Windex Wonders: Goalie Edition

[RizzeeDizzee]

Following in the tradition of cleverly coming up with the name "Band-Aid Boys" to describe frequently injured players, last year Dobber and his readers coined the term "Windex Wonders" to refer to players that are notoriously streaky scorers. But can the Windex concept be successfully applied to goalies too? The answer is yes, and doing so can help GMs look beneath the numbers a bit more and in turn can help them in drafting, trading, and setting line-ups.

Most starting goalies these days finish the season with a GAA between 2.00 and 3.00, but how they end up there can be very different. Some goalies are remarkably consistent, stringing together a number of steady streaks of good games, whereas others have stretches of several games where they're either amazing or terrible. Looking not just at the GAA number, but at the game by game tallies of goals that a goalie gave up can tell you just how "Windexy"� (or "non-Windexy"�) the goalie is.

To actually calculate goalie "Windexy-ness"�, you have to account for stretches of solid consistency in net, as well as departures - both good and bad - from that consistency. Thus, you can look at (a) the total number of times where the goalie gave up 2 goals or fewer in 3 consecutive starts, and (b) the total number of times the goalie had 4 game stretches where he gave up 14 or more total goals, and (c) the total number of times where the goalie had 2 or more shutouts within a 7 game stretch. Yes, technically these are arbitrary numbers, but they are analogous to Windexy stats for offensive players and when consistently applied they can paint a useful picture.

To make the data even more relevant, we should only look at the 25 goalies who played in at least 45 games during the 2010-2011 season and had a GAA in the 2.00 to 3.00 range, since it is likely those are the guys who will be on rosters (or your draft/trade radar) for the upcoming season. Also, to make (a) more meaningful, we only want to count games where the goalie played 40+ minutes. Note that for calculation purposes, let's say a goalie had a 6 game stretch where he gave up 2 or fewer goals in each game, then that would count as "4" for purposes of (a), since the 6 games in a row is really 4 separate 3 game streaks. One shortcoming with this method of calculation is that is does not fully account for the awfulness of games where a goalie gets pulled early after letting in something like 3 goals on 7 shots.

So what's the definition of "Windexy-ness"�? Looking at all 25 goalies that met the 2.00 to 3.00 GAA and 45+ games played criteria for 2010-11, the average for (a) - (b) - (c) was about 3.1, so by definition any goalie with a number above that is not Windexy, and anyone with a number below that would be Windexy. Just how much above and how far below will determine the degree of (non-) Windexy-ness. But that alone is not the final analysis. If a goalie's "good"� Windexy-ness - denoted by (c) - is a lot greater than his "bad"� Windexy-ness - denoted by (b) - then the goalie can be as valuable, if not more valuable, than a goalie where (a) is much greater than (b) - (c).

Without further ado, here are the numbers and categories. For each goalie, his total games played are listed, then his (a-b-c) calculation:

Goalies that are very Non-Windexy

Cam Ward 74 gp (9-1-0) = 8
M.-A. Fleury 65 gp (12-2-0) =10
Roberto Luongo 60 gp (15-2-0) = 13
Tim Thomas 57 gp (17-1-8) = 8
Corey Crawford 57 gp (12-0-1) = 11
Sergei Bobrovsky 54 gp (13-1-0) = 12

Comments - It was surprising to see Crawford and Bobrovsky in double digits, although that was likely helped by their lack of shutouts and playing in front of good teams. Those two, plus Ward and Fleury, were about as consistent as you could get - not a lot of shutouts, but not a lot of bad stretches either. Thomas being here despite how many shutouts he recorded is a testament to just how amazing his year was.

Goalies that are somewhat non-windexy

Kari Lehtonen 69 gp (11-3-1) = 7
Pekka Rinne 64 gp (10-3-2) = 5
Jonathan Quick 61 gp (11-2-3) = 6
Martin Brodeur 56 gp (11-3-2) = 6
Jonas Hiller 49 gp (11-4-2) = 5
Michael Neuvirth 48 gp (6-0-1) = 5

Comments - It was surprising to see Lehtonen with a higher number than the other guys here. This bodes well for him if he can stay healthy, although the Dallas team in front of him has taken some offseason hits. Rinne and Quick probably would've been higher if they had better defenses in front of them like Luongo and Bobrovsky. It was interesting to see that Neuvirth was one of only 3 goalies to put up a zero for (b), with another being the guy who's slated to be the starter in front of him - Vokoun.

