Is it worth trading up in the NHL draft?
we'll likely never know ...
Zeev Buium is the talk of this year’s defense prospect class, for good reason. He’s an effortless skater who creates offense in droves. Even if his defensive acumen isn’t world renowned, it’s good enough. Buium’s scouting reports read like those of Evan Bouchard, Quinn Hughes, and Adam Fox. He’s the sort of defenseman worthy of a top-10 selection. Which represents a quandary for the Sharks. Unless the Penguins win the lottery and hang onto this year’s pick, Mike Grier and Co. will also own the 13th selection in the 2024 NHL Entry Draft. And after Buium’s performance in the NCAA Frozen Four, he’s unlikely to be available at 13, where he was located at the season’s midway point. If the Sharks would like to add the flashy blueliner to their arsenal, they’ll have to consider trading up.
But is that a smart move?
What do the numbers say about trading up?
To move from 13th to 7th, say, the Sharks would also need to fork over the 33rd pick. If Grier wants to move up to 10th, he’ll need to add the 42nd pick. If you can guarantee that Buium turns into the next offensive phenom from the blueline, an early second-round pick is pittance to pay. Therein lies the rub. Despite his promising profile, no one really knows how good Buium can or will be in the NHL. This uncertainty is the first reason that trading up might not be such a good idea.
The second reason is that GMs tend to overvalue draft picks relative to how often they become star players.
This chart from the Sound of Hockey blog post linked below shows how a draft-pick-value curve based on actual GM trades compares to one based on player value once in the NHL:
At the very top of the draft and especially in the middle of the first round, GMs have given up more value than they’ve acquired in the trade up. Another approach to the same exercise using a goals-above-replacement (GAR) model demonstrates the same phenomenon.
If we use the pick values from the GAR approach, the two hypothetical Sharks trade-ups really look like this:
Pick 13 + Pick 33 + the value of pick ~61 = Pick 7
Pick 13 + Pick 42 + the value of pick ~78 = Pick 10
By trading up, Grier will surrender an undervalued second-round pick. He’ll also give the receiving team the equivalent of a late 2nd- or early 3rd-round pick in terms of the likelihood of finding an NHL star.
How have GMs fared in real terms during trade ups?
The value of a 2nd-round pick is abstract and doesn’t exist in reality. It just means that those 2nd-round picks turn into stars at a higher rate than GMs presume. By looking at historical draft-pick trades we can see how well GMs perform when they move up for a player.
Inspired by Bill Barnwell’s recent work on the NFL draft, I used NHL Trade Tracker and CapFriendly to log all the draft-day trades between 2003 and 2014 involving only current-year selections. That means I did not include trades at the 2013 NHL draft involving picks in the 2014 draft. Likewise, I excluded trades where an NHL player or prospect was involved.
I wrote down every player selected at the picks in question then looked up their prime-age (age 23-27YO season) goals above replacement, per Evolving Hockey’s model. For goalies, I used their age-25-29 seasons. I looked at all 85 transactions and marked how often either side ended up with the better NHLer in the trade. In many cases, none of the players ended up as NHLers. In other cases, one or two players played some games over the course of 2-3 seasons. In those two scenarios, I considered the trade a push.
Of course, it’s never really a push. Teams moving up give their trade partners a higher likelihood of drafting a star player, regardless of how said players actually pan out.
Here’s an example from the 2003 draft:
This trade was obviously a win for the trade-up team, Anaheim. There was one weird case, where Nic Deslauriers ended up the best player of the bunch with a cumulative -9.7 goals above replacement during his prime. Now, technically, -9.7 goals above replacement is worse than getting 0 goals above replacement from someone who never makes the NHL. But Deslauriers has played more than 600 games, so clearly NHL teams think he has some value. Even according to the model’s distribution of talent, Deslauriers never had a single season during this span where he was an obvious replacement candidate. As a result, for this exercise I counted Deslauriers as the best player of the group.
Here are the major results from my quick-and-dirty manual study.
Teams trading up for players got the better player about 33% of the time.
Teams trading back ended up with the better player 28% of the time.
If you restrict the hit rate to just trades involving first-round picks, trade-uppers got their guy 43% of the time.
Teams that traded down in the first round ended up with the better player 31% of the time.
When teams traded up for players in the first round, those players surpassed expectation at a similar rate, about 40% of the time. I define expectation as the expected prime-season goals above replacement at a given draft slot.
When GMs traded up for a player in the first round, the next player selected had a better prime-age career 43% of the time.
I used the 2003-2014 drafts because those are all the drafts for which GAR data is available for everyone’s age 23-27 seasons. We can look at a few more recent first-round trade ups to see not much has changed.
At the 2018 draft, for example, the Blues traded picks 29 (Rasmus Sandin) and pick 76 (Semyon Der-Arguchintsev) for 25, where they selected Dominik Bokk. Sandin’s performance to date hasn’t been too impressive, but he’s logged 68 NHL games during his prime-age seasons. Meanwhile, Dominik Bokk is playing in Germany. Two-thirds of the players (Nils Lundkvist and Jacob Bernard-Docker) the Blues could’ve selected between picks 25 and 29 have solidified themselves as bottom-of-the-lineup blueliners already. Of course, the Sens acquired the Bernard-Docker pick when they traded back with the Rangers, who moved up to select K’Andre Miller.
Those are only two more recent examples, but they illustrate the same push/pull of first-round trade ups. Win some, lose some. Yet, in most cases handing over the additional draft-pick value probably isn’t necessary given how uncertain it is the team trading up will end up with the best player when all is said and done.
What to do about Buium?
It’s clear moving up in the first round can cut both ways. If historical rates are at all predictive, however, Grier and Co. have about a 43% chance of landing the best player in a trade-up swap and a 40% chance he outperforms expectations given his draft slot. I will say these are higher hit rates for trade-ups than I guessed heading into this exercise. After counting up all the outcomes, I was a little more hesitant to proclaim that trading back in the middle of the first round is always the way to go.
Then I thought about it another way. Could you imagine walking up to your favorite team’s GM on draft day and asking him to flip a coin to decide whether or not to trade up for a player in the first round? And then telling him that, no matter which prospect the team ends up with, they’ll also have to give up the value of an 2nd- or 3rd-round pick in the process?
When couched like that, I’m not sure trading up is such a wise proposition. If a team does decide to trade back, they’ve historically had about a 57% chance of at least breaking even with their trade partners, in large part because of that extra trade value. In fact, half of the teams that “won” the trade-down scenario did so because of a player they selected with a lower pick they earned in the trade.
There you have it. Using either the abstract idea of pick value or real-world trades, trading down is a better option than moving up, simply because it leaves you with more outs. I hope Grier and his team understand this and make a wise decision come June should this option present itself.