Flames possession numbers

[WhereIsIt]

Hey guys, I don't know if any of you caught Thomas Drance on the Fan 960 Friday morning. That's the guy who pens the Score power-rankings and refuses to remove the Flames from his bottom-10. You can check out that conversation here.

Anyways, Drance was talking like the Flames were one of the worst and luckiest teams in the league, so I wrote an article in response (excuse the awkward layout, site just got redesigned): http://alongtheboards.com/2014/12/flames-friday-numbers-dont-die/

The Flames are not a very good possession (or perhaps more accurately, shot-attempt differential) team. Actually, they are getting slayed in that department. They are getting badly outshot , and that’s not up for debate.

What is up for the debate is how to fill the gaps between the Flames shot-attempt differentials and their actual record. Luck is a popular theory, and that certainly plays a role — especially in their third period comebacks. There’s been more “they didn’t really deserve that, but oh well†wins than “they deserved better†losses. You can subtract a handful of points from the Flames’ season total in that context and see that they should probably be just outside of the playoff picture in a fair and perfect world.

But there are other factors that Corsi doesn’t attempt to cover.

Lust wanted to here your guys's thoughts, because we don't seem to discuss analytics a lot/at all on these boards.

Click to read more...

 

[slappipappi]

"The stats and other factors imply a regression should be coming, not a complete collapse".

You concluded you argument with this statement, but I didn't see anything that talked about what stats and other factors imply a small regression.

 

[Volica]

I think... that people need to stop using advanced stats like scripture.

Do other teams attempt a lot of shots, yeah; but that's our defensive structure. We clog lanes and make things very difficult to get through. It's why although 'attempts' are bad, actual shots against per game has us in the top half of the league. This again, is by design. We have a ridiculously mobile group; a blocked shot is an offensive rush that has provided this team with a lot of sparks, and points. It's definitely how this team works.

I will also add the Flames do A LOT of work setting up the proper shoot scenario, so much so that it makes me feel we have Tanguay back on the team at times. I don't think the shooting percentage is a fluke; I think it's by design.

Calgary is once again a team that is absolutely bottom feeding in the league again at Faceoffs. We're not a team that can slide a puck on net, then expect to come away with a win at the dot. This team holds pucks and sends them deep if there's no play... There's honestly very little 'put it on net at any situation' (annoyingly so).

I think some guys are assuming that these stats make or break teams and players; I mean, Edmonton went out this summer and picked up two high end advanced stat players. Guess what? They're still bottom feeding; 11 games in a row, but we can chop that up to Calgary stealing their 'Puck luck' right?

 

[WhereIsIt]

"The stats and other factors imply a regression should be coming, not a complete collapse".

You concluded you argument with this statement, but I didn't see anything that talked about what stats and other factors imply a small regression.

I didn't say small regression, I said not a complete collapse

Some are predicting a complete collapse (e.g. missing the playoffs by a wide margin) given Calgary's Corsi is 29th in the league, but like I said, that doesn't take into account blocked shots, injured players, limiting shots to the outside, having two starting-quality goaltenders instead of one and riding the hotter one, etc. All of those factors were in there.

 

[WhereIsIt]

I think... that people need to stop using advanced stats like scripture.

Do other teams attempt a lot of shots, yeah; but that's our defensive structure. We clog lanes and make things very difficult to get through. It's why although 'attempts' are bad, actual shots against per game has us in the top half of the league. This again, is by design. We have a ridiculously mobile group; a blocked shot is an offensive rush that has provided this team with a lot of sparks, and points. It's definitely how this team works.

I will also add the Flames do A LOT of work setting up the proper shoot scenario, so much so that it makes me feel we have Tanguay back on the team at times. I don't think the shooting percentage is a fluke; I think it's by design.

Calgary is once again a team that is absolutely bottom feeding in the league again at Faceoffs. We're not a team that can slide a puck on net, then expect to come away with a win at the dot. This team holds pucks and sends them deep if there's no play... There's honestly very little 'put it on net at any situation' (annoyingly so).

I think some guys are assuming that these stats make or break teams and players; I mean, Edmonton went out this summer and picked up two high end advanced stat players. Guess what? They're still bottom feeding; 11 games in a row, but we can chop that up to Calgary stealing their 'Puck luck' right?

