clock menu more-arrow no yes

Filed under:

The Other Side of the Trade Down

New, comments

Why do teams trade up?

Columbus Blue Jackets v Philadelphia Flyers Photo by Patrick Smith/Getty Images

In the last few years, the notion of trading down at the draft has become more and more popular among the NHL analytics community. The idea is simple - NHL draft pick value decays exponentially. When analysts study the draft to determine the individual value of a pick, it’s commonly found that there’s almost no difference in value between picks after the first 20 or so. Essentially, the top-end of the draft is disproportionately valuable.

NHL draft pick value, from SportsNet (Steve Burtch’s model)

The idea in trading down is that you’re able to obtain a higher quantity of picks, and on average, the downgrade in pick quality isn’t that large. If you play it right, the math works out in your favour.

So let’s take for granted that trading down, as a general idea, is a smart strategy in the NHL draft - we haven’t shown this rigorously, but it’s quite plausible, when looking at the research done about draft pick value.

What I’m curious about is the psychology behind trading up. Why do teams do it? Are they just stupid? That’s certainly a possibility. It wouldn’t be the first time NHL GMs systematically improperly valued an asset. To me, the more compelling reason is that when a NHL team trades up, they’re generally trading up for a player, one that they think would be undervalued at the position they’re trading up to.

The argument for trading down is much like the argument for passive investment in the stock market. No one can consistently pick winners, so don’t waste your money. Similarly, in the NHL draft, trading down is about recognizing that most of your draft picks will fail, no matter how good your scouting staff is, and that by having a large quantity of picks, you maximize your chances of producing a NHL player. Trading up flies in the face of that logic.

You trade up because Travis Konecny is just sitting there. You trade up because oh shit, the Bruins could have gotten Barzal (and Connor, and Kylington).

You’re generally not trading for the pick anymore, you’re trading for the player. In the case of Barzal and Konecny, early returns indicate the team trading up made the right move - even though on average, those trades may have negative expected value (in fact, they do, if we use the Schuckers draft pick value linked earlier).

Aside: This is a problem with draft value charts in general. They’re averages, and while averages can be applied to individual cases, said cases will almost certainly deviate from the average in one direction or the other (to be clear, no one who has come up with a draft value chart has advocated using it, and it alone for decision making).

So how do you adjust for specific situations? The obvious answer is to adjust chart based on where you think the player you trade up for ranks.

When the Flyers or Islanders were trading up, they likely had some draft pick value chart to reference their decision. And that chart probably said the trade they would make, on a pick for pick basis, was a bad one. However, Konecny was probably higher on the Flyers’ big board than 24th, and Barzal was probably higher on the Islanders’ big board than 16th. In their minds, they weren’t trading for the value represented by the 24th or 16th pick - they were probably trading for a pick that had top 10 value to them. Conventional wisdom indicated they were right - both Konecny and Barzal were rated by scouting services and draft models to be lottery picks.

Of course, the same is true of the unsuccessful trade ups. The Leafs almost surely thought Tyler Biggs represented tremendous value at 22nd overall. So how do you separate ‘good’ trade-ups from ‘bad’ ones, without waiting 10 years for the results? Is it just based on how players were viewed pre-draft by rankings? If that’s the case, why have scouts at all - just trade up when someone drops well below their ranking on Bob McKenzie’s list.

The real answer is, there isn’t a way to do so with certainty. Trading down is an admission that you understand you’re more likely to fail than succeed in picking winners in the draft. Trading up is a vote of confidence in your scouts. Depending on your scouts, and the player, that could end up being a great call, or a terrible one.

Personally, I feel that the more you study a particular draft, the more you can get seduced by a prospect. You find someone you like, and start thinking of all the things that set them apart, and what make them a good pick. It’s easy to fall in love with players in the draft. They’re all good hockey players - you can talk yourself into any of them. You can start to believe that only you see the real potential in this player, how could everyone else be so dumb! Not to mention, it’s much sexier to be the hotshot who can pick the best prospects than to be the nerd counting draft picks. Perhaps that’s why teams get roped into what is, on average, not a smart decision. After all, no one at the high-stakes poker table thinks they’re the sucker.