What should the Sharks' defense pairings look like?

Applying a purely objective model to solve for the theoretically best pairings.

In a break from the mold, I thought I'd follow Shutdown Line and create a mystery pairing post, with a reveal at the bottom in the comments section. Do your best not to peek! It will ruin all the fun.

First, the data thanks to stats.hockeyanalysis.com. The pairings represent 2 defensemen. 1 or more defensemen may appear in more than 1 pairing. I filtered the possible pairings to those over 48 minutes of 5v5 together, the average amount of time spent 5v5 per game by all teams this year. Note: Burns is not included in this data.

5v5 Defensive Pairings 2012-2013

Pair TOI 5v5 GF 5v5 GA 5v5 Corsi Diff 5v5 Corsi/60 5v5 Corsi%
A 66.2 2 2 25 22.65 0.611
B 249.2 6 9 88 21.19 0.593
C 113.9 6 3 27 14.22 0.561
D 51.1 1 1 1 1.17 0.505
E 204.2 11 4 -9 -2.64 0.488
F 83.7 3 6 -6 -4.30 0.477
G 156.6 4 3 -30 -11.50 0.454
H 129.0 4 7 -29 -13.48 0.442
I 54.7 0 3 -33 -36.20 0.351

Note that the data above is not adjusted for score, zone start, or forwards. Clearly, by the way I've sorted the table you can tell I'm biased towards Corsi, and not plus-minus. Especially given the limited amount of time each of these pairings have had together.

  • Pairing A: An unlikely pairing that hasn't been used frequently. I'm hesitant to label them as the Sharks best pairing, as they have played so little together. And in fact, one member has sat a fair number of games in the press box. At this point, they've played a substantial amount of time to believe their numbers to an extent, but I would expect their numbers to regress heavily if they were given top pairing.
  • Paring B: A new pairing this year, with a high event rate. This pairing has been impressive to many fans since its inception, with some calling 1 of the defensemen a key to the Sharks success.
  • Paring C: This has been the Sharks best pairing for a very long time. A great stay at home defensemen combined with one of the best puck moving defensemen in the game today. As you can see they aren't paired together often, mostly because of the sharp decline in depth after these two.
  • This pairing has had limited time together, and I can't honestly remember a game in which they started out partners. Nonetheless, the formula here clearly works. Combine a good puck mover, with a more traditional stay at home, and you get some progress in the shot department.
  • Paring D: Not often seen together either, this pairing at least in theory should work. They don't share the ice because they've both shown to perform better with other partners.
  • Pairing E: Neutral's favorite pairing in the beginning of the year, they've combined for a team best +7 when together. Neutral finally got what he wanted. But the saying, "be careful what you wish for," as this parings puck moving ineptitude has been surpassed by paring H.
  • Pairing F: An unlikely pairing, nothing else worth noting.
  • Pairing G: A bottom pairing that has struggled at 5v5. One member has been routinely criticized by FTF for his continued presence on the roster.
  • Paring H: Also a bottom pair, featuring the same defensemen mentioned in G. At -3, and -13 Corsi/60, this is probably the worst pairing the Sharks have routinely trotted out, and they're definitely not facing the toughest of opponents, nor starting a significant amount of shits in their own end. Everyone knows its a lack of zone exit ability.
  • Paring I: A bottom pairing that has barely seen time together, and probably shouldn't unless an injury occurs.

Now to the operations management part.

Given the constraints of limiting each D men to 18min or less of 5v5 ice time (the current amount skated by the top pair), and assume an average of 48 minutes of 5v5 per game for the sharks, we can ask a computer to maximize the time each defensemen spends with each other to achieve the highest theoretical Corsi per game possible. Below is the output of that program.

Solver Output

Pair TOI/Game Corsi/TOI Corsi%
A 18 6.80 0.61
B 18 6.36 0.59
G 12 -2.30 0.45
Total 48 10.85 0.56

Yup, a +11 Corsi per game is what we can expect if we optimize the pairings generically as above. It turns out that as a whole the team has averaged a Corsi/60 of about +2 over the course of the season at 5v5, and a Corsi% of 51%.

To be pretentious there may be some room here, but I would count on the Sharks coaching collective to spot favorable matchups over something as mechanical as the analysis above. And of course, they have to deal with injuries, zone starts, PK/PP, and a host of other factors we aren't accounting for.

Also, it's important to note that we're dealing with a limited amount of ice time, and coupled with the injuries, it's tough to say these would look the same at the end of the year. The coaching staff also has the benefit of evaluating players through training camp and team practices.

***Check the comments section below for the reveal.