Monday, January 09, 2006

 

Flames Midpoint Evaluation

Nice. Real nice.




What, you want more? Uh, OK. After going 4-7-2 in October, the Flames then went 21-5-2 over the next 28 to hit the halfway point. That's wins in fully three-quarters of their games, and by my count, 18 were in regulation: only 2 were in shootouts, plus the OT win over Ottawa.

Some of the underlying statistics are frankly awful. A sample, with a bit of mediocrity thrown in:
There's a few impressive stats too, I suppose:
And some aspects of their record are simply phenomenal:
I'm also very happy with the less measurable things. Even when Jarome is bad, he's good: he's an excellent defensive forward who hardly ever makes mistakes. Robyn Regehr is giving Lidstrom and Pronger a run as the best defenseman in the Western Conference. Jordan Leopold might be the greatest goal-less player in NHL history. Phaneuf makes mistakes, but he keeps getting better. Ference is finally proving that he belongs, and is giving up the puck much, much less than in the early season.

Tony Amonte and Daymond Langkow have become Calgary Flames. They didn't really fit early on, but they're with the program now, and they're part of the juggernaut. Yelle and Nilson are as good as ever. Huselius and Lombardi give us something we didn't really have last season, and that's hands that can make tough goals look easy, and can create something out of nothing.

The Flames are on track for somewhere between 110 and 115 points, which will translate into 1st or 2nd spot in the conference, and most likely home-ice advantage for as long as they keep winning. Times are good in the C of Red, and there's no good reason to believe that they won't keep rolling. Go Flames.

Comments:

Holy crap! You got some extra time on your hands, or what? I should get you to do my taxes.

A simple "Peca sucks" would do.

Great posts, just the same, but I think the only ststs that really matter are WINS, LOSSES, and POINTS.
 


I was just going to mention that it's refreshing to see a Flames fan take note of this stuff and furrow his brow worriedly.

Great series of posts, really interesting stuff. I had no idea that Ottawa had fallen off the pace so much.

jhuck, the point of it all is that wins, losses and ties follow from the stats that Matt has cited. While no one would argue that they'd rather have a team that had 80 points at the end of the season and a goal differential of +40 as opposed to a team that had 100 points at the end of the season and a goal differential of 0, for assessing a team's strength (which for most fans is more interesting than arguing over how many wins a team has), goal differential is better than wins.
 


Worried? You're being sarcastic, right?

The Flames G.D. was -11 after that 4-7-2 start (it was -8 after their 1-3 start). Now it's up to +7; that's a decent swing for their record over that period.

Plus, I just can't account for why the Flames can't, or won't, blow anyone out, ever. Most teams have at least a handful of these to prop up their average at the top end, the Flames do not.

As "things to worry about" go, that's low on the list. We'll see how it all goes in the 2nd half.
 


Worried? You're being sarcastic, right?

I like to say "furrow his brow" and will leap at any opportunity to do so.

The Flames G.D. was -11 after that 4-7-2 start (it was -8 after their 1-3 start). Now it's up to +7; that's a decent swing for their record over that period.

Well if we're going to use artificial end points, the Oilers are now at +9 after being -14 through 10. The Flames may be good, but they're no Edmonton Oilers.

In any event, they're +18 and 21-5-3 since the 4-7-2 start. How many points (old school points) should you expect from a +.64/game goal differential over 28 games? My math says 34-the Flames have actually earned more than 40+ traditional points (if the league awarded no point for an OTL and no bonus point for an SOW) during this stretch. I don't think that the goal differential not supporting the record can be attributed to the Flames struggles early on in the season.

Plus, I just can't account for why the Flames can't, or won't, blow anyone out, ever. Most teams have at least a handful of these to prop up their average at the top end, the Flames do not.

It's entirely possible that the Flames are unique in hockey history and that Sutter is some sort of genius to whom the laws of goal differential do not apply. My personal belief is that a team that can't or won't blow teams out just isn't an excellent team. I've looked long and hard for evidence that teams can outperform their goal differential though, and I can't find it. Teams simply don't achieve at the level that the Flames are achieving at year after year.

Does Sutter's record disclose any special aptitude in this regard? He's been the head coach of a team for a full season 9 times, 8 if we ignore the lockout shortened 1994-95 season. Here are the differences between his team's expected and actual points for those seasons in which he was head coach for the whole year:

2003-04: 0
2001-02: -2
2000-01: 1
1999-00: -6
1998-99: -4
1997-98: -2
1993-94: -2
1992-93: 7

To sum up, the Flames, as of today, are on pace to exceed their expected traditional points by about 12 or so. Nothing in Sutter's history or hockey history suggests that this reflects anything but a team enjoying a bizarre one year positive aberration. It's not explained by the early season struggles.

I'd be furrowing my brow if I was a Flames fan, although I'd figure that Iggy will start rolling at some point.
 


Mudcrutch

"The Flames may be good, but they're no Edmonton Oilers."

If you're trying to convince Matt that the Oilers are not as good as the Flames, you're too late. Total points and the season series results so far have confirmed that already.

Not sure where you grabbed the "expected points" regarding Sutter's years of coaching from (expected by whom?), but following the 2001-02 season Sutter became only the second coach in NHL history (Al Arbour) to improve his team's point total for five consecutive seasons.

The Flames had a good year last year (better than the Oilers) and have improved again this year.

As for good or great teams having a small goal differential....

