Not to worry; this won't last says Myrtle

[Resolute 14]

Originally Posted by _Q_

But see the Flames are 6th in the league in goal differential. Why is it that a drop in PDO will affect Calgary but not Pittsburgh?

I'm actually looking forward to April when are analyst pulls out the "is anyone really surprised this team has made the playoffs? Their PDO is at the top of the league, I'm actually surprised they didn't finish higher in the standings." then everyone will think he's a genius.

A drop in PDO (relative to other teams) will affect Calgary and Pittsburgh and anyone else. But when you look at the roster of the two teams, which do you think is more likely to regress further? Almost everyone would pick Calgary.

We are where we are primarily because our shooting percentage is crazy high. Hockey Analysis has our 5 on 5 shooting percentage at 10.16%. Tampa Bay is at a similarly unsustainable 9.98%. Third is Montreal at 9.09%. The last team to shoot over 10% at 5 on 5 over the course of a season was Toronto in 2012-13 at 10.57%. In an 82-game season, Washington in 2009-10 at 10.35%. Losing one point of shooting percentage is about a 20 goal difference (based on last year's shot totals). We are currently on pace for 168 goals at 5 on 5. That 1% drop would cut our offence in this situation down by nearly 12% And at 9.16%, we would still be an elite offensive team. Fall to middle of the pack, and you see where this is going.

We can, of course, cover for this expected drop in other ways. We are already one of the top power play teams in the league (22.8% - 8th), so not much room for growth there. Other situations are statistically insignificant, so can't make any argument. (Incidentally, we lead the NHL with 4 goals at 4 on 4).

Or, our goaltending can improve. We're currently 14th in the league at 5 on 5 with a .926 save%. Honestly, I don't personally see much improvement there.

Where the Flames can cover a loss of PDO without a loss of results is in possessing the puck more. Your shooting and save percentages can fall, but if you are taking more shots and giving up less, then the impact of the changes are blunted. We are currently one of the worst teams in the league in terms of possession, which is why stat heads are looking at Calgary as being a 'lucky team'. We could well do what the Avs did last year and defy the odds to make the post-season. Personally, I would rather our underlying numbers improved. They appear to be, incidentally, but slowly.

[Textcritic]

^
This gets down to what this means for the Flames as far as I am concerned: the regression towards the mean for this team is not significant enough to make such a difference that it will see them plummet to near the bottom of the standings. Even as the shooting and save percentages continue to normalise (as I have said, we have already seen this taking place for a couple of weeks now) the Flames are providing an increased assurance that their success is not so dependant on these unsustainably high percentages.

[Textcritic]

^
This gets down to what this means for the Flames as far as I am concerned: the regression towards the mean for this team is not significant enough to make such a difference that it will see them plummet to near the bottom of the standings. Even as the shooting and save percentages continue to normalise (as I have said, we have already seen this taking place for a couple of weeks now) the Flames are providing an increased assurance that their success is not so dependant on these unsustainably high percentages.

[Itse]

Once I see a consistent pattern of analysts using advanced stats doing a better job at predicting development in the league than those who are not using them, I'll believe that they actually provide value.

So far I haven't seen even a hint of this. From what I've seen they're just as hit and miss as everyone else.

[Resolute 14]

Originally Posted by _Q_

So what's a reasonable shooting percentage then? Anywhere between 9 and 10%?

We were 13th last season in shooting percentage at 9.2%. We're obviously a better team this season, so can we expect to reasonably be closer to 10% and finish in the top ten in shooting percentage? And really, assuming we maintain a 26 shots on goal average that we're at right now, a drop in shooting percentage of say 1 or 1.5% only means a drop of two or three goals every 8 games. Is that a big enough drop to make a difference?

Most advanced stats rely on even strength data, usually 5 on 5. stats.hockeyanalysis.com is a good place to look at shooting percentages in this situation, and all others.

Here is this year, so far: http://stats.hockeyanalysis.com/team...T&sortdir=DESC

We are at 10.16%, first in the league. For comaprison, 10th place is 8.49%. That is a crazy large difference.

Last year, First was 9.83%, 10th was 8.01%. In 2012-13, first was 10.57%, 10th was 8.31%.

In short, we are currently on pace for 168 goals at even strength on 1657 shots. If we end the year at 8.49% - 10th as of right now, and still a high-end offence - we would score only 141 goals total, assuming the same number of shots. And that 27 goal difference would come out of the remaining 63 games. That would be a huge hit going forward.

[_Q_]

Originally Posted by Resolute 14

A drop in PDO (relative to other teams) will affect Calgary and Pittsburgh and anyone else. But when you look at the roster of the two teams, which do you think is more likely to regress

Sorry to cherry pick, and I can see some useful points in your argument, but see this is what bugs me about these advanced stats.

You're essentially saying Pittsburgh's PDO will drop, but they'll weather the storm better than Calgary because they're a better team. How is it then that PDO, if looking at it as a purely predictive tool as some writers use it telling us anything?

"LA has a low PDO, but that will improve because they're a good team."

"Buffalo is a horrible team, their PDO clearly shows that. "

" Calgary has a high PDO, but don't be fooled, they're not that great of a team and they'll come back down to earth."

