There’s a phenomenon in economics and business known as the sunk-cost fallacy. In short, organizations tend to over-value their own assets and hold on to them for far too long because of previous investments, economic and emotional. This prevents them from being able to make the best objective decision and ends up being costly in the long run.
How does this apply to hockey, and particularly this frustrating incarnation of the Dallas Stars? It’s actually pretty straight forward. Players, especially those who have been internally developed, are often held on to far too long or overvalued internally, because of the amount of resources already poured in and the emotional attachment to what they might be or were supposed to have been.
These games after the trade deadline were supposed to be a “tryout” of sorts for the defense, made up now of two players who had more than 100 NHL games coming into this season and an assortment of younger pieces. And the longer it goes, the more the defense as a whole is being shown to be incredibly problematic. If this tryout is designed to test what the Stars hope are the cornerstones of the future, they are failing pretty spectacularly.
Is this stretch exposing a sunk-cost fallacy in the internal defensive hierarchy? The trick with evaluating whether or not this is happening is to take a step back and look at the objective and current performance, with no name or history attached. So let’s play a game of that and see where it leads us.
All numbers are just from this regular season and collected from Hockey Reference for consistency. This does not include newcomer Greg Pateryn because of his very low sample size in Dallas.
Defenseman #1 0.07 PPG, Corsi 49.9%, Corsi Rel 0.2, 50% D-zone starts
Defenseman #2 0.22 PPG, Corsi 47.4%, Corsi Rel -2.5, 54% D-zone starts
Defenseman #3 0.20 PPG, Corsi 49.2%, Corsi Rel -2.9, 53% D-zone starts
Defenseman #4 0.61 PPG, Corsi 50.0, Corsi Rel 0.1, 52% D-zone starts
Defenseman #5 0.17 PPG, Corsi 50.3%, Corsi Rel 0.0, 56.1% D-zone starts
Defenseman #6 0.23 PPG, Corsi 52.7%, Corsi Rel 4.0, 52.0% D-zone starts
Okay, so there’s at least one obvious guy in here. Defenseman No. 4 is pretty clearly John Klingberg, who really struggled to start the year and has individually come back around (though still struggles with some possession numbers as his pair takes a lot of shots against).
For whatever his weaknesses are this season, the value is obvious. The points per game is ridiculously high for a defenseman, and his possession numbers in a down season and with somewhat tough starts are average or better for this team. He isn’t making the team as much better as he did in past years, but that’s as much a consequence of his pairing as well as his play.
The other player who is clearly succeeding this season is Defenseman 6, none other than offseason acquisition Dan Hamhuis. After a little bit of a rocky adjustment period, he has become the player Dallas pairs with the player they seem to trust the least (to pull them up to his strong possession game). No matter the situation he’s been put it, his pairing has been generally steady.
The other four defensemen are a combination of “fine but not impressive” and “worryingly underwhelming,” particularly from a possession point of view, which, of all the imperfect ways we have of judging defensemen, is most relevant to their primary job.
From that possession perspective, Defensemen 1 and 5 are clearly ahead of that pack. Both are average for the team as it stands this year (which isn’t necessarily high praise, but it’s better than the alternative). Defenseman 5 even gives a fairly average amount of offensive production to boot while taking the toughest average zone starts on the team.
These numbers probably project as defensemen who will top out in the 4/5/6 range, which isn’t a lost cause at all. It isn’t, perhaps, what one would hope from a high draft pick, but it is perfectly serviceable provided its viewed in the right context.
Defenseman 5 is Stephen Johns, the individual defenseman who as probably seen his stock fall the most throughout the season, probably unfairly. While he has had some problems in individual games with mistakes, his overall impression is very consistent with his performance from last season. The issue is, the team around him simply got worse, which makes him look worse in context.
Defenseman 1 is Patrik Nemeth, who has both decent numbers on a bad team this season and also is continuing a trend of improvement in possession relative to his teammates. Again, the ceiling for those numbers is probably a 5/6 player (especially when you consider he is in his third “full” NHL season, albeit with one very much hampered by injury), but that’s fine. He doesn’t produce a lot of offense, but you don’t need to in that role.
That leaves us with the players who make the team worse when they are on the ice, and the ones where the sunk-cost fallacy really comes into question.
Jamie Oleksiak is Defenseman 2, a former first-round pick now playing in his fifth season in the NHL. Two things are concerning from his development standpoint. The first is that, from a possession standpoint, he has shown no improvement over those five seasons. He is on track for his lowest career CF percentage (his highest was in his rookie year) and second-lowest Corsi Rel. He is relatively luckier this year than he has been in the past as well, at least according to his PDO. And his penalty minutes are trending upward as he is more and more exposed.
All the numbers, and the eye test of many, point to Oleksiak, rapidly aging out of “prospect” territory, being a below-average bottom-pairing defenseman with no upward trajectory. But this doesn’t match the impression of many, who point to his size, his hitting, and his brief flashes of goal scoring this season as showing definite improvement.
After all, he needs to be good, right? If he’s not, that first round pick would be wasted.
But that’s the sunk cost fallacy in a nutshell. The pick was spent six years ago - its value cannot be recouped at this point, and therefore its value should not play into the impression of Oleksiak as a whole. His current play, and therefore current value, is a negative contributor to a bad team.
So what of Defenseman 3, who is clearly Esa Lindell? He is much earlier in his career trajectory than Oleksiak at this point, but he has had a very, very rough start. That may not match the narrative being pitched for him, as he plays Top 2 minutes and on the flailing second power play unit, but it’s the story essentially every number tells.
His possession numbers are among the worst on the team. His offensive production is on par with Oleksiak and Johns despite playing with Klingberg and often the team’s best offensive forwards, very much a disappointment in context. His defensive minutes are no harder than most, which means he’s struggling more with better teammates in average minutes while someone like Johns is treading water and Hamhuis succeeding in much more difficult scenarios.
The difference in Lindell and Oleksiak is this is Lindell’s rookie season, and many players improve over time (though statistically more do not). He may become a legitimate, stastical 3 or 4 defenseman yet, but it would take some serious improvements both tactically and with some key skills like footspeed.
The trick here is not to fall into the Oleksiak morass of sunk-costs. The three years of development the Stars have put into him, the draft pick that was spent, none of that has any bearing of how he should be evaluated at this point. He is who he is, and the resources that have gotten him to this point cannot be recovered. Right now, his contribution is pretty much a failure, but it is reasonable to give him further shots in less demanding roles to see if it can lead to success. And that success or failure needs to be judged only on what he currently brings, not who he was developed to be.
That is the sunk-cost fallacy in a nutshell. It’s incredibly easy to fall in to - who wouldn’t want to see a return on previous investment? But to view prospect development and sports personnel management from that angle is to get lost in the forest of inevitable stalls.