2003-2004
Boston 82(gp) 41(w) 19(L) 15(T)
7(OTL) 104(P) 209(GF) - 188(GA)

1999-2000
Dallas 82(GP) 43(W) 29(L) 10(T)
6(OTL) 102(P) 211-184(GF/A)

The late 90's saw the perrenial powerhouse Colorado finish with a +30ish goal dif. three times.


I don't know if Calgary is an "excellent" team, but they're damned good. They can win without Iggy at his best, or with half the team injured. It's not that they're so much better than other teams (which they aren't, skill-wise), but rather that they're very, very tough to play against and beat.

Plus, one-goal games are way more exciting than blowouts.

Winning, not goal differential, is the Flames' strength.
 


If you're trying to convince Matt that the Oilers are not as good as the Flames, you're too late. Total points and the season series results so far have confirmed that already.

Last I checked neither of those were finished. The Flames have played 4 of the 5 BoA games at home and have a 3-2 lead in the series? Who cares. Neither of those are determinative.


Not sure where you grabbed the "expected points" regarding Sutter's years of coaching from (expected by whom?), but following the 2001-02 season Sutter became only the second coach in NHL history (Al Arbour) to improve his team's point total for five consecutive seasons.

There are a bunch of formulae out there to produce expected points. I use one that is based on goals being scored according to a poisson distribution, which has been historically more accurate than the Bill James version of it.

As for the point about Sutter increasing his teams points it has no probative value to a discussion of whether or not the Flames points are consistent with their goal differential. That's the question at hand. As I stated previously, hockey history doesn't suggest it's possible for teams to continually outperform their goal differential. Why you'd raise the whole "increased his teams points for five years" is beyond me.

As for good or great teams having a small goal differential....

2003-2004
Boston 82(gp) 41(w) 19(L) 15(T)
7(OTL) 104(P) 209(GF) - 188(GA)

1999-2000
Dallas 82(GP) 43(W) 29(L) 10(T)
6(OTL) 102(P) 211-184(GF/A)


The Bruins went out in the first round of the playoffs to a team with nearly the same goal differential, despite finishing 11 points back. The Stars team you've listed is interesting-they had the good fortune not to have to face StL or DET in the playoffs because they got upset.

Honestly though, the two examples you've posted are proof of nothing. I can post a list of hundreds of teams who finished within 5 points of their expected points. Goal differential matters-show me some teams that managed to put together two or three seasons of significantly outperforming or underperforming their expected points. It never happens.

Winning, not goal differential, is the Flames' strength.

Well, we'll have to see where they sit at the end of this year and where they sit at the end of the season after. I'd be willing to bet that if their points are significantly more than their goal differtial suggests this year, it won't be next year.
 


First of all, Mud is basically right. In accordance with years of numbers than show that 2.8 goal diff = 1 point, the Flames have earned more points than their ability to score and prevent goals. It's not a situation that's particularly tenable or repeatable.

Second of all (that said), this same Poisson model is more limited this year than ever because of the SO. We suspect W/L in the shootout is roughly 'random', but we don't really know yet; we suspect being tied after 60 is a 'random' occurrence as well, but we don't really know that. I like it, but the reductio is slightly absurd to think that the Flames record would 'suit' their GD better if more of their wins had been in shootouts..

But third and most important, the reason I'm not worried is that while you can't escape your GD, you can certainly improve it! I don't really think the Flames are the 28th-best scoring team in the league as of today. Their D and Kipper are still excellent. As such, I think they're going to outscore most of their opponents going forward, and probably have a better GD number while doing it...
 


ha

When you were talking about "expected points" under Sutter, I thought you meant something in regards to total points (W-L-T). Hey, it was late and I just watched a boring debate on TV - so my brain was shut off.

Anyways, talk all you guys want about goal dif. As long as the Flames keep winning I'm happy. They went to the 7th game of the Stanley cup finals after beating out 3 very, very good teams who each had good goal differential. This year they're first in a tough division and have beat a lot of good teams to get there. Obviously goal dif. isn't a strength that's as important to the Flames as it is to other teams. Defence, goaltending, checking back and forth, boardwork, hard work, etc. etc. are all important strengths that teams need in order to win.

As for other teams that don't have exceptional goal dif. right now yet are doing quite well.....
Buffalo +12 (56 points), Toronto +7 (51 points), Nashville +13 (55 points), Vancouver +6 (51 points), Edmonton +10 (50points),
L.A., Colorado and Carolina are around +20.
Calgary shouldn't be any more worried than any of those teams.
 


Poisson? I'll pass on the seafood. If you just do a simple linear regress of current points against GD for the whole league, it's easy to spot the major outliers. Biggest overperformers right now are Nashville (+6.1 pts) and Calgary (+6.0 pts). Informally I would regard these teams as due for a downturn, but if you assume that the excess "luck" is nonrandom and heavily dependent on good goaltending, you can exempt Calgary.

Edmonton is right on the dot with points vs. GD: no reason to think they aren't as good as they look.
 


Informally I would regard these teams as due for a downturn, but if you assume that the excess "luck" is nonrandom and heavily dependent on good goaltending, you can exempt Calgary.

What makes a team with good goaltending and poor offence any more likely to outperform than a team with good offence and poor goaltending?
 


Shootouts might have that effect. Think of a model where every team's top 3-5 shooters are about equally good, no matter how good the team's overall in-game offense is. Somewhat credible, no? In that event, if there were a persistent nonrandom influence contributed by shootouts, it would be almost 100% goalie-dependent.
 


That said, Calgary is only 2-3 in shootouts, so obviously you can't exempt them on those particular grounds.
 

Post a Comment

<< Home

This page is powered by Blogger. Isn't yours?