So, About Those Goalies: I Was Wrong Edition

For a while now, there have been a lot of questions about the Lightning goaltending, and I've been as obsessed with it as anyone. Maybe more. And whenever someone has said that this isn't the way Dwayne Roloson was last season, I've been skeptical because I've seen him struggle before. I've held onto the idea that this slow start was explainable by Roloson's history of inconsistency and the way the goalies have been "hung out to dry" so often by the team this year

But without looking at the data, none of us could really be sure whether what we thought we were seeing was what was really happening. So I bit the bullet. Even though numbers are so very much not my thing (side note: I hate making data tables; tables suck) I looked for the data, and it turns out that I was wrong, but I was sorta right, too.

Dwayne Roloson's stats are, in fact, down from the numbers he put up with the Lightning last season. Between January and April of 2011, Roli played in 34 regular season games. In that time, he gave up 85 goals on 967 shots. So his SV% was .912 and his GAA was 2.56. So far this season, in 14 games, he has a SV% of .887 and a GAA of 3.46 (42 goals against on 346 shots). That's a significant drop. Yes, I was wrong.

But Roloson has a reputation as a slow starter, so I looked up how he did in October and November historically. (The data here includes all October and November games, while the data for 2011-12 only includes games through November 26, because (a) that’s the way hockey-reference.com splits up their data, and (b) I haven’t finished the time machine yet.) Last season with the Islanders, Roli had a SV% of .912 and a GAA of 2.58 in October and November, almost identical to his numbers with the Lightning later in the year. In his career, his October-November splits are .912 SV% and 2.64 GAA. Um, oops. I admit it. I was wrong. He's not such a slow starter after all.

(The question of inconsistency lingers--his SV% throughout his career ranges from .884 (1999-00 in 14 games) to .933 (2003-04 in 48 games). His GAAs range from 3.05 in 2007-08 to 1.88 in 2003-04. But I haven't done the research to find out what the typical range is for goaltenders with similar length of service.)

And yet... The drop in Roloson's stats since last season are not unprecedented, and the next comparison I did was pretty informative.

It turns out that Roloson is doing about the same as Dan Ellis was at this time last season (better save percentage but more goals allowed) and quite a bit better than Mike Smith was. When combined, this year's goalies are performing better than the Ellis-Smith combo from this time last season.

That's not great news for Roloson, considering Ellis's eventual fate with the team. In addition, he's doing somewhat worse than his career averages, or his cumulative record over the last five years, or his career splits for October and November. That's bad in the sense that he is obviously struggling so far this season. It's good in the sense that there is considerable potential for improvement; he's capable of better. (Don't get overly excited about Garon's numbers right now. Two games ago, he was at 2.84 and .907.)

However, given the similarities between the goaltending records of last season's start and this season's start, it's difficult to say that the issue is solely the netminders. You've got 4 different guys doing more or less the same in similar circumstances. You've got 3 of those guys clearly underperforming in about the same way in those circumstances and 1 guy doing somewhat better than he usually does (Garon's currently just above his career averages: career totals of .904 and 2.81 and career October-November splits of .905 and 2.77).

To me this indicates that there's a problem beyond the goaltending, something that folks have been arguing since the season began. So in this, I was right. The problem with Roli may be less that he's getting old and can't cut it anymore, or that he personally is in a funk, than it is that the team has been in a funk. For instance, the team is giving up 30 shots per game right now compared to last season's 26 in Oct.-Nov., and about the same number of goals per game (3.12 GPG in Oct.-Nov. of 2010; 3.05 GPG this year).

A big part of the goaltending problem this season has been that the only way that a Dwayne Roloson-Mathieu Garon combination ever made any sense was if Guy Boucher's system worked. That is to say, Yzerman and Boucher had to have planned for the team to be more balanced towards the front end than the back end, or they wouldn't have signed goaltenders who could hold down the fort but who couldn't steal games. (Keep in mind that if Roli/Garon were performing at .908, which is their composite career SV%, the Lightning would have given up 62 goals and would still have a zero goal differential on the season, including the last two games.) The system hasn't been working, and the goalies are only the most visible illustration of that.

So, what to do about it? Well, first of all trading for a new goaltender under these circumstances is unlikely to fix what's ailing the Bolts, unless the goalie is a game stealer, in which case, it's going to cost—a lot. Putting a new guy into the same bad situation hasn't worked out yet and it likely won't unless the new guy is an elite or near-elite talent.

In the short term, give Garon game time. (Again, this was most likely the plan all along.) Garon's been the only one able to handle the load so far, and he should be rewarded for that. But fans should be prepared to see a lot of goalie switching in the near future. Second, fix the offense. Get the team to give the goalies the offensive support they need. Saturday's game against the Panthers illustrates how much help that would be. Third, fix the defense. This may be harder, given the way this roster is stacked with offensive power and weak on defense, but it has to be done. Reduce shots on goal to something manageable and the goals against will go down. Straightforward? Cliché? It still works. If the offense can consistenly score like it is capable of, and the defense can help the goalies out, then the team can look again at the goaltending with less desperation and more options.

Either way, whether problems get fixed or not, December is likely to be make-or-break time for this team and for these two goaltenders as well.

(All data is from hockey-reference.com and nhl.com. You can see my spreadsheet—and check my work—here. Note that GAA is calculated using TOI/60, not the number of games played. GP doesn't take into account either overtime or partial games. Note also that all currently accepted goalie stats are deeply flawed in that they don't sufficiently separate the individual's performance from his team's performance. That might be fodder for a third post, though.)