Really, is any of this saying anything we don't know? That good team will improve in the standings and bad teams will drop? Do we need an advanced stat to tell us that?

I'd also argue with you that Pittsburgh is a bad team being propped up by two generational talents and that Calgary is a much better built team. But that is for regular old fashioned hockey analysis and observation, it has no room in this thread.

[Cleveland Steam Whistle]

I think this POD stuff can maybe be useful, if put in context. As I said earlier, of course I'm worried the Flames can't sustain this pace. But I'd be more worried if I felt the POD stuff was reflecting one of the following (which is what I think it's meant to reflect):
- Is the team getting above average goal tending that is stealing them games.
- Are they scoring at a ridiculous pace because they are either sniping or getting good bounces that end up with the puck in the net beyond the level of which their control of the play should earn.

The answer the both the questions above from a pure watching the game perspective from me is a NO to both questions. At the start of the year, our goaltending stole us some games, for sure, but since then, the Flames IMO have been full marks for their record and it's not like I feel they are scoring more than they should or that the goalies are standing on their heads. Until that's how it starts to feel while watching the games, I'm not going to worry about these POD stats.

[Resolute 14]

Originally Posted by _Q_

Sorry to cherry pick, and I can see some useful points in your argument, but see this is what bugs me about these advanced stats.

You're essentially saying Pittsburgh's PDO will drop, but they'll weather the storm better than Calgary because they're a better team. How is it then that PDO, if looking at it as a purely predictive tool as some writers use it telling us anything?

"LA has a low PDO, but that will improve because they're a good team."

"Buffalo is a horrible team, their PDO clearly shows that. "

" Calgary has a high PDO, but don't be fooled, they're not that great of a team and they'll come back down to earth."

Really, is any of this saying anything we don't know? That good team will improve in the standings and bad teams will drop? Do we need an advanced stat to tell us that?

I'd also argue with you that Pittsburgh is a bad team being propped up by two generational talents and that Calgary is a much better built team. But that is for regular old fashioned hockey analysis and observation, it has no room in this thread.

You actually answered your own question in the first few words. Cherry picking. Beware people using a single stat to make points like this.

Calgary has a huge PDO despite poor underlyng fundamentals. So too does Tampa Bay. Pittsburgh isn't as bad, but are still a bit out of shape. If the underlying stats do not change, all three will plunge relative to the league in PDO. But of the three, Pittsburgh is in the best shape.

The flip side is a team like Chicago, which has extremely strong underlying figures (i.e.: best Corsi For% in the NHL), but an absurdly low shooting percentage of 5.94% at 5 on 5 (dead last). Chicago is on pace to take 2378 shots at 5 on 5 and score only 141 goals. If they were to rise to that same 8.49% for 10th place at the end of the season, they would score 202 goals on the same number of shots. And that +61 difference from now to the end of the season is a full goal per game.

So when I look at the numbers as a whole and view in the lens of PDO, I see teams like Calgary and Tampa that should drop a fair bit, Pittsburgh who should drop a little, and Chicago, who should jump significantly. But in all cases, these predictions would change if the first three possess the puck more, and if Chicago controls it less.

[Fan in Exile]

The thing is - as Text Critic pointed out above - the Flames are improving as it relates to PDO increasing their shots for and cutting down shots against. So it may well be that the PDO regresses towards the mean but the Flames are also trending towards better possession stats.

It doesn't take advanced stats to establish that the Flames stole some games early in the year, against Chicago in particular. And the Flames have built confidence, have started believing in themselves and have played better and better as the year goes on.

I like what I see when I see the Flames play. I don't expect them to win the President's Trophy and the Cup but I can't imagine a total collapse either. I would have been more inclined to accept Mirtle's analysis and predictions ten games in looking at the next ten but much much less so now.

[Parallex]

Originally Posted by _Q_

You're essentially saying Pittsburgh's PDO will drop, but they'll weather the storm better than Calgary because they're a better team.

No, Pittsburgh's PDO will drop, but they'll weather the storm better than Calgary because they don't rely on the abherrent percentages to acheive results. When percentages normalize they'll still win games because they regularily outshoot their competition, Calgary doesn't.

As for what it's good for it's a good tool to explain how short term spikes and dips in percentages can leave a team looking better or worse by W-L standing then they actually are.

[bob-loblaw]

What were our stats like after January 25th of last year for the rest of the season?
If the season started then we would have made the playoffs last year. We've just continued playing well since then. Cammy got hot after the trade deadline and our goaltending was Ramo and Ortio.

It's easy to say it's unsustainable, but the line-up has changed by about 30% and our defence and goaltending is arguably better. The fact of the matter is that we've been playing better much longer than people think so perhaps we should be able to keep it going.

As far as advanced stats goes, and I'll admit I don't know much about hockey stats but I do know something about statistics, and I just think there's way too many variables in every hockey game to make some stats meaningful with a high degree of correlation.

[BACKCHECK!!!]

PDO is a descriptive stat, not a predictive one.

There is a fallacy that many people repeat, which claims that PDO strongly regresses to the mean over large sample sizes.

