unit maintaining as much possession as possible. I find it extremely interesting how they have had somewhat similar results in the actual standings, but the game-play and resulting advanced metrics very different. At this point, neither side should either be believed or discounted, even without the available empirical data.

No side seems to have the required data to explain anything definitively. So I don't mind the speculating from both sides - except the shouting 'my eye tells me' and 'luck' that kind of goes on at times.

[Jiri Hrdina]

Originally Posted by nik-

Guys, if you don't take the maximum amount of time possible to arrive at the same conclusion, you're an irrational hothead and your opinion means nothing.

Yup that's what people are saying.

[Fan in Exile]

Amusing stuff here. We're now in a philosophical quandary over whether it's better to have the right result or have an established thought process which produces the wrong result. Supposedly those who came to an early opinion and then saw examples confirming that early opinion are more biased than those who stuck to a position supporting everything the coaching and management say and do, and stuck to stats that indicated that somehow the results had it wrong and things would even out.

I say hiring Gulutzan was wrong. Not many people supported it when it happened. He did not appear to be the best candidate, even among coaches in the Canucks organization. After one year of Travis Green as Canucks' HC I would agree.

There are numerous reasons to reach the conclusion that Gulutzan is not a good coach. These have been in evidence from the outset, and as time has passed, it has crystallized. Even so, as recently as three weeks ago, there were still some true believers that thought the Flames were about to turn it around.

It is an aspect of human nature that people are reluctant to admit they are wrong. Treliving maintains that Gulutzan is a good coach, as do a number of posters here. Or at least that the poor coaching is somehow disconnected with how the Flames have regressed. May as well tune in to Sarah Huckabee-Sanders.

[djsFlames]

Originally Posted by Classic_Sniper

Just posted this in the other thread, but outside of 1 season, Hartley's teams came back less than Gulutzan's teams. So I don't think you can hang that on GG. 14-15 was a very fun, but very fluky season. Everything went right, we were getting all the bounces and I wouldn't put too much stock into Hartley's come back capability.


Hartley - Comeback wins:
12-13: 2
13-14: 11
14-15: 23
15-16: 10

Gulutzan - Comeback wins:
16-17: 16
17-18: 12


Lastly, this whole ipad crap needs to die. He's looking to see if there was goaltender interference and an offside on the play to potentially call for a review which is way more important and beneficial than the alternative. Like what do you people want? For him to stare down the players who aren't even paying attention to him anyway for the sake of your own preconceptions.

What are you qualifying as a 'comeback'? Games that were tied up in the 2nd or early 3rd are borderline.

We tied up MAYBE one game with the EN this season. Tkachuk helped the team even up that one game with a couple goals in the middle of the third against Arizona I think it was? But unless we've evened up a game in the first few minutes of the third period, where the gap was no more than one goal, the Flames have typically folded like a cheap tent to finish games.

We ####ing sucked this season when pulling the goalie. And I distinctly remember the team being far, far better at pushing to tie in those situations under Hartley. There's no comparison when it comes to pushback.

[Shazam]

Originally Posted by Calgary4LIfe

This is where a disconnect is forming with advanced metrics to me. The empirical data SHOULD be showing something different than what we as fans are seeing.

So, do we believe in the empirical data provided and find some relief in that the system implemented is indeed a good system that the Flames should continue with, and that it is predicting a turn-around for next season?

Or perhaps the empirical data is incomplete as we know it, and that the 'eye-test' not aligning with the advanced metrics is providing enough of a discrepancy whereby we should be questioning the data in the first place, and/or looking for differing/additional data?

Yes, it's very incomplete.

There is this concept. It is called the information shadow. When you amass data about anything at all, you're actually just making a representation of that thing. The problem is that people take that set of data and presume that it is the thing itself, when all you really have is a shadow.

It's one of those fancy IT things that everybody ignores.

[Looch City]

Originally Posted by Pointman

Darryl Sutter's interview when he was hired as Flames' head coach:

"This is not a short-term deal for me," Sutter said. "I want the identity of the Calgary Flames back to where it was -- this is a tough town, a tough place to play. Just to restore that is really important."

http://www.vindy.com/news/2002/dec/2...ames-debut-is/

Dammit I read this quote and got excited.