I agree, especially with the bolded part. Even if Calgary's offense was better at generating shot attempts, the defensive structure might preclude good possession numbers anyways. Shots are welcome against Calgary, just not very good ones. Seems to be working fine.

If the Flames do collapse it will because of a complete offensive meltdown, not because we allow too many shot attempts.

 

[slappipappi]

I agree, especially with the bolded part. Even if Calgary's offense was better at generating shot attempts, the defensive structure might preclude good possession numbers anyways. Shots are welcome against Calgary, just not very good ones. Seems to be working fine.

If the Flames do collapse it will because of a complete offensive meltdown, not because we allow too many shot attempts.

Calgary's save % is actually not particularly out of line with expectations at about .915. Their GA/GP and SA/GP are about the same and above average in the NHL.

Where Calgary's advanced stats don't align with reality is in the scoring percentage.

They sit 5th in the NHL in G/GP, and 28th in S/GP. Historically teams in the NHL don't do that.

It's seams reasonable to suggest that Calgary will start scoring less unless it starts generating more shots on goal.

 

[Johnny Hoxville]

Calgary's save % is actually not particularly out of line with expectations at about .915. Their GA/GP and SA/GP are about the same and above average in the NHL.

Where Calgary's advanced stats don't align with reality is in the scoring percentage.

They sit 5th in the NHL in G/GP, and 28th in S/GP. Historically teams in the NHL don't do that.

It's seams reasonable to suggest that Calgary will start scoring less unless it starts generating more shots on goal.

Brad Treliving even stated those are among his expectations. With that said, because he expects that, the coaching staff has been trying to prepare the team accordingly.

 

[Volica]

Brad Treliving even stated those are among his expectations. With that said, because he expects that, the coaching staff has been trying to prepare the team accordingly.

It's kind of funny. If Calgary was putting up 3 more shots a game nobody would be saying anything. It's almost kind of funny, to be honest. 3 shots. What's 3 shots most nights? Those off-wing lazy wristers to the goalies?

 

[berto14]

I couldn’t find a source for team blocked shots but, according to the Sportsnet broadcasts, the Flames are top in the league in that category.

Here you go:
http://www.nhl.com/ice/teamstats.ht...AAll&sort=blockedShots&viewName=realTimeStats

 

[slappipappi]

It's kind of funny. If Calgary was putting up 3 more shots a game nobody would be saying anything. It's almost kind of funny, to be honest. 3 shots. What's 3 shots most nights? Those off-wing lazy wristers to the goalies?

3-4 shots will generally result in about a third of a goal extra per game. Which is about 27 goals per season, which is actually quite a huge amount.

And 3-4 shots aren't always off-wing wristers, they are sometimes tap-ins.

The fact of the matter is, teams that average less than 27 shots per game are usually near the bottom of the team's goals scored list, not at the top. There will be some regression, we just hope it's not too severe.

 

[WhereIsIt]

Here you go:
http://www.nhl.com/ice/teamstats.ht...AAll&sort=blockedShots&viewName=realTimeStats

Thanks man, I have no idea why I couldn't find this haha.

 

[SmellOfVictory]

I think... that people need to stop using advanced stats like scripture.

Do other teams attempt a lot of shots, yeah; but that's our defensive structure. We clog lanes and make things very difficult to get through. It's why although 'attempts' are bad, actual shots against per game has us in the top half of the league. This again, is by design. We have a ridiculously mobile group; a blocked shot is an offensive rush that has provided this team with a lot of sparks, and points. It's definitely how this team works.

I will also add the Flames do A LOT of work setting up the proper shoot scenario, so much so that it makes me feel we have Tanguay back on the team at times. I don't think the shooting percentage is a fluke; I think it's by design.

Calgary is once again a team that is absolutely bottom feeding in the league again at Faceoffs. We're not a team that can slide a puck on net, then expect to come away with a win at the dot. This team holds pucks and sends them deep if there's no play... There's honestly very little 'put it on net at any situation' (annoyingly so).

I think some guys are assuming that these stats make or break teams and players; I mean, Edmonton went out this summer and picked up two high end advanced stat players. Guess what? They're still bottom feeding; 11 games in a row, but we can chop that up to Calgary stealing their 'Puck luck' right?

