Before we start off on our quest for the weakest link on the Leafs, I have a confession.  This is the only thing I know about the television game show that has made that proverb up above famous:

We aren’t going to fry anyone with death rays here. Instead, this is a very easy quest.  You get to pick now who you think is the weakest link on the Leafs’ roster.  I can imagine some of your selections, so we’ll do a pretend poll where I say:

  1. Roman Polak
  2. Matt Hunwick
  3. Matt Martin
  4. Zach Hyman
  5. Curtis McElhinney (because I usually have my own answer somewhere in a poll)
  6. Tyler Bozak
  7. Ben Smith — he’s still there! Just because he’s recovering from surgery and not playing doesn’t mean he might not come back, OMG!

And you pick one, and we each look at the results and scoff at how wrong everyone else is.

What Really Matters

Now that we’ve found our weakest link, and we can each have a different one, it doesn’t matter, we need to ask if it matters if we do fry him with a death ray.  Or less dramatically put, sit him in the pressbox with Josh Leivo or trade him to Vegas.

That was exactly the question that Alex Novet at Hockey Graphs set out to answer.  Does it matter if you change up your depth?  Strong and Weak Links: Talent Distribution within Teams is very much worth your time.  Do not fear the math! You can go and read the whole article and easily grasp the point of it. As always, I’ll wait.

First, as the article explains in detail, you need to understand the concept of strong- and weak-link sports.  The strong link is the best player, the weakest link the worst.  A strong-link sport is one in which a team is most improved by improving their best player and a weak-link sport is one in which improving the weakest player makes the most difference to team success.

Novet uses a range of techniques, including business analysis methods, to determine which sort hockey is.  He determines that it is a strong-link sport, and methods aside, it is a premise that most people will intuitively support:

Hockey is a strong link game. Getting the very best players is essential to success. Phrased this way, it sounds obvious. But the above shows that this is the case even at the cost of creating weaknesses elsewhere in the lineup. This has implications for many of the major decisions that general managers make.

To put this concept in context: having Auston Matthews is more important than getting rid of Ben Smith.

Well, duh.

So, to get Auston Matthews, you tank.  Novet’s analysis fully supports the idea that tanking in hockey is a good and beneficial tactic, possibly a required tactic, unless you can come accidentally last like the Avalanche have done.  Annecdotally, Fulemin looked at how cup-winning teams have high draft picks over here.

Novet’s analysis also says that stressing over your worst player and spending resources to upgrade him is likely a waste of those resources.  The Leafs could have replaced Ben Smith with Byron Froese or Frederik Gauthier and it would not have made the team worse in a meaningful way.  We should not expect that swapping Smith for Brian Boyle will change game outcomes for the better in a meaningful way despite how many have hopes for that very thing.

We might have more fun watching the games with Boyle in them—I do—and his Corsi will almost certainly be better, but the reality is, it won’t make the team better by an amount that matters.  This is not an indictment of the trade. I think the trade was about things not precisely related to game outcomes, and it was a fine move to make, in part because it raised the overall value of the team as a whole.  That’s still a good thing, and if you can do that for low cost, you should. But to get off the bubble and into contender range, you need to focus on the top end.

This is where this gets complicated.  Because Novet’s analysis also ties into other work that shows that total team worth (as measured by whatever single number type stat you like) is important for success.  At the same time, what his analysis zeros in on is that an unbalanced lineup is necessary to win in the NHL.  There is a tension between these two realities and finding the balance is difficult.

The salary cap makes it inevitable that to have very strong links on one end of your chain, you have to have weaker, and therefore cheaper ones on the other.  And you can only cheat that system so much with players on ELCs.  Most of them are not yet very strong.

A team cannot be Matthews and 22 Ben Smiths, but 23 Connor Browns won’t get you a cup.

Novet mentions the Carolina Hurricanes as an example of a balanced lineup that is good overall but can’t win. I might put the Los Angeles Kings in that bin too.  And I would put the Colorado Avalanche and the Arizona Coyotes down next to them as unbalanced lineups that are just too bad overall to win.  The Vancouver Canucks go in that bin as well.

That’s their problems to solve though. We want to make the Leafs into a contender.  How can this information help us?

The concept here is that changes of equal value at different ends of the scale have different effects.  Some of the why of that is obvious: ice time, usage, etc.  But to put it into real terms, imagine this: Brian Boyle is exactly 6.7 player value units better than Smith, and Matthews is exactly 6.7 PVU better than the player the Leafs would have drafted if they’d lost the lottery. By getting Matthews they improved the team by much more than getting Boyle has.

You want to put your money where it has the most bang for the buck.  Great in theory, but what isn’t clearly understood is how deep into a lineup that strong-link boost effect travels.  The point at which the improvement just raises the overall value of the team and has no extra kick to it is not something anyone can point at.

I might want to say, “Ah ha! You don’t need to change Zach Hyman for a better player, it’s much more important to sign Kevin Shattenkirk instead.” And you might say, “Wow, you sound like Fulemin all of a sudden.”  And we might all agree that if it you had to choose one or the other, Shattenkirk would be the better option.

But take a trickier choice: should the Leafs swap out Leo Komarov or James van Riemsdyk now?  Should they trade a player as good as van Riemsdyk and bring in a prospect of lesser current value, or should they trade the declining Komarov and upgrade him?  Which path should you pursue?  (This question assumes you aren’t going to get a major asset back for van Riemsdyk, more on that idea is over here.)