Goalies that are somewhat Windexy

Carey Price 72 gp (9-3-4) = 2
Mikka Kiprusoff 71 gp (10-7-3) = 0
Ilya Bryzgalov 68 gp (6-2-4) = 0
Henrick Lundqvist 68 gp (8-2-7) = -1
Ryan Miller 66 gp (7-4-2) = 1
Jimmy Howard 63 gp (3-4-0) = -2
Antii Niemi 60 gp (6-3-2) = 1
Tomas Vokoun 57 gp (2-0-3) = -1
Dwayne Roloson 54 gp (9-4-3) = 2
Niklas Backstrom 51 gp (10-11-1) = -2

Comments - Lundqvist is the guy who stands out here, since his 7 for (c) is only lower than Thomas's 8 and is 3 more than the next highest person. But it does say something that despite all his shutouts he did not put together as many strings of 3 straight good starts. As great a goalie as he is, he's not as consistent as a Luongo or Thomas. Price's numbers surprised me. His (a) of 9 is a lot lower than some goalies who many would argue are "worse" than him. It might be an indication of Price being a guy who is not as great - at least not yet - as some here seem to make him out to be. Vokoun was the only goalie to have all his numbers be 3 or less, which means he's never amazing and never terrible. It will be interesting to see how he does with a much better Washington team in front of him. I would've thought Howard's numbers were worse, but he mixed up his good and bad games pretty well.

Goalies that are very Windexy

Ondrej Pavelec 58 gp (6-12-1) = -7
Jaroslav Halak 57 gp (4-5-4) = -5
Craig Anderson 51 gp (1-8-0) = -7

Comments - It's not a shock to see Pavelec and Anderson here, as they had several stretches of bad games. More surprising is Halak, but his 4 for (c) in only 51 games is a positive sign, as is the consensus that the St Louis team will be healthier and better this year, and that Halak's numbers last season might have been due to a lingering hand injury that he tried to play through.

Conclusion

So what value is there in this information? Plenty! Digging deep like this can pay nice dividends, especially in leagues where you rotate your goalies in an out of the lineup as opposed to drafting them to play every game. For example, if you've retained a goalie that is very Windexy then you almost definitely want to compliment him with a very non-Windexy goalie, especially if you have weekly roster changes. Also, if you're thinking about making a goalie trade, you should see whether a non-Windexy goalie has been Windexy or a Windexy goalie has been uncharacteristically non-Windexy, since that could be useful data. And in leagues that don't reward shutouts, this stat can be very useful, since there you'd likely want the more consistent, very non-Windexy goalie as opposed to a similar, more Windexy counterpart. This data also helps you see goalies who cluster together their shutouts (or their bad games), as opposed to ones who spread them out.

Lastly, I should note that I put together all this data by hand, so there might be errors. Good luck in all your pools!

[ross10019]

Nice job bud!

[GMGates]

Pretty interesting!

Definitely putting this on twitter.

[Pengwin7]

Solid.

Writing Tip:
Put the lists at the top. (right after the first paragraph)
Put your written explanation afterwards.

People like to see the groups/names/scores immediately.
It gets them hooked.

Paragraphs often lose readers.
If somebody gets hooked by the names & group titles... they'll read the rest.

I think it is very good though!

[RizzeeDizzee]

Solid.

Writing Tip:
Put the lists at the top. (right after the first paragraph)
Put your written explanation afterwards.

People like to see the groups/names/scores immediately.
It gets them hooked.

Paragraphs often lose readers.
If somebody gets hooked by the names & group titles... they'll read the rest.

I think it is very good though!

Thanks - all tips (and readers) are definitely appreciated. I'll keep this in mind for future articles.