[Psytic]

Originally Posted by Galakanokis

I like a lot of the Flames players but there is little doubt that changes need to be made. The make up of the team is just not right. Now, I will also say I would make or would have made the same moves even if the Flames went on a run. I like the roster to a degree but you always look to improve not just make a trade for the sake of a trade.

Flames don't make a big trade unless it is a pretty significant piece coming back.

I'm not as optimistic given the Flames history. I don't want another Greg Gilbert/ Marc Savard situation or Phanuef for spare parts. Maybe I wasnt clear. Trades are needed but the return better be another star player this time around. We shouldn't trade Star players JUST for the sake of the coaches system itself. GG should be gone as well.

[nik-]

Originally Posted by Jiri Hrdina

Yup that's what people are saying.

It's how they're coming across. And it's been happening often this season.

How many times has the term "level headed" been used in counter to people who wanted this coach gone long ago?

The evidence hasn't really changed all year, so taking an exceptionally long time to come to a conclusion others have isn't a virtue, and it's definitely not something which should be used to diminish the opinions of others.

[squiggs96]

Originally Posted by Cali Panthers Fan

In short, I trust Bingo's conclusion far more than yours because he did the work to understand the problem objectively. Just because you were too lazy to support your conclusion doesn't mean we should give it the same validity.

The Flames have the 3rd best CF and 2nd best CF%. They have the best FF and 4th best FF%. They are 6th in shots per game and 7th in net shots per game. You can harp on these stats all you want, but when you look at the standings, these stats don't seem to add up. They should be near the top of the standings if this data were useful. Doing lots of work to understand corsi and fenwick, and concluding they were unlucky doesn't make the analysis better. Saying I'm lazy because I don't type out an essay to ensure that you can understand it, is false.

Here's a chart we can both agree on:


Here's my conclusion. The red snake is off the bottom of the grid. The team is nowhere near the standings. The games are painful to watch. The powerplay is effing abysmal. The standings reflect all of these things. I don't think I need to take a GB of data to explain that, and now that his conclusion matches my conclusion, we are now saying I got lucky. I disagree. I arrived at it a lot earlier because I was able to eliminate a lot of the noise, and focus on what mattered.

[Galakanokis]

Isn't one problem with advanced stats or stats in general is that they themselves are subjective? Is it a high danger scoring chance because of where it is on the ice or is it a high danger scoring chance because you have traffic in front of the net and you have the D and goalie out of position? Flames were great at getting shots off from certain areas of the ice. What they weren't great at is getting the D and goalie moving. When the Flames shot the goalie was typically square and there just isn't much there to shot at. Those stats are meaningless. Slow and methodical puck movement might have lead to a lot of shots on net but not real high scoring chances.

The sense I get about Gulutzan is that he himself is not emotionally engaged enough with his players throughout the game to have any influence on it. The way he conducts himself on the bench only seems to reenforce that. Any proof you need is in the results and Treliving's words in the article. How did Scotty Bowman put together so many great teams without the help of advanced stats?

If the Flames were a crappy team on paper and not expected to do much the process becomes important. You look for things they are doing right and you look to build on aspects of the game. If you are expected to win, be a playoff team the only thing that really matters are the results. From the start of the season it appears the Flames never really built on anything. They didn't improve in any aspect and quite frankly regressed a lot in others. Not a good sign for the coaching staff.

[Galakanokis]

Originally Posted by Psytic

I'm not as optimistic given the Flames history. I don't want another Greg Gilbert/ Marc Savard situation or Phanuef for spare parts. Maybe I wasnt clear. Trades are needed but the return better be another star player this time around. We shouldn't trade Star players JUST for the sake of the coaches system itself. GG should be gone as well.

I am trying to be optimistic but deep down I fear the same thing.