To make this question harder: what if the Leafs did both?  That is the right question anyway; it’s just a matter of time before both players get moved.  What if doing both left the Leafs with the identical overall player value when they were done?  Does an upgrade to Komarov give you the strong-link kick?  Is a downgrade on van Reimsdyk worse than it seems because he is a such a strong link?

As an aside in Novet’s piece, which focuses on overall team strength, he notes that this concept might apply at the forward line level:

These findings also nicely complement DTMAboutHeart’s RITHAC presentation, in which he suggested it is better to spread out top players on different lines rather than putting them together. That suggests that individual matchups are also strong link games, and it is best to have as strong a player as possible on the ice at all times.

If we look at the Leafs top nine, we see:

  1. Matthews
  2. Kadri and Nylander
  3. Marner and van Riemsdyk

Those are my very strong players on each line.

Then there’s the defence:

  1. Rielly
  2. Gardiner
  3. Zaitsev ??

Novet’s work didn’t dig into the very meaty and interesting issue of measuring forward and defender changes separately. But considering how much time they spend on the ice, defenders might be more important, not less.

The easy conclusions here are, obviously, easy.  Worrying about better third pairing defenders isn’t going to push the team to the next level. The lamented deadline trades that never happened for Roman Polak and Mat Hunwick would not have made the team a contender or even much better.  Exposing Matt Martin and playing [your favourite prospect here] won’t help.  It wouldn’t hurt, either, but it isn’t necessary for success.

The tough question is: should the Leafs put a lot of resources into Shattenkirk? Because now we’re finally talking about the important thing, which is money. All of this is about allocation of resources. Are there enough resources to get, not just a stronger link at the top end of the defence, but to also improve the top nine so there are always two really good players on the ice at all times?

This is the other lesson here.  Trading out ageing players for average-quality prospects is only okay if they’re your depth players. Even some decline in overall current value is okay there.  But if you never add very good players at the top end, you never get better.

The best thing about this for the Leafs is that most of their strong links are still climbing the growth curve, and so Mitch Marner being one year older and better next year is very meaningful, whereas the decline as Matt Martin ages is not.

Conclusion

Do I have one? Sort of. I’m still thinking about this.  But, here goes.

The Leafs have to move on from Bozak, Komarov and van Riemsdyk at some point soon.  They could let all three walk as UFAs with zero return and, as long as they acquire one strong-link forward, they’ve lost no ground.  Bozak is a high-quality points scorer, but his overall value is just not quite enough that he can’t be replaced at minimal expenditure or with normal growth by multiple other players in the system.

However, if the Leafs trade van Riemsdyk for futures (picks and prospects) now or at the next deadline, and they do not add someone of equal or greater high value from outside the organization, they have taken a step back greater than his loss seems on the surface. There is no one in the system who is, right now, his equal. His defensive weakness does not come close to negating his offensive contribution, which is prodigious. Trading him for the sake of trading him would be the biggest mistake the Leafs could make this summer.

Trading Komarov and Bozak for futures when they are at peak value is likely a good idea.  Van Riemsdyk has to stay, perhaps through to the end of his contract, until that replacement is found.  This is a change in my opinion on this. I was thinking it was Komarov and his defensive skills and Bozak and his ‘at least he’s an effective centre’ status that needed to be kept.  I’m moving away from that idea.

I think the Leafs should have six strong-link forwards, not five. I think they need to add one over and above replacing van Riemsdyk, but this might be asking for too much right away. However, I’d offer a very large pile of medium quality middle six forward or middle pairing defender players/prospects for Gabriel Landeskog.

The biggest impact this article had on my thinking was that I value those muddy middle players even less than I did before, and the Leafs have lots of them.  The Leafs have more than enough to keep the marginal positions filled at a good enough level for a long time.  Using some of them in a reverse Kessel trade seems the only way the Leafs can land any high-end forwards.

Zach Hyman can play in the top nine forever if that kind of upgrade occurs.

I think the Leafs need a very strong-link defender, but they don’t grow on trees.  However, the cap hit of Komarov plus Bozak is equal to about what they’d need to sign Shattenkirk.  The problem with him is his term, much like with Stamkos, but it might be a worthier gamble.  If the Leafs can’t get Shattenkirk or pry someone out of Anaheim who is at least high-ish end, then a stop-gap like Cody Franson is a good idea until the market, or fate, delivers another option.

I think the Leafs should pass on signing Brian Boyle even if he wants a deal, and I think they should worry about their third pairing defenders on the twelfth of never.

My central idea is that if the Leafs tinker with the edges around the core they have now, they will not get better enough to do more than maintain their position as a mediocre team.  They will become less unbalanced, and the overall strength of the team may rise, but it won’t translate into wins.

The Pittsburgh Penguins won that Stanley Cup because they added Phil Kessel to a team full of strong links, not because they signed Matt Cullen.

The GMs who make the easy choices, even if they are good decisions, to the exclusion of the hard, meaningful choices.  They’re the guys who tinker with the third pair and upgrade their depth.  They assume the core is fine as they age into decline and get replaced with players only as good as those retirees were in their dotage. They gradually, slowly, let the team get worse and worse, even as they seem to be doing smart things.  They are the weakest link because no one ever fires them until the ship has sunk all the way to the bottom of the sea.

So far the Leafs haven’t done this. They are only just embarking on the acquisition phase of their rebuild, but if they don’t add a high-value player before they start losing their stronger-link veterans, they will be heading in that direction.

Analysis