Better Late than Never: Colorado Avalanche

By now, you've probably gotten the hang of this series…

Not much was expected from the Colorado Avalanche this season. They were coming off a 69-point season, last in the Western Conference and 28th overall in the NHL. But they jumped out of the gate 9-2-3 to start the season, which gave them a head start on a postseason berth. They played close to .500 hockey for the next 55 games (25-21-9) and then, just to instill confidence for the playoffs, went on an 0-7-6 slide to close the season, finishing four wins above .500. In the first round of the playoffs, the Avs redeemed their late-season losing streak and gave the powerhouse San Jose Sharks a run for their money, eventually falling four games to two. All things considered, 2009-10 was a nice season for the Colorado Avalanche. Right?

The truth is that it was a very lucky season for the Avs. Colorado allowed so many shots against at even-strength that they ended up with a historically bad Fenwick % for a post-lockout team:

Team	Season	Fenwick%	W	L	T
LA	2007-08	45.9	32	43	7
NYI	2008-09	45.9	26	47	9
COL	2009-10	45.8	43	30	9
FLA	2009-10	45.8	32	37	13
EDM	2009-10	45.6	27	47	8
PHO	2008-09	45.5	36	39	7
PIT	2005-06	45.0	22	46	14
ATL	2007-08	43.0	34	40	8

They were saved by very good goaltending and one of the highest team shooting percentages in the last five seasons. In general, goaltending is, to some extent, sustainable: Colorado made a brilliant decision to sign the clearly-talented by unheralded Craig Anderson to a value contract. But high shooting percentages tend to crash to earth after a season-long high.

Now it’s true that Colorado’s team talent isn’t captured by what happens in general to our group of “average” bad teams. Teams with winning records have the lead a lot and tend to allow more shots than they take, and vice-versa, and they also take higher-percentage shots than their opponents.

That last statement is the assertion that was most commonly-associated with Colorado, and we can investigate it in several ways. We have recorded shot distances for every shot taken by Colorado and their opponents. And we have a database of scoring chances for 45 of Colorado’s regular-season games, primarily from the first half of the season, and all six of their playoff games. If a Colorado shot was more likely to be a high-quality scoring chance, then we’d expect to see them shooting from closer to the net than their opponents, and we’d expect them to be closer to outchancing their opponents than their Fenwick totals would suggest.

And Fenwick percentage does, in fact, systematically under-estimate Colorado’s performance – they got the puck closer to the goal than their opponents when they shot, and their scoring opportunities were proportionally more dangerous than their opponents’:

	Pct
Fenwick	45.0
Shots<30 ft	45.8
Chances	47.8

The Avalanche look better in this analysis, but owning 47.8% of overall scoring chances is not a good figure. Edmonton was the only Western Conference team to post a total that was clearly worse. Indeed, Colorado had virtually no top-flight talent on their roster. After they traded away Wojtek Wolski, they had – at most – two forwards and three defensemen who could play against another team’s top lines without giving up a disastrous number of shots:

Player	Age	Chance%	Ozone%
Hannan	31	48.5	42.3
Quincey	24	46.1	42.8
Stastny	24	50.6	45.2
Hejduk	33	43.8	47.5
Foote	38	46.4	48.2

Colorado’s vaunted “young players” didn’t figure into the mix on the top unit. They saw average competition and produced somewhat below-average results:

Player	Age	Chance%	Ozone%
O’Reilly	18	44.6	43
Cumiskey	23	52.7	47.9
Galiardi	21	47.3	48.4
Duchene	19	49.6	50.3
Stewart	22	44.0	51.8

The older players facing middling competition didn’t fare much better:

Player	Age	Chance%	Ozone%
McLeod	25	44.6	51.9
Hendricks	28	42.2	44.2
Svatos	27	50.9	54.7

At the bottom of the roster, the Avs also had three defensemen who needed to be sheltered to get even mediocre performance out of them. Liles, in particular, was a catastrophe:

Player	Age	Chance%	Ozone%
Wilson	22	45.9	51
Clark	33	51.4	52.1
Tucker	34	51.2	54
Liles	29	44.3	56.2

Overall, the first unit out-performed expectations, while the other two groups did not:

	Chance%	FO%
Top Competition	47.1	45.2
Avg Competition	47.0	49.0
Sheltered	48.2	53.3

Colorado opened the 2009-10 season with four NHL players who’ve showed the ability to match up against top competition: Stastny, Hannan, Quincey and Wolski. As we’ll see from their performance against the San Jose Sharks, that’s not enough to get the puck moving in the right direction against a good team.