You're creating a false dichotomy between the percentages and actions by players/teams. What is described as "luck" by analytics is simply anything that cannot be sustained over the long term and cannot easily be measured or accounted for. Even if the Flames are currently keeping most shots to the outside, that doesn't mean it's sustainable. Every single year there's a team that "keeps teams to the outside" while getting badly outshot and outperforms their analytics. And every year, or the year after, that team's performance takes a nosedive when things balance out.

Why does it balance out? Who knows? Some of it is almost certainly blind luck or hot streaks (e.g. the ridiculous goaltending on Hiller's part), some of it could be other teams adapting to a novel strategy that works over a short span of games, etc. But whatever it is, the story is always the same: you get out-attempted over the long term and you will eventually start to lose.

 

[slappipappi]

You're creating a false dichotomy between the percentages and actions by players/teams. What is described as "luck" by analytics is simply anything that cannot be sustained over the long term and cannot easily be measured or accounted for. Even if the Flames are currently keeping most shots to the outside, that doesn't mean it's sustainable. Every single year there's a team that "keeps teams to the outside" while getting badly outshot and outperforms their analytics. And every year, or the year after, that team's performance takes a nosedive when things balance out.

Why does it balance out? Who knows? Some of it is almost certainly blind luck or hot streaks (e.g. the ridiculous goaltending on Hiller's part), some of it could be other teams adapting to a novel strategy that works over a short span of games, etc. But whatever it is, the story is always the same: you get out-attempted over the long term and you will eventually start to lose.

I think the theory that Calagry is "keeping teams to the outside" is simply that, and likely true.

if it was, teams could utilize this strategy and many teams would post such stats for many years. Since they don't, you can likely assume that in fact they aren't doing a particularly better job of this than anyone else.

Colorado last year out performed their advanced stats over the course of the year, until the playoffs. This year, with more or less the same team, they are no longer outperforming their stats.

Does this mean that Calagry has to regress this year?

Absolutely not.

But there is no real reason to suggest that going forward Calagry will out perform their advanced stats any more than the next team.

We shouldn't expect that Calgary will get bad luck since they've had a run of good luck this season. That good luck has been banked. From here on in, Calgary should have average luck, and should finish the year on the plus side all things being equal.

But I doubt that suddenly that Calgary has found some magic formula to out perfroming their advanced stats. Teams do such a thing, but it's never been sustainable for more than a year.

I'm a big proponent of the advanced stats, but Calgary can make the playoffs this year. They just need some help to do so. And th elonger their winning goes, the less help they will need. But we've still got a long ways to go. And if this was the lockout year, when the season was 48 games long, it would be that much easier.

 

[Anglesmith]

I've seen several shot heat-maps and shot location charts that show without a doubt that the Flames take shots from better positions than their opponents. I don't know why these aren't considered when looking at advanced stats.

 

[WhereIsIt]

I've seen several shot heat-maps and shot location charts that show without a doubt that the Flames take shots from better positions than their opponents. I don't know why these aren't considered when looking at advanced stats.

I don't think they've found a good/easy way to integrate those yet, so often that aspect is wrongly dismissed. Should become easier with some of the technology the NHL and Sportsnet are rolling in during the next few years. The analytics we have now are so primitive compared to what's coming in my opinion.

 

[Johnny Hoxville]

I don't think they've found a good/easy way to integrate those yet, so often that aspect is wrongly dismissed. Should become easier with some of the technology the NHL and Sportsnet are rolling in during the next few years. The analytics we have now are so primitive compared to what's coming in my opinion.

It should also really add to fantasy leagues.

 

[SmellOfVictory]

I've seen several shot heat-maps and shot location charts that show without a doubt that the Flames take shots from better positions than their opponents. I don't know why these aren't considered when looking at advanced stats.

Partially because they have to be individually collected and they're not collected for all teams nor are they easily integrated into something like a Corsi table.

 

[Anglesmith]

Partially because they have to be individually collected and they're not collected for all teams nor are they easily integrated into something like a Corsi table.

I feel that that's garbage, though. In the case of comparing the Avs to the Flames, for instance, it could really make a case for whether they really are comparable or not.