This is not correct. Good teams routinely carry high PDO's over entire seasons, or even several seasons. Poor teams do the opposite.

There is nothing inherently unsustainable about an outlying PDO.

[Itse]

Originally Posted by Resolute 14

So when I look at the numbers as a whole and view in the lens of PDO, I see teams like Calgary and Tampa that should drop a fair bit, Pittsburgh who should drop a little, and Chicago, who should jump significantly. But in all cases, these predictions would change if the first three possess the puck more, and if Chicago controls it less.

So... Calgary and Tampa are probably not Presidents trophy contenders, Chicago could be despite a poor start, and Pittsburghs goal differential is probably not really going to reach the levels unseen since the Oilers Dynasty days?

Wow.

Where would we be without all the new info advanced stats give us.

[azzarish]

Absolute rubbish.

How the Leafs performed last year has absolutely no bearing whatsoever on how the Flames are going to perform for the rest of the season. There could have been many reasons in Toronto's downturn last year that don’t have to equate to this Flames team. It seems like some of the media think this start is a fluke. What they seem to forget is that we're missing 4 or 5 key players and still performing at this level. I have been watching the Flames since 2000 when I first moved here and this season for the first time in about 5 or 6 years they look like a team who is going to compete on a nightly basis, and one that could sustain some semblance of this form for the rest of the season. This is also the best defensive core that I think I've seen over that time. The most noticeable aspect this season is that there are more key pieces in the team to cope with other key pieces either; 1) getting injured or 2) losing form. There are more players that stand out to me that have done in a good few years. Which means there are more difference makers, and if you have difference makers you are more than likely going to win more games. There a quite a few more players that seem to stand out every game than we had last year, or the year before.

Yes, we've fallen away before and been left disappointed but I think there is a different mentality in the team now. We've finally seemed to have eradicated all elements that were rotten to the core that were left overs from the Sutters/Iginla period. We've moved on from that era now, and this is a new team, and the new version of the Calgary Flames. We've built a new identity and the signs from the play last year, coupled with the start this year are very promising indeed. I suspect we're not going to fade away as easily as some of these so called experts are expecting.

[Street Pharmacist]

Originally Posted by BACKCHECK!!!

PDO is a descriptive stat, not a predictive one.

There is a fallacy that many people repeat, which claims that PDO strongly regresses to the mean over large sample sizes.

This is not correct. Good teams routinely carry high PDO's over entire seasons, or even several seasons. Poor teams do the opposite.

There is nothing inherently unsustainable about an outlying PDO.

You're wrong, though. I suppose it depends on your qualifier on "outlying", but almost universally a PDO that is significantly higher than 1 will regress to the mean. That's not fallacy, that's fact.

For example. Take any PDO over 1.02 for a season. Now add the season before it. Guess what, whatever sample you chose will now be lower. The top PDO and bottom PDO in the league all get smaller as the sample size gets bigger.

If you are calling a PDO of over 1.02 an outlyer, then yes, for one season that's possible. Colorado did it last year. How did that work out?

[calumniate]

Originally Posted by Street Pharmacist

You're wrong, though. I suppose it depends on your qualifier on "outlying", but almost universally a PDO that is significantly higher than 1 will regress to the mean. That's not fallacy, that's fact.

For example. Take any PDO over 1.02 for a season. Now add the season before it. Guess what, whatever sample you chose will now be lower. The top PDO and bottom PDO in the league all get smaller as the sample size gets bigger.

If you are calling a PDO of over 1.02 an outlyer, then yes, for one season that's possible. Colorado did it last year. How did that work out?

That's like saying a teams' ERA was good one year and did well, but then the next year it was not as good and they didn't do as well. On average they will do 'okay'.

[_Q_]

^^ Colorado had an incredibly good season last year and as a result had a high PDO. This season they're terrible and have a low PDO. I fail to see how it predicted anything.

It seems to me that good results drive high PDO numbers and not the other way around.

[Street Pharmacist]

Originally Posted by calumniate

That's like saying a teams' ERA was good one year and did well, but then the next year it was not as good and they didn't do as well. On average they will do 'okay'.

No, because good ERA is much more reproducible than PDO....

[Beatle17]

Originally Posted by Tinordi

When criticising advanced stats please continue to demonstrate limited or no understanding of the stats themselves and the arguments as to why they're useful.

Please give us a condescending lecture about not understanding something, with no corresponding ability to explain to the uneducated how they work. Or better yet, please explain how advanced stats, that are reporting on past games, will tell us who is going to win all the games/cups so we can get our bets down in Vegas.

[wingedwheel]

I haven't watched any Flames hockey this year. My team is in the East and just lost to the Habs. Habs are a quick team with good goaltending, but they are going to fail. They won't finish first in the East. Just because this article is about the Flames "failing" doesn't mean no other team is going to falter.
That being said, the Habs made the playoffs and had a legitimate shot at winning the EC last year, so there is more proof that they can continue to win (with sustainable numbers or not) than there is for the Flames.
I'm hoping Flames keep it up all year. My first winter here in Calgary and a playoff run would be great