Calgarypuck Advanced Stat Summary

[edslunch]

Originally Posted by Street Pharmacist

No, you're missing the point. Their 5v5 is driving the team, but their 5v5 is bucking the corsi expectations. It's still 5v5 that is driving success

I don't know. The Flames are 15th in the league for 5-5 goal differential, just barely positive - sounds like 5-5 is not hurting but hardly driving their success. On the other hand a stunning 25% of their wins have come in 4-4 OT - seems like that's a bigger differentiator vs other teams.

[sworkhard]

Originally Posted by edslunch

I don't know. The Flames are 15th in the league for 5-5 goal differential, just barely positive - sounds like 5-5 is not hurting but hardly driving their success. On the other hand a stunning 25% of their wins have come in 4-4 OT - seems like that's a bigger differentiator vs other teams.

The flame are 14th in the league in standings (and evening up the games played won't move them more than 2 places higher), so it seems that their place in the standings is exactly where you'd expect it to be given their 5v5 goal differential.

[Street Pharmacist]

Originally Posted by edslunch

I don't know. The Flames are 15th in the league for 5-5 goal differential, just barely positive - sounds like 5-5 is not hurting but hardly driving their success. On the other hand a stunning 25% of their wins have come in 4-4 OT - seems like that's a bigger differentiator vs other teams.

If you look at it that way, yes. My point is that the Flames play 5v5 for the vast majority of the time, so it's their play in 5v5 by bucking the possession game, that has allowed for the success.

But yes I'd agree that special teams and 4v4/OT have really made a difference in their wins.

Semantics I suppose

[edslunch]

Originally Posted by Street Pharmacist

If you look at it that way, yes. My point is that the Flames play 5v5 for the vast majority of the time, so it's their play in 5v5 by bucking the possession game, that has allowed for the success.

But yes I'd agree that special teams and 4v4/OT have really made a difference in their wins.

Semantics I suppose

Fair enough

[PeteMoss]

Originally Posted by Resolute 14

The system, basically, is to either outscore or outsave your possession stats.

Over the last four full (82-game) seasons, seven teams have made the playoffs with a CF% below 48. All but one did it through amazing goaltending, amazing shooting, or both:

Season  TM   CF%  Sv%  Rk   Sh%  Rk
2013-14 COL 46.9 .930 ( 6) 8.80 ( 2)
2013-14 MTL 46.7 .930 ( 5) 7.44 (21)
2011-12 NYR 47.7 .927 ( 8) 8.30 ( 8)
2011-12 NAS 46.6 .925 (12) 8.51 ( 6)
2010-11 ANA 44.4 .923 (16) 7.79 (20)
2009-10 MTL 47.2 .929 ( 3) 7.58 (25)
2009-10 COL 44.8 .926 ( 8) 8.84 ( 4)

2014-15 CGY 44.8 .922 (16) 8.88 ( 2)

(percentages and ranks at 5 on 5, per hockeyanalysis.com)

That Ducks team in 2010-11 had no business being anywhere near the post-season...

Montreal is the poster child for the system that defeats Corsi. It's called "Have a gold medal winning goaltender". The Habs are 21st in CF% this year at 49.4, 26th last year at 46.7, 7th in 2012-13 at 52.8, 22nd in 2011-12 at 48.0, 11th in 2010-11 at 51.6 and 27th in 2009-10 at 47.2%.

Goaltending is reasonably predictable from season to season, so it can consistently overcome poor possession. Calgary is in a playoff spot because of shooting percentage, and honestly, it is far more random. Washington, for instance, has ranged anywhere from 10.3% to 7.6% between 2009-10 and this season. And a 3% swing on 1800 shots is 54 goals. They are an extreme though.

Interesting from that list that Anaheim seems to be a team that finishes better than their advanced stats would say on a fairly regular basis. They've changed coaches but still seem to finish higher than expected based on Corsi.

[johnnybegaudreau]

A couple of things to clear up. First of all a correlation merely shows that there is or is not a trend. It can show wether that trend occurs positively or negatively. It does not "PREDICT" anything. Just that there's a trend between 2 variables.

I think what the majority of the adv stats gurus are advocating here is not that there is "never" any exceptions to the rule. OR that a team with a low CORSI for and high PDO can't make the playoffs -ever.- There is always exceptions to the rule in anything. That's the beauty of stats is they don't predict things with 100% certainty. Look at your standard bell curve there is no 100% category it goes to 99.9999999999... Nothing does predict out at 100% certainty. You can be an EMT and do the exact same procedure on 2 people who have the exact same personal stats and one may die and one may live. It's the nature of being immersed in our own worlds and not seeing the bigger wider picture of all the saved lives versus all the lives lost for that procedure over a 20 year time period.

In research/stats there's two types of studies. Cross-sectional which have predictive power just not as high a power as a longitudinal study. Looking at one season is analogous to a cross-sectional study. Longitudinal study would be like looking at 10 seasons. Logitudinal has the most predictive power of any study.

The point Adv. Stats gurus are making is that using adv stats is a good way to be cautious and do your diligence. If you take the Flames this season for example. Yes. They're the exception to the rule. However, examples like Toronto and Colorado show why it's important to consider what those numbers mean. So as not to make the mistake of those teams and go out in the Market and make rash or unrational UFA signings and or rental signings based on the hype and luck of this season for the playoff push. That's really why Toronto, Colorado, etc have fallen to where they are. Instead to stay the course with the plan that has been set out in building from the ground up. If you find value in the UFA market or can make a trade for value with a team then you feel comfortable doing that.

and that's what adv stats gurus are advocating is that Stats show trends and have predictive capabilities. The flames are in a situations where the stats show they could end up like Toronto/Colorado but they dont' have to.