Advanced stats are, in my opinion, simply a way to analyze games without needing to watch the games. People who follow sports have always known that often the scoreboard does not indicate how the game really played out, and you could tell if a team was playing well only if you watched the game. Now people are trying to replace that because they want to be able to talk about the performance of all teams in all games without actually being obligated to watch all the games. It's how people like Drance make a living.

What's going to drive the field forward, however, is critical analysis of how the advanced stats might skew perception of how teams played just as much as the scoreboard can. Obviously some of that has been discovered already: there are things like Adjusted Corsi, Fenwick Close etc. which seek to correct for obvious manipulative factors on the numbers. However, now that the field has surged to the mainstream and people are making money off of the notion, it seems to have become a very defensive field. Eventually that will break down. There are clear flaws in the system that can be sussed out by simple "extreme case" thought experiments.

From my own perspective, as a scientist, the way that shot quality is pointedly ignored as a factor by the field is completely ridiculous. If I was running an experiment that I knew was affected by temperature, I would never manage to publish a paper on that experiment unless I showed that I had accounted for temperature fluctuations. From that perspective, it's very difficult to take the field seriously as it is today, and even moreso when the people leading the charge are often smug buffoons such as Thomas Drance instead of the rational thinkers who are actually trying to advance the science.

 

[HighLifeMan]

Shot blocking is a very legitimate and much needed skill in the game today. That is an area in which Calgary excels at more so than any other team in the NHL, and I personally find it silly to dismiss them for relying on a defensive structure that helps promote that area of the game.

In addition to that Calgary has actually outshot their opponents over the last 23 games of the season (albeit only by a +3 differential). The team got absolutely dominated in shot numbers throughout the first five games of the season (-62 differential!!) yet still managed to win three of those contests. It's not as if we have been consistently outshot by the opposition throughout the year. Corsi numbers can be misleading, and I think the Flames are a great indication of that.

The only concern I have in regards to these "stats" is the teams shooting % on the whole. I do expect some regression in this area but also believe the team will find more ways to get pucks on net when the likes of Stajan, Backlund, Raymond and Colborne are back in the lineup on a full time basis which will help ease that regression over time.

Overall if this team continues to play as it has all year long I have an extremely tough time believing they are not a playoff team.

 

[SmellOfVictory]

Advanced stats are, in my opinion, simply a way to analyze games without needing to watch the games. People who follow sports have always known that often the scoreboard does not indicate how the game really played out, and you could tell if a team was playing well only if you watched the game. Now people are trying to replace that because they want to be able to talk about the performance of all teams in all games without actually being obligated to watch all the games. It's how people like Drance make a living.

The funny thing is, a lot of analytics guys watch way more hockey than the average fan. The reason scoring chance info is publicly available is normally because there's an analytics nerd out there who's painstakingly going through each game and watching/rewinding/pausing/etc. to get that info. Or the new thing, zone entries: all a guy going in and watching for literally every zone entry made, recording and compiling the data.

And I fully subscribe to the opinion that stats are better than the average eye. We watch hockey for entertainment; how can you possibly expect someone to accurately gain useful information with one set of eyes that are naturally going to be drawn to the action while watching? Unless you specifically set out with the mindset of "I am going to watch player x all game" or you rewatch a game three or four times over, the eye test is complete garbage.

From my own perspective, as a scientist, the way that shot quality is pointedly ignored as a factor by the field is completely ridiculous. If I was running an experiment that I knew was affected by temperature, I would never manage to publish a paper on that experiment unless I showed that I had accounted for temperature fluctuations. From that perspective, it's very difficult to take the field seriously as it is today, and even moreso when the people leading the charge are often smug buffoons such as Thomas Drance instead of the rational thinkers who are actually trying to advance the science.

They've done analysis on scoring chances vs Corsi vs Fenwick, and they've looked at percentages over the long term and found that, while shot quality is a factor, it's a minor one in the long term. There is so much parity in the league that even a garbage fire like the Oilers and a borderline dynasty like the Hawks aren't far enough separated that one will consistently give up or generate a significantly larger proportion of high quality shots than the other. As high quality shots increase, so do low quality shots on a roughly proportionate basis.

