106
Photo Credit: NHL.com

Get Out While You Still Can: Erik Gudbranson and the Sunk Cost Fallacy

Coming into this season, there was a whole lot of discussion about Erik Gudbranson. The pending unrestricted free agent was awarded a one-year $3.5 million extension in the summer after missing the majority of his first season with the Vancouver Canucks.

Gudbranson has been a polarizing player throughout his young career, with some lauding his physical, stay-at-home style, and others criticizing his horrendous possession numbers. The conflict between these two camps seems to be coming to a head as time ticks away on his one-year pact with the team, bringing on serious questions about what to do with the 6-foot-5 Ottawa native going forward.

On their TSN 1040 AM radio show on Friday, Mike Halford and Jason Brough broached this topic. Rightly, they pointed out that, as a pending free agent on a single year term, the Canucks will be able to re-sign him come January 1st. They then went on to state that the Canucks had but two choices: trade Gudbranson at the deadline, or re-sign him; they simply cannot lose him for nothing in free agency.

That’s not an uncommon line of thinking, and the exact quote made an appearance in Saturday’s edition of The Provies:

I’m here to tell you not only that the Canucks can lose him for nothing, but that’s it’s actually the second best option of the lot. What we’re looking at here is the Sunk Cost Fallacy. But before we get to that, let’s see how we got to where we are.

Acquisition

Gudbranson was Jim Benning’s big acquisition in the 2016 offseason, shipping former first-round pick Jared McCann to Florida in return, as well as swapping a few drafts picks which saw Vancouver lose significant potential draft value.

Losing McCann has been enough of a blow. The forward had five points in his first eight games with Florida so far this season and was beginning to push his way up the lineup before a lower-body injury forced him out of a handful of games. Those draft picks, though, and especially the second rounder, could really sting.

The term “second-round pick” hardly does it justice – it was, in fact, a 33rd overall selection, just barely outside the first round, where there is still a wealth of potentially impressive players. The Canucks drafted Kole Lind at 33 in 2017, and fans are rightly excited about his terrific start to the season. At the 2016 draft though, 33rd ended up being Rasmus Asplund, a Swedish centre who’s off to a fine start in the Swedish Hockey League with nine points in 17 games (only Elias Pettersson is having a better start to the season among under-20 players in that league). A couple of players selected shortly after that pick, such as Alex DeBrincat (Chicago) and Sam Girard (Nashville, since traded to Colorado), are already playing in the NHL and producing.

Fans around here are savvy, and popular opinion for a significant portion of the Vancouver fanbase was that the Panthers thoroughly fleeced Canucks general manager Jim Benning. Others were willing to delay judgement until seeing how Gudbranson played in Vancouver. Well, we’ve now seen Gudbranson play in Vancouver, and the results were have not been very appealing.

Results

In 179 minutes at 5-on-5 this season, Gudbranson has been one of the NHL’s worst defencemen. His shot shares are far and away the worst on the Canucks (aside from some very uncharacteristic, small sample Alexander Edler numbers), and the only other players that are even close are the ones that Gudbranson has spent significant time with this season. His shot attempt differential of minus-53 is more than double the next closest Canuck (Derek Dorsett, minus-24), and his CD60 of -11.5 means that the Canucks are allowing 11 and a half more shot attempts against than for, for every hour that Gudbranson is wandering the ice.

His five-on-five numbers are ghastly enough: only 23 defencemen with at least 100 5v5 minutes have a worse Corsi-for percentage right now. But that’s just the half of it. The Canucks have put up decent possession numbers this season, and are even top ten in the league in adjusted CF% and xGF% as a team. That really puts a dent in Gudbranson’s team and teammate-relative numbers, putting him near the absolute bottom of the league.

Player CF% FF% SF% GF% xGF%
5v5 44.10 43.99 43.58 48.70 42.44
5v5 Rel -11.33 -10.67 -10.32 -12.64 -13.95
5v5 RealTm -11.12 -10.79 -11.71 -12.64 -14.70

Source: Corsica.hockey

His -11.3% relative CF%, for instance, is the second worst among all defencemen with at least 100 minutes at five-on-five. His relative Fenwick ratio is seventh worst, and his relative Expected Goals ratio is eighth worst. In each case, the culprit is his complete inability to suppress shots. The Canucks allow 15.6 more shot attempts every hour that he is on the ice, relative to when he is on the bench, as well as nearly half an Expected Goal. Again, that’s the relative difference every hour.

