Breaking Down: The Curious Case of Nemeth & Compher
Plusminus is a widely derided statistic in the modern view of hockey. On the surface it seems logical, player X was on the ice at even strength (and a couple special situations) for this many goals more/less than a net zero. That’s not exactly how it works and the flaws begin showing up quickly the more you look at it.
Even strength is a catchall term for any time there is not a man advantage for either team. It includes 5v5, 4v4 and 3v3, which make sense, but it also includes empty net situations because technically each team has the same number of players on the ice. These days we’ve shifted to looking at 5 skaters vs 5 skaters when analyzing teams statistically and lumping the other situations into more or less the same territory as special teams since they have specialized personnel and strategies along with small sample sizes.
An even bigger flaw than using empty net or less than 5v5 even strength situations for plusminus is that shorthanded goals against on the power play and for on the penalty kill are included for reasons both arcane and forgotten. As we’ll see, this is not only silly but it further skews an already flawed statistic.
Patrik Nemeth ended the regular season at +27, good for 11th in the NHL, 4th among defensemen and #1 on the Avs. Gabe Landeskog was 2nd at +16 and the next closest regular defenseman was Big Z at +4. Seems like a bit of an outlier.
JT Compher finished up at -29, tied for 9th worst in the NHL and bottom of the Avs by a factor of 2. Out of the top 50 worst +/- in the league only 2 players come from playoff teams. To refresh your memory, the Avs as a team had a +20 goal differential overall. Needless to say being -29 on a 95 point team is impressive.
Even though plusminus is a garbage stat, outliers inspire curiosity so I’m going to pick apart these two fellows’ reasons for ending where they did and see if there are any conclusions to draw.
Nemeth (+27)
5v5: +47/-34, net +13
4v4: +3/-1, net +2
3v3: none
ENG: +14/-7, net +7
SHG: +5/-0, net +5
Compher (-29)
5v5: +16/-30, net -14
4v4: +1/-5, net -4
3v3: +1/-2, net -1
ENG: +4/-9, net -5
SHG: +2/-7, net -5
Tackling these one by one, the most obvious discrepancy is at 5v5. For both players, about half of their surplus/deficit comes from straight up normal hockey with 5 skaters a side. Both guys have GA/60 rates that are around mid-pack on the Avs, Nemeth slightly better at 1.91 to JTC’s 2.25. Where Nemeth kills it is in GF/60. He is on ice for 2.64 goals per 60 minutes which is yards ahead of Compher’s 1.19. Nemo’s rate is about the same as Tyson Barrie’s, which makes sense since they play together frequently, and is only surpassed by the top line guys (and Yak if you want to count that). Compher’s goals for rate is the lowest on the team among regulars, we’re talking behind Bourque at 1.4/60 and Wilson/Jost at 1.77/60 so he’s lagging by a mile. Also interesting to note that both Nemeth (+15/-15) and Compher (+7/-7) are even 5v5 when the score is not within one goal, so this isn’t a bunch of score effect tomfoolery.
At 4v4, the Avs were a net +6/-9 and at 3v3 were a net +3/-8 so not too good overall. Nemo is an outlier at 4v4 and there’s not enough data to suggest why. Compher was bad at 4v4, suggesting he might be a possible cause, and better than team average 3v3.
The Avs were a net even in empty net situations but breaking it down they were fantastic at scoring ENG’s (19) and craptastic at scoring 6v5 (2). There’s no breakdown on personal 6v5/5v6 scoring but Nemeth’s usage was primarily to shut things down at the end of games while Compher spent more time in the futile pursuit of trying to tie the game late. The Avs were +7 when the opponent pulled their goalie and strangely enough so was Nemeth. They were -7 with their own goalie pulled and Comphs was a -5 so that tracks. On JTC’s side this might go back to having the lowest GF/60 rate on the team and that affecting their overall performance.
The last and silliest component of plusminus is shorthanded goals. The Avs tied with the Islanders for letting up the most shorties with 11 and Comphs was on the ice for 7 of those. Since Nemeth doesn’t play PP he was on for none. They scored a respectable 6 SHGs themselves, Compher was on the ice for 2 (of his own) and Nemeth for 5 (with 2 assists). Net special teams numbers had Comphs at -5 and Nemeth once again yarding him at +5.
Patrick Nemeth doesn’t drive play that we can see easily. He generates about a SOG per game and a point every 4½. His CF% isn’t great although his xGF% is a lot better. From what I see his swollen +/- comes down to being on the ice with the right teammates and at the end of games where the Avs scored empty netters every 4 nights. Couple that with not being in situations where the team struggled like the power play and 6v5 and voila! +27. It’s a little bit of individual skill and a lot of role & usage, nothing wrong with that at all.
Compher’s situation is a bit more murky. He was definitely an anchor for a large part of the season but it’s odd because he’s a decent shot generator that had a gruesome shooting percentage outside of the power play. In contrast to Nemeth, he had a satisfactory CF% and bad xGF%. He’s a decent defensive forward that, from these numbers, you do not want on the ice in offensive situations. The question is whether this will always be the case and it’s too early to tell.
Looking at both these guys side by side the minus components of their +/- were pretty similar. Nemeth just had a huge amount of plus that Compher didn’t which translates into their on-ice goals for rates. This is something the staff will look at over the summer, they already made a few adjustments towards the end of the season that helped a little. If Nemeth’s +27 is mainly role & usage then it’s stands to reason that a large part of JTC’s -29 lands there too.
Of course now I really want to know why Compher’s GF/60 is abysmally low so look for an examination of that next time.