That's not to say that people don't want to include shot quality, or that it shouldn't be included. It's just that ignoring it is not enough to discredit the entire set of statistics. Unfortunately shot quality data is also not widely available and has not been compiled in a useful fashion yet.

Analytics are by no means perfect, and they have a long way to go before they're truly mature, but they're still better than the average dude's eye test.

 

[Anglesmith]

The funny thing is, a lot of analytics guys watch way more hockey than the average fan. The reason scoring chance info is publicly available is normally because there's an analytics nerd out there who's painstakingly going through each game and watching/rewinding/pausing/etc. to get that info. Or the new thing, zone entries: all a guy going in and watching for literally every zone entry made, recording and compiling the data.

And I fully subscribe to the opinion that stats are better than the average eye. We watch hockey for entertainment; how can you possibly expect someone to accurately gain useful information with one set of eyes that are naturally going to be drawn to the action while watching? Unless you specifically set out with the mindset of "I am going to watch player x all game" or you rewatch a game three or four times over, the eye test is complete garbage.

...

Analytics are by no means perfect, and they have a long way to go before they're truly mature, but they're still better than the average dude's eye test.

I just don't agree with these points in particular. I feel that the eye-test is the best way to analyze who's playing better. There's nothing that numbers pick up that you can't pick up by watching. If a team is getting more shot attempts, you see that watching the game. If a team is having trouble maintaining possession, you see that watching the game. If a team can't enter the zone when a certain player is on the ice, you see that.

And it goes beyond that. The eye test allows you to sort this data qualitatively in exactly the ways that analytics can't. Case in point, the comparison between Toronto last season and Calgary this season. Everyone is able to compare their advanced stats, and we've seen a lot of that piled on us along with words like 'unsustainable' and 'collapse' and 'regression' all because the advanced stats say similar things. However, last season while the Leafs were successful everyone who watched the games could see they were getting outplayed often and were getting bailed out repeatedly by their goalie. This season, we've had plenty of games where we've lost the advanced stats yet felt like we were in control the whole time and allowed next to no chances (the first game of the back-to-back against the Coyotes stands out in that regard). And it isn't just Flames fans who have noticed this.

One easily noticeable point with regards to the Flames is just how few odd-man rushes happen against the Flames. I think that's the main factor in why we are defying odds at this point. The game in Tampa Bay was a complete anomaly because we gave up about 5 in a 5 minute stretch in the third, but in most games you're hard-pressed to find even one 2-on-1 or breakaway against. Meanwhile we can usually rely on getting at least two such chances a game because of our speed and Brodano beating the other team down the ice.

And the Oilers are another case which defies statistics. They're possession champions, but when you see them play, you can see why they are losing. They consistently hand their opponent quality chances with horrible defensive plays and breakdowns, while not being handed those same chances in return.

 

[Shawnofthedeadz]

I just don't agree with these points in particular. I feel that the eye-test is the best way to analyze who's playing better. There's nothing that numbers pick up that you can't pick up by watching. If a team is getting more shot attempts, you see that watching the game. If a team is having trouble maintaining possession, you see that watching the game. If a team can't enter the zone when a certain player is on the ice, you see that.

And it goes beyond that. The eye test allows you to sort this data qualitatively in exactly the ways that analytics can't. Case in point, the comparison between Toronto last season and Calgary this season. Everyone is able to compare their advanced stats, and we've seen a lot of that piled on us along with words like 'unsustainable' and 'collapse' and 'regression' all because the advanced stats say similar things. However, last season while the Leafs were successful everyone who watched the games could see they were getting outplayed often and were getting bailed out repeatedly by their goalie. This season, we've had plenty of games where we've lost the advanced stats yet felt like we were in control the whole time and allowed next to no chances (the first game of the back-to-back against the Coyotes stands out in that regard). And it isn't just Flames fans who have noticed this.

One easily noticeable point with regards to the Flames is just how few odd-man rushes happen against the Flames. I think that's the main factor in why we are defying odds at this point. The game in Tampa Bay was a complete anomaly because we gave up about 5 in a 5 minute stretch in the third, but in most games you're hard-pressed to find even one 2-on-1 or breakaway against. Meanwhile we can usually rely on getting at least two such chances a game because of our speed and Brodano beating the other team down the ice.