Worse yet is that he’s managed all of this in relatively sheltered minutes, lining up for roughly as many faceoffs in the offensive zone as the defensive zone, and ranking fifth on the Canucks regarding the average minutes played by opposing forwards (an approximation of quality of competition). He’s been a significant drag on whichever poor defenceman draws the short straw and has to babysit him on the ice, as evidenced by the following table, showing all d-men that he’s spent at least ten minutes with at five-on-five.

Teammate TOI With TOI Away CF% With
Gudbranson
CF% Without Gudbranson
Michael Del Zotto 112.25 57.32 40.1 60.5
Ben Hutton 37.42 132.15 45.7 53.2
Derrick Pouliot 13.27 114.12 48.7 56.9

Source: Natural Stat Trick

The analytics look terrible on Gudbranson, but he hasn’t looked any better to the naked eye. While some hoped that mid-season surgery to an injured wrist would improve some of his underlying numbers from last year, that hasn’t been close to the case. It makes sense: wrist surgery was never going to fix questionable decision making, poor reads, awkward pivots and transitions, and a noticeable lack of foot speed.

Any direction that you look at it, Gulbranson’s results have been brutal, which is pretty much in tune with what he’s accomplished in his career to date. Now, let’s double back and explore the idea of sunk costs, and how Gudbranson fits the bill.

Sunk Costs

The mere fact that the Canucks cannot recoup what they paid for Erik Gudbranson does not mean that they are forced to keep him.

Professional sports, particularly one as divided from educational avenues as hockey, tend to struggle with critical thinking, and in turn, lend themselves to a host of logical fallacies. Appeal to authority is a particularly common example. In the case of Gudbranson, the Sunk Cost Fallacy is most relevant.

A sunk cost is something that has already been paid for, and the payment for which cannot be recuperated. Having already invested something of value into this object, the emotional mind will go to great lengths to avoid feeling like the investment was wasted — even if that means sinking more capital into that investment.

The fallacious part of it is the lack of awareness that while you think you are simply trying to break even, you are really only making the problem worse.

Gudbranson is a sunk cost. Convincing yourself that you need to hold on to him just because you paid a hefty price for him does not help the situations — it just digs you deeper and deeper into a hole.

What this means

Obviously, I don’t want the Canucks just to lose an asset for nothing. We often talk about asset management, and with good reason. But asset management is still the point here; sometimes walking away is better asset management than doubling down.

For many, like Botchford, Halford, and Brough, the order of preference goes like this:

Trade > re-sign > walk away

I see the logic behind this preference, simply because walking away from an asset that you paid a hefty price to acquire is seen as a deadly sin.

My order of preference is similar, with an important distinction:

Trade > walk away > re-sign

As with most people, my first choice would be to get something for Gudbranson. If there are teams out there that still value what he brings to the table, and we know that there are, by all means, the Canucks should exploit that. The difference for me is that walking away isn’t my worst case scenario; re-signing Gudbranson is.

Part of this requires some context. I’m not necessarily saying that Gudbranson isn’t an NHL player; I just don’t believe he’s a top-four defenceman, and I think he’s drifting dangerously close to replacement level. He’s not worth $3.5 million a season, and he certainly won’t be worth the big, long contract he’ll inevitably get as a pending unrestricted free agent next summer. There’s an amount that I’d be willing to pay Gudbranson to play on the bottom pair, with no guarantee of a spot in the lineup each night. But it would be close to the league minimum – definitely not an amount that he or his camp would accept, which makes going down that avenue pointless.

Trade Expectations

I’ll make one more distinction from popular opinion, and that is to lower your expectations of a return. They aren’t getting McCann back. They aren’t recouping the value they lost in the pick swaps. The deal they made last summer is, and always will be, a loss. Canucks management can’t go into future trade negotiations with the mentality that they need to recover that lost value.

