BOSTON - SEPTEMBER 29: Milan Lucic #17 of the Boston Bruins and John Carlson #74 of the Washington Capitals fight for the puck on September 29 2010 at the TD Garden in Boston Massachusetts. The Capitals defeated the Bruins 4-1. (Photo by Elsa/Getty Images)
I saw this comment from Brian McNally, the Caps beat reporter for the Washington Examiner:
"Boudreau loves to knock wood when talking about PK. But luck be damned. Caps have killed 67 of last 76 penalties (88.2%). That's just good."
Forget that he's making a big deal out of a small sample size. Something I find really funny is that reporters can't seem to tell when NHL coaches speak the absolute truth. There's no question that Washington has been getting lucky at 4-on-5:
There's no reason to believe that Washington - which has seen its even-strength save percentage drop this season - has suddenly figured out how to substantially improve its save percentage on the PK. And what Boudreau might have in mind is his incredible luck when he's down two men, having given up just one goal so far this season in situations where the penalty-kill is largely helpless:
We know that the Caps haven't been a great PK team over the last three seasons. They parted ways with Joe Corvo and Shaone Morrisonn, who frequently played on the second PK pairing last season, and they've dealt with various injuries to the rest of the crew, including Tom Poti, who has been their #1 penalty-killer for three years. The big change was to add John Carlson and Karl Alzner to the top of the rotation and to trade for Scott Hannan. This improved the penalty kill's shot suppression abilities to nearly the top 10 in the league. However, Washington still suffers from legitimately below-average goaltending, which hurts them more than the defensive improvements help.
If we assume an .870 save percentage a man down and a .750 save percentage two-men down, then Washington would have allowed an additional eight goals. That would give them a PK efficiency of 80.7% on the season, which is below the league-average. That...is not good.