Playoff Series vs San Jose

Colorado’s 4-2 loss to the Sharks in the first round of the playoffs was their season in a microcosm. They took 36% of the shots in the series – and scored 37% of the goals. And this time, scoring chances at even-strength tell the same story:

Game	COL	SJ	Result
1	9	10	W 2-1
2	10	13	L 5-6
3	5	19	W 1-0
4	12	27	L 1-2
5	11	19	L 0-5
6	9	17	L 2-5
Total	56	105	11-19
Pct	34.8

In retrospect, this series doesn’t look close: Colorado was thoroughly dominated, particularly after Game 2. But while the series was in progress, the conventional wisdom was that the Avs had the Sharks’ number – they won two of the first three games, after all!

Even though those two wins came via fluky goals – Chris Stewart scored from behind the goal line in Game 1, and the Sharks shot the puck into their own net in Game 3 – the “Sharks choke again” story was already being written by journalists nationwide. One contention was that even though Colorado had one third of all goals, one third of all shots and one third of all chances, their chances were still of higher quality than the Sharks chances. If this were true, then we’d expect to see it in the chance breakdown:

Chance Type	COL	SJ
Dump-In	17	16
Ftc/Gv	14	13
Skate-In	11	38
Rebound	2	17
2-on-1	0	3
3-on-2	5	7
Breakaway	1	2
Faceoff	3	4

‘Ftc/Gv’ refers to situation where the defensive team either failed to clear the zone or gave the puck away directly. As with all observed data, this table isn’t a perfect assessment of what happened, but it tells us a lot about the differences between Colorado’s style and San Jose’s. For example:

– Colorado was twice as likely as San Jose to dump the puck in order to gain the zone

– Colorado generated virtually no chances off rebounds

– San Jose’s defense had a lot of trouble breaking out; even though Colorado rarely had puck possession in the Sharks’ zone, they got just as many chances from defensive miscues

– There were virtually no odd-man rushes during the series, and San Jose had 2/3 of the few that there were

Overall, Colorado lacked the skill to carry the puck into the offensive zone but they could dump-and-chase and take advantage of San Jose’s relatively weak defensemen.

But that wasn’t enough. Colorado’s top line did extremely poorly at controlling chances against San Jose’s top line:

Player	Chance%	Ozone%
Hannan	41.1	38.9
Quincey	42.3	39.2
Stewart	36.7	41.4
Galiardi	34.5	44.3
Stastny	33.9	47.8

If this playoff series was Colorado’s year in a microcosm, then one particular shift was Colorado’s series in a microcosm. In the second period of Game 3, the Avs sent out their top line against Joe Thornton, Patrick Marleau, Manny Malhotra, Dan Boyle and Douglas Murray. They didn’t get the puck out of the zone for 1:45 – the Sharks put 15 players on the ice before Colorado was able to change their defensive pairing.

Not that the players who drew the Sharks’ second line did any better:

Player	Chance%	Ozone%
Hejduk	40.0	37.5
Cumiskey	24.2	39.6
Foote	28.1	41.2
Duchene	36.9	47.9
Yip	23.5	61.5
Svatos	37.9	66.7

Brandon Yip may have scored two goals in this series, but he certainly gave it all back at the other end of the ice.

The bottom of the roster looks even worse – only Ryan O’Reilly fared well against San Jose’s 3^rd and 4^th line.

Player	Chance%	Ozone%
Yelle	31.6	31.3
O’Reilly	37.5	35.6
Tucker	28.1	45.8
McLeod	29.2	55.6
Hendricks	40.9	60
Liles	37.2	65.8

John-Michael Liles’ performance was pure ugliness despite complete sheltering. His contract, which pays him $9M over the next two seasons, is one of the underrated bad deals in the league.

The Bottom Line

The bottom line is that Colorado did not play very well this season. They were better than their underlying numbers appear, but substantially worse than the standings would suggest. Craig Anderson will keep them from crashing completely to earth next season, but this was a 12^th– or 13^th-place team that rode a wave to a playoff berth.