When the next option in line if a trade doesn’t occur is letting him go for nothing, you need to be willing to take whatever you can get for the asset. Even if you go into negotiations with higher expectations, they should be lowered accordingly as the search for a deal proceeds.

If it comes down to it, even a late draft pick or a warm body on a low salary should be acceptable deals. If they can get a serviceable player like Alex Petrovic in return, as our editor J.D. Burke outlined this morning, all the power to them then that’s an absolute win for Vancouver in my books.

And, again, if a deal cannot be reached, then it’s time to cut bait and walk away. Signing this player for the salary that is going to command, and for any length of term, will absolutely come back to bite the Canucks, and should be avoided at all costs.

The $14 million that will be freed up as the Sedins come off the books after this season will look like a lot to start with, but healthy raises for the likes of Sven Baertschi, Brock Boeser, and Troy Stecher will be in the offing in the next two off-seasons, and any term longer than that could start to impede the first non-ELC contracts of players like Elias Pettersson, Jonathan Dahlen, Adam Gaudette, and Olli Juolevi, should they command any serious sum of money (and the hope is obviously that they will).

Having three of four million dollars tied up in a replacement level defender is never going to be a good look. Merely not having him on the roster at that salary next season is the very definition of addition by subtraction, both in terms of the salary cap structure, and the team’s ability to control the flow of play on the ice-ice.

So, as we approach the January re-sign window, and then the trade deadline, keep these things in mind. Getting something in return is still a priority, but if that isn’t feasible, they do not have to re-sign Erik Gudbranson, and they shouldn’t do it either.

If Gudbranson is on the Canucks’ roster next September, it means in all likelihood that a mistake has been made, and they could suffer from it for years to come. The only safe solution is to get out while they still can, be that on or before February 26th at the trade deadline, or on June 30th when the current contract expires.

      • Hi Jeremy, I normally really like your articles and I don’t doubt your stats, but here is where I have an issue. I’m an ex goalie and have noticed that in recent years defence seem to at most tie up the opponents stick in front of the net. Often they allow the players to stand in front of the net uncontested so as not to add another player screening the goalie. I have on multiple occasions this year watched Gudbranson remove the player from the front of the net. In Detroit, he moved a player from the front of the net to below the goal line as the defence man was winding up for a shot. As a goalie I have no problems facing lots of shots when I have a clear line of sight. For all his faults, Gudbranson removes players from danger areas and he’s the only Canuck I have seen do that in years. How does that get measured in the stats? I know it has value.

        My second question is regarding the stats. My understanding is that the purpose of the stats is to predict future performance and success. If you have more possession you should score more, etc. You just told us what Malkin’s CF%, shot share and scoring share chances were against Gudbranson this year. What were the actual results. How many more goals did Malkin and his line score against Gudbranson than the Canucks score during that time against Pittsburg?

        We can hope fancy stats can predict how things will go, but in the end the only thing that matters is the results. I don’t think Gudbranson is a top d man, but I agree with lots of the other posters that I didn’t see that poor of a performance by the eye test.

        • You measure it by digging deeper into the Corsi stats and looking at the quality of shots against. Heat maps are avaialable that show this kind of data but they were not included in the analysis. I’m guessing that not a lot of shots would show up close to the crease when Gudbranson was on the ice.

      • Malkin didn’t get any points so there is that. I would be very interested to see what the zone stats were for those nine minutes that Gudbranson faced him at 5v5.

  • Very logical and reasonable article. The analytics for Gudbranson only apply to regular season games though. In 2011, when the referees suddenly decided not to call anything – not even the attempted murder of Mason Raymond, a player like Gudbranson on the Canucks D would have been invaluable.
    Benning came from the team that won that series. The league seems to have changed its playoff officiating since 2011, but who can know how they will call it should the Canucks ever return to the playoffs. It might be an idea to have a guy like Gudbranson around just in case.

    • I hand-track certain aspects of games, which is extremely difficult if you don’t watch. If you think Gudbranson looks good, I have no idea how you’re seeing that. The eye test is as bad as the stats, as mentioned in the article.

