A Look at Belleville’s Penalty Kill

STATS

Back in late December I took a look at the BSens powerplay¬†and with less than thirty games left in the season it’s time to take a look at the team’s PK. The raw percentage for the team is terrible (77.2%, tied for 28th in the league, but giving up the second most goals with 49–only Laval is worse). There’s nothing unique about the BSens strategy on the PK (a fairly passive box), although as the season has gone on they’ve often chosen to use just two or three defensemen on the rotation when trusted players are injured.

In the absence of useful stats from the AHL what we can do is look at deployment (in this case a shift count)–who is being used and to what level of success? Given injuries and call-ups there’s been significant variety so there’s a lot to break down. First, let’s look at usage broken down by shifts-per-game (minimum 20 shifts and averaging 1 per game):
Forwards
Sexton 4.4
McCormick 3.97
Kelly 3.93
Blunden 3.77
O’Brien 3.71
Flanagan 3.0
White 2.84
Perron 2.02
Paul 1.93
Randell 1.86
Rodewald 1.18

Defense
Sieloff 4.7
Harpur 4.66
Englund 4.51
Burgdoerfer 4.23
Jaros 2.58
Erkamps 1.00

Small sample size is something to keep in mind, as Sexton has only played in 10 games, Flanagan 14, Harpur 15, etc. There’s a clear preference on defense (four primaries with a regular sprinkling of Jaros) and at forward (four–five with the departed Kelly–getting the heaviest usage).

So that’s frequency, but what about effectiveness? Here’s how they stack up based on goals against per shift:
Forwards
O’Brien, White, Paul, Sexton 0.06
Blunden, Flanagan 0.07
Randell 0.08
Rodewald 0.11
McCormick, Kelly, Perron 0.12

Defense
Erkamps 0.06
Sieloff 0.07
Harpur 0.10
Burgdoerfer 0.11
Englund 0.13
Jaros 0.16

The small sample size (in this case infrequency of use) means we can’t trust Erkamps’ numbers, but in general this is a fair if broad representation of relative success. Clearly this doesn’t quite match up with usage, so here they are together:
Forwards
Sexton 4.4/0.06
McCormick 3.97/0.12
Kelly 3.93/0.12
Blunden 3.77/0.07
O’Brien 3.71/0.06
Flanagan 3.0/0.07
White 2.84/0.06
Perron 2.02/0.12
Paul 1.93/0.06
Randell 1.86/0.08
Rodewald 1.18/0.11

Defense
Sieloff 4.7/0.07
Harpur 4.66/0.10
Englund 4.51/0.13
Burgdoerfer 4.23/0.11
Jaros 2.58/0.16
Erkamps 1.00/0.06

For the most part with the blueliners’ success follows usage, but at forward Kleinendorst’s favouritism impacts what he does, as Kelly and McCormick play more than their effectiveness justifies. Looking at this you’d say the top four PK forwards should be Sexton, O’Brien, Paul, and Flanagan (White is in the NHL, but if not, he bumps Flanagan), with the top-four D remaining as is with some Erkamps experimentation to see if his ability is sustainable. How does this hold up to actual pairings? We can break down who helps or hinders by seeing results depending on partners, so let’s take a look at the most frequent combinations (the dates in brackets aren’t necessarily absolute timeframes–they can also indicate the heaviest usage; the numbers are shifts vs goals against):