And the Oilers are another case which defies statistics. They're possession champions, but when you see them play, you can see why they are losing. They consistently hand their opponent quality chances with horrible defensive plays and breakdowns, while not being handed those same chances in return.

I personally have a bunch of problems with advance stats, one of them is a lot of people don't understand statistics in general (including a large amount of analysts). Remember Corsi is based solely shot differentials. In my mind to saying one team is better than the other (or player) based on shot differentials is absurd. For example, a team which takes a lot of bad shots a game from everywhere on the ice (Oilers), will be considered a better team who would style is to hold the puck and rather force into a good shooting lane before shooting (Flames). In addition, the same can be said in reverse. Defensive styles may vary, growing up playing hockey as a D, I was taught to keep shots to the outside and as long as they are from the outside (sides or blue line corners) your goalie should be expected to save them. Teams who play a defence similar to this style are going to suffer more under Corsi than teams who take more attempts at preventing all shots yet leaving better scoring opportunities open.

Not only is this season and a lot of seasons a good example of this, but when you play NHL with your friends isn't one of the biggest excuse always "well I out played you" despite losing? There are a lot of reasons why shot totals might change in game such as goals, penalty kills, etc.

 

[SmellOfVictory]

I just don't agree with these points in particular. I feel that the eye-test is the best way to analyze who's playing better. There's nothing that numbers pick up that you can't pick up by watching. If a team is getting more shot attempts, you see that watching the game. If a team is having trouble maintaining possession, you see that watching the game. If a team can't enter the zone when a certain player is on the ice, you see that.

And it goes beyond that. The eye test allows you to sort this data qualitatively in exactly the ways that analytics can't. Case in point, the comparison between Toronto last season and Calgary this season. Everyone is able to compare their advanced stats, and we've seen a lot of that piled on us along with words like 'unsustainable' and 'collapse' and 'regression' all because the advanced stats say similar things. However, last season while the Leafs were successful everyone who watched the games could see they were getting outplayed often and were getting bailed out repeatedly by their goalie. This season, we've had plenty of games where we've lost the advanced stats yet felt like we were in control the whole time and allowed next to no chances (the first game of the back-to-back against the Coyotes stands out in that regard). And it isn't just Flames fans who have noticed this.

One easily noticeable point with regards to the Flames is just how few odd-man rushes happen against the Flames. I think that's the main factor in why we are defying odds at this point. The game in Tampa Bay was a complete anomaly because we gave up about 5 in a 5 minute stretch in the third, but in most games you're hard-pressed to find even one 2-on-1 or breakaway against. Meanwhile we can usually rely on getting at least two such chances a game because of our speed and Brodano beating the other team down the ice.

And the Oilers are another case which defies statistics. They're possession champions, but when you see them play, you can see why they are losing. They consistently hand their opponent quality chances with horrible defensive plays and breakdowns, while not being handed those same chances in return.

The eye test is fine if it's a trained eye that's explicitly looking at a game from a critical standpoint (and again, probably rewatching parts of it multiple times) but unless that's being done and notes are being taken game-to-game so that fallible human memory and its inherent cognitive fallacies aren't at play, I think the eye test is pretty bad. And it should be done by multiple people who combine their viewpoints, to reduce personal bias.

And that's the thing: nobody that I'm aware of does anything remotely rigorous with the eye test (unless they're collecting data for statistics). Certainly, I'd bet money that very few, if any, HFBoarders do.

 

[slappipappi]

For example, a team which takes a lot of bad shots a game from everywhere on the ice (Oilers), will be considered a better team who would style is to hold the puck and rather force into a good shooting lane before shooting (Flames).

If this were true, it would make sense.

I'm not certain it is true though.

If it were, certain teams who play that style and have for years would be "beating" the asanced stats consistently.

And I don';t know of any team that does it for that long of a period.

It can be said that Colorado beat the advanced stats last year. They haven't changed their style of play this year, and the results are now matching up to the asdvanced stats.

The same thing happened to Toronto.

Have Calgary suddenly discovered style of hockey no one has ever done before?

 

[Johnny Hoxville]

The one thing I think the Flames will be unable to sustain is their shooting percentage. With that said, I expect them to compensate in other ways and still find ways to win games.

This is the kind of the thread I had in mind when I first suggested it in the OT thread.