      • Respectfully suggest that there may be some confirmation bias on your part when watching him play. Meaning that you (and the rest of CA) are virulently anti-Gudbranson and are going to see that when you watch him.

      • Agree to disagree Jeremy.
        As dozens of posters here are pointing out with more than frequent regularity – there are constant and sometimes staggering differences in what is written about the games and what the rest of us see.
        I am not biased but I am a Canucks fan.
        I believe the reverse is true for almost everyone that contributes articles here on CA.

        • “Looks good” – this means he makes a couple big hits a game. When he connects it does look really good! He’s massive, strong etc. A few plays like this have a way of standing out for anybody watching the game is a ‘normal’ fashion. Problem is what is he doing for the other 95% of the game which I assume you’re watching just as intently? Enlighten us, Locust.

    • agree to disagree – you sound like Ron Burgundy insisting “a whale’s vagina” is the correct translation of San Diego. This article is correct, yet I laughed when I saw the title because I knew how much bloo-bloo the comments section was going to be. I really do not understand how such a generally savvy website can have such bad commenting. The “eye test” is again, not a quantifiable thing. If you can’t quantify something, you can’t track it. It reminds me of the scouts in Moneyball complaining that a player has an ugly girlfriend.

  • Looking forward to Bud Poile popping up and yelling about how Gudbranson’s poor numbers are all Ben Hutton’s fault.

    Of all the bad moves Benning has made since taking over as Canucks GM, the Gudbranson trade is probably the worst. Giving away a centre prospect with top-six upside and a top-40 draft pick for a guy who looked to be, *at best*, a second-pairing defenceman was a bad move on its face. Given that Gudbranson has been a replacement-level defenceman in his time with Vancouver, it was just a colossal misstep.

  • Analytics have obviously taken over as the primary evaluation method for players in today’s game, however, hockeys played by human beings, it’s not just a simulated algorithm. You cannot quantify every facet of the game! Stop trying to! There are more factors involved in the game of hockey than what you can record on a stats sheet. Any author who’s played the game should be able to realize that.

    • I played hockey my entire life. You don’t need to quantify everything. He looks bad out there. Poor reads, questionable decisions, glass-and-out plays, routine turnovers, misplays on odd man rush defences, slow pivots, lack of foot speed. He’s big and strong. That’s about it. Other than that, his skills are about that of an replacement level NHL player.

      • Playing D is a world away from skating/floating up ice.
        Here’s the Guddy stats I compiled five days ago when replying to an earlier CA hit piece:
        #42 of all NHL d men-SH Hits
        #81 of all NHL d-men SH Blocked Shots
        #102 of all NHL d-men SH TOI
        #37 of all NHL d-men in Hits-leads Canucks d-men
        #28 of all NHL d-men Hits/GP-leads Canucks d-men
        #79 of all NHL d-men BS per GP
        #97 of all NHL d-men Shots per GP

        “He’s big and strong. That’s about it.” Jeremy Davis

        Jeremy,come playoffs when the whistles are put away and hockey becomes a battlefield games are won with men like Guddy.
        That’s the bottom line.
        Dorsett,Sutter and Gudbranson are warriors you win with.
        These men will never be “replacement level” players.

        • I didn’t know that the Canucks were getting outscored. You need to watch all the other D, when you single out one D for your ability assasination it is very easy to pick things you don’t like, but hockey is a team game. I don’t think your stats work very good for a hard hitting D that keeps the other team wary. To see how pure stats are working for player selection just look at where the coyotes and panthers are in the standings. That is the true test of where the rubber meets the road.

          • So you’re saying that Guddy’s hard-hitting play tires out the stars thereby allowing the other D to put up better fancy numbers? So, he’s like a martyr?

        • It literally isn’t. Edmonton’s outshot their opponent almost every single game by a large margin, almost twenty shots and is still thirtieth overall. The best corsi teams have missed the playoffs by fifteen to twenty points. Corsi is basically do well enough, outshoot your opponents significantly, and your guaranteed eighty points. You took a loose correlation with winning and shots taken and completely misapplied it. Every post indicates your complete lack of understanding for why there was the relationship to begin with.

          • The reason why there were additional shots was because scoring chances typically create shots, not vice versa.