Forward Pairs
McCormick-O’Brien 0.06 100-6 (Nov-Jan)
Perron-White 0.05 39-2 (Nov-Jan)
Kelly-Blunden 0.08 34-3 (Dec-Jan)
Randell-Blunden 0.08 23-2 (Dec)
McCormick-Sexton 0.09 22-2 (Oct-Nov)
Paul-Rodewald 0.05 19-1 (Oct)
Kelly-Randell 0.13 15-2 (Dec)
Paul-O’Brien 0.00 14-0 (Oct-Nov)
Sexton-Blunden 0.08 12-1 (Feb)
McCormick-White 0.16 12-2 (Nov)
White-Flanagan 0.00 11-0 (Jan)
McCormick-Paul 0.09 11-1 (Oct)
Perron-Flanagan 0.00 10-0 (Jan)
O’Brien-Blunden 0.00 9-0 (Oct)
McCormick-Flanagan 0.11 9-1 (Jan-Feb)
Perron-O’Brien 0.12 8-1 (Oct-Nov)
Perron-Randell 0.12 8-1 (Oct/Dec)
Perron-Paul 0.14 7-1 (Oct/Jan)
McCormick-Blunden 0.14 7-1 (Oct/Jan)
Flanagan-Blunden 0.14 7-1 (Jan)
Sexton-O’Brien 0.00 6-0 (Oct)
O’Brien-Randell 0.00 6-0 (Oct-Nov)
McCormick-Randell 0.20 5-1 (Dec)
Paul-Sexton 0.40 5-2 (Nov)
Paul-White 0.00 4-0 (Dec)
McCormick-Rodewald 0.25 4-1 (Nov)
Perron-Kelly 0.50 4-2 (Dec)

While some of this bewildering variety is due to injury and experimentation, it’s far above what any coach should want in terms of stability and consistency. Kleinendorst has feels a compulsive need to play aging vets like Kelly beyond the bounds of reason. It’s also difficult to understand the decision to either get away from or stick with particular combinations. Conclusions: White, O’Brien, and (so far) Sexton make their partners better, while McCormick tends to be a drag on his partner and Perron is someone who reflects his partner (for good or ill). Paul, who has mostly played well in this role, was largely removed from the PK in mid-November for no apparent reason. In White’s absence and looking at the data, what we should be seeing is Sexton, O’Brien, Paul, and Flanagan getting the bulk of the time (we already know Paul-O’Brien works from the above).

Defense Partners
Englund-Burgdoerfer 0.14 87-13*
Sieloff-Burgdoerfer 0.08 69-6
Harpur-Burgdoerfer 0.06 43-3 (Dec+)
Englund-Sieloff 0.05 37-2
Englund-Jaros 0.17 35-6
Sieloff-Jaros 0.05 18-1 (Oct-Dec)
Englund-Erkamps 0.00 17-0 (Oct-Nov)
Sieloff-Harpur 0.13 15-2
Murray-Jaros 0.14 14-2 (Nov-Jan)
Englund-Harpur 0.16 12-2 (Nov-Jan)
Murray-Erkamps 0.00 10-0 (Nov-Dec)
Englund-Murray 0.20 10-2 (Nov-Dec)
Murray-Burgdoerfer 0.22 9-2 (Nov-Dec)
Lajoie-Jaros 0.00 6-0 (Jan)
*the duo had a horrific January (34-9), making them 43-4 otherwise (0.09)

A few things become apparent looking at it this way: Sieloff makes everyone better, with Harpur helping to a lesser degree; Burgdoerfer and Murray are drags on whoever they play with, while Englund’s impact seems more of a neutral factor–he can’t make up for his partner, but he doesn’t hurt either. Jaros is heavily effected by who he plays with (good with Lajoie and Sieloff, struggles elsewhere). Looking at what’s been tried I’d put Sieloff back with Englund and keep Burgdoerfer with Harpur until Jaros is healthy (and why not mix in some Erkamps so you can find out if the guy is actually a decent penalty killer or not?).

A few other player-specific comments:
-White, who has easily been the team’s best penalty killer, only started getting top rotation in January–Kleinendorst is so reluctant to trust younger players that even with a first-rounder like that he needed two months of watching before pulling the trigger
-Rodewald’s usage fell off at the same time as Paul’s, but his numbers aren’t nearly as good suggesting this was a smart coaching decision
-Randell received a huge spike in usage in December (60% of his deployment for the year) which has since dropped off to his more typical levels
-A third of Erkamps’ shifts come from a single game in October (4-2 loss to Syracuse) and over half from October in general–so buyer beware, but that’s not a reason to ram Murray into the PK (which was what Kleinendorst switched too)

This article is written by Peter Levi (@eyeonthesens)

Advertisements