[Resolute 14]

Originally Posted by edslunch

I'm confused. You and Resolute make the point that 5v5 is by far the driver for every team, yet 5v5 stats don't line up with the Flames' (or the Oilers' recent) success. Do we conclude that they are outliers doomed to fall back or do we look at something else to explain it?

Mostly, I am arguing what is believed by the analytics community. They have found that 5 on 5 play correlates most strongly to success. As JohnnyB notes, that does not predict the future though. A lot of advanced stats guys really wish it did though, because they tend to take outliers as personal insults, despite the fact that two or three teams will qualify for the playoffs every year with CF%<50.

It can be a dangerous game to look at one stat (Corsi 5 on 5, for instance) and try to make all kinds of judgments. I hadn't considered this until responding to the earlier question, but you do raise a good point about how our 5 on 5 scoring stats don't seem to match up with our success. The largest reason why our special teams scoring stats look so good is the fact that we take so few penalties 5 on 5. That is hidden by most analyses of Corsi.

We have played by far the fewest minutes 4 on 5 - 214mins. The median for the league is about 280 minutes. At our season rate of 6.73GA/60 minutes shorthanded, that would translate to 7-8 more goals against if we were closer to average in terms of penalties taken. So instead of a goal differential of +6 combined between 5 on 4 and 4 on 5, we would instead by -1 or -2. And those extra goals would probably turn two or three wins into losses. And given how razor thin our margin is for making the playoffs, that alone could be fatal.

[Resolute 14]

Also, for those who were upset by Yost's commenting on the Flames last week, he picks on Winnipeg this week. Notably, he argues that Tyler Myers' defensive partners were better off without him than with:

http://www.tsn.ca/jets-taking-risk-b...myers-1.209396

I rather wish he had done that comparison for Bogosian in Winnipeg too.

[GranteedEV]

Originally Posted by johnnybegaudreau

That's really why Toronto, Colorado, etc have fallen to where they are. Instead to stay the course with the plan that has been set out in building from the ground up.

Colorado's problem wasn't rushing ahead, it was failing to retain Statsny.

You don't need advanced stats to know that was a dumb move.

[Bingo]

Advanced Stats update for March 10th.

Will do nothing to quell the Flames as the luckiest team in professional sports sentiment!

[edslunch]

Interesting how the spread between Corsi and Fenwick increased dramatically in March. Holy shotblocking!

[Bingo]

really amps up the debate about shot blocking as an art or skill versus a team in trouble

[CorsiHockeyLeague]

Seriously though those are Buffalo-like corsi numbers for March. Wow that's bad.

By the way, the individual teams graphs at the bottom, I'm assuming those are the Flames' numbers vs those teams?

[Street Pharmacist]

Originally Posted by CorsiHockeyLeague

Seriously though those are Buffalo-like corsi numbers for March. Wow that's bad.

By the way, the individual teams graphs at the bottom, I'm assuming those are the Flames' numbers vs those teams?

Those are against the opposition, yes. At least if I read it correctly.

Interesting to note, bingo has a list of Calgary's best and wrist corsi showings. They've lost almost all of their best ones and won most of the worst. It makes no sense. Absolutely dominated Montreal and Buffalo in every facet but goaltending and lost. Got ragdolled by LA and won

[EldrickOnIce]

Luck can kiss my ***, but those numbers are brutal.
Mind you, I guess without Giordano it wasn't going to be good.
Need to keep getting results, so need to be a lot better.

[Bingo]

those numbers are 5on5 and not all, they get a bit better but not great if you get away from 5 on 5 only

Yes the teams at the bottom are Flames versus.

Here is the top 5 lists ... seeing a big change between corsi and fenwick now with the shot blocks on the road trip

Corsi
1 BUF Dec 11
2 BUF Jan 27
3 NYR Dec 16
4 TOR Dec 09
5 N.J Feb 25

66 CHI Oct 15
65 L.A Feb 12
64 DET Mar 06
63 DAL Dec 19
62 BOS Mar 05

Fenwick
1 BUF Jan 27
2 EDM Jan 31
3 N.J Feb 25
4 BUF Dec 11
5 MTL Oct 28

66 CHI Oct 15
65 L.A Feb 12
64 VAN Jan 10
63 PIT Feb 06
62 L.A Dec 29

PDO
1 CAR Oct 23
2 ANA Nov 18
3 EDM Oct 09
4 EDM Dec 27
5 ARI Nov 13

66 BUF Dec 11
65 PIT Feb 06
64 ANA Nov 25
63 CAR Nov 10
62 ANA Jan 21

[Resolute 14]

Originally Posted by EldrickOnIce

Luck can kiss my ***, but those numbers are brutal.
Mind you, I guess without Giordano it wasn't going to be good.
Need to keep getting results, so need to be a lot better.

Without Gio, all on the road, against a few Corsi darlings. And somehow, 4-2-1. Crazy.

[CorsiHockeyLeague]

Originally Posted by EldrickOnIce

Mind you, I guess without Giordano it wasn't going to be good.

Mark "Delta Corsi Impact" Giordano. The team's performance with and without him on the ice is a good argument for him being the best defenseman in hockey.

[Street Pharmacist]

There's some discrepancy with the "eye" test too. The flames did NOT get wildly outplayed by Detroit despite it being our 3rd worst corsi. We won the game and even Babcock said that Calgary didn't give them anything. The flames allowed almost no shots in the third, but lots of "corsi" events from the point

[bomber317]

dumb question here, are the stats taken from what is recorded in-game?

For example, we all know LA @ home has a discrepancy for shot counting. So will that affect the corsi rating for that game?