[Joborule]

I've said it multiple times during the season, the only stat that I felt truly reflected how a team plays, and their position in the standings is their goals differential. The Flames were in the red all season long. That stat clearly indicated that the Flames have difficulty getting more goals than the opposing team; despite all the advance stats being greatly in their favour. I think there's a few games a season where you could use the unlucky argument, but the amount of times the Flames lost yet our corsi'ing the opposing team, indicated that the advance stats doesn't correlate to the scoring result at the end of the game, at least entirely.

Perhaps with how slow and conservative the system is, it leads to positive corsi numbers, but not enough goals because it's not aggressive and reactive enough. It's about the quality of the shots, rather than the quantity after all.

There's also a stat flying around about Gulutzan regulation record over the past two seasons, and how abysmal it actually is. If it wasn't for the team overtime success last season, this would be two straight seasons of missing the playoffs under him.

So while the advance stats may be in GG's favour, the important stats that impact teams positions in the standings, are not in his.

[Psytic]

Regardless of the possession stats does anyone actually enjoy watching this team? Serious question, if it wasnt for hometown pride would anyone actually enjoy watching this? Even if they were winning teams like the Pens, Vegas and the Hawks (before they fell off a cliff) have shown that you can win with a more exciting system. Do we need to subject ourselves to more of this? Win or lose, good possession stats or not I hope Treliving can see the lack of entertainment value for the fans GG hockey provides. I guess being molded for years in the Phoenix system maybe he has lost sight of what entertaining hockey looks like.

[Royle9]

Originally Posted by Galakanokis

Isn't one problem with advanced stats or stats in general is that they themselves are subjective?

Yes, its ALL subjective which is why this argument is funny because both sides are trying to convince each other they are more right and the way they got there is smarter because it took more time to analyze.

Subjective thats the bottom line.

[stone hands]

The flames problem wasn't that corsi/Fenwick are flawed statistics, it's that they have 1 guy on the entire team who can shoot the puck and hit the net and he had a wrist injury for the last 6 months***

You could have 95% time of possession, but if nobody can bury it doesn't make a lick of difference

*** as it pertains to the advanced metrics debate. There are an assload of other issues this team faces, like its abysmally flaccid powerplay strategy

[Bingo]

Originally Posted by The Fonz

I think it's pretty much exactly what many of us have been saying for the majority of the season. Don't discredit someone else's analysis simply because it took you longer to arrive at the position.

Show my your analysis then.

I didn't discredit anyone. I'm saying I think it should be looked at. It would be prudent.

Arriving at a conclusion without an analysis isn't an analysis

[Royle9]

Can one not visualize and analyze at the same time? Does it have to be mathematically calculated in %'s and CORSI to count as legit analysis?
Honest question.

[smiggy77]

Originally Posted by nik-

I'm glad things are never as bad as they seem, because they seem pretty ####ing bad.

Bad has an evolving meaning based on what happens up north.

[Bingo]

Originally Posted by Red

Yes, this is very valid. I may have been biased all along thus my take is not accurate. It's fair game to question my abilities to properly analyze it.
But my analysis being ineffective does not change the facts about how stats people collect their data. By watching the games like I do. Those stats are a compilation other peoples impressions. Is it not?

Unless we are agree that by looking at heat maps you can tell a good shot from bad one etc. Because I do not agree with that.

So in the end, there is no scientific way to determine which shot was more dangerous than another. There is no micro chip there, or some other advanced tech or analysis that tells us how shot 1 is better than shot 2. We are all left with our eyeballs.

I watch them play and come to some conclusions. I see the results and they match my conclusion.
Stats do not disprove my conclusion. Stats conclude many things, but the results don't match those conclusions.

One takeaway is that it is very hard to predict hockey. A lot of random events and momentum swings that effect results.

And maybe the thing you guys don't see in me is my obsession with getting my bias out of things.

I love having stats that take passion from the view.

With the advent of advanced stats it gives me an overlay to watch the game. When I see a period of play I check the stats when I'm live or later if I'm not to see if the numbers support what I'm seeing.

More often than not they do.

But I'm not arrogant enough to think my view is the only view, or that my view is the correct view so I suggested what I'd do to get rid of another layer of bias is look at those high danger chances.

So I'm not judging anyone.

I'm open to the stats being wrong the whole time.