So... how goes your rebuild?

[mr grieves]

So? given what we?ve seen this year, and before Lou does anything crazy, where do y?all think this rebuild is at?

Last summer, after the lottery, I offered what was then a wildly optimistic projection, not only because I dreamed of them signing Stamkos and using JvR to sort out the defense in the off season. For 2016-17, I imagined solid systems, good rookies, bounce-back years for unperforming vets, and average goaltending would get them to 85-90 points, then a wildcard spot next year, and finally contending the year after that.

During the season, as the rookies exceeded expectations, I pointed to some ways in which the team very closely echoed the 2007-8 Blackhawks. Our #1C looked at least as good as Toews, our 1W just as good as Kane (he then faded), our 2C as good as Sharp, and our fourth forward better than either Ladd or Byfugilen, etc. Others pointed out some pretty clear differences (neither Rielly nor Gardiner is Keith or even Seabrook, but then neither was Keith or Seabrook, as we think of them today, in 2007-8?).

But the 2016-17 Leafs blew by my optimistic prediction, even though they didn't make either of the big moves I'd hoped for (get Stamkos and another top 4 defenseman), and, even if individual contributions aren?t perfectly equivalent, they accomplished something even the Hawks didn?t manage their first season after drafting Toews and Kane. They made the playoffs. And took the best team in the league to 6 games in a pretty evenly played series.

That?s a high water mark without precedent, surely.

Not quite. Of all the teams who?ve bottomed out since the Hawks ? the Islanders, the Lightning, the Oilers, the Panthers, Buffalo, Edmonton yet again ? only one?s made the playoffs the season after drafting a franchise center or defenseman 1OA (give or take a spot): the 2013-14 Avalanche, who had an astonishing 45 point* improvement from 2012-13 to 2014-15 (the Leafs managed an almost(?)-franchise-record improvement of 26 points).

Now, that Colorado team was one that a lot of smart folks thought was poised to become perennial contender, though they had some serious reservations about Roy and the team?s system.

They, of course, fell apart? and hit a new low this season.

So, where do you think we are, measured against either the post-lockout contenders (Chicago, LA, Boston) or the teams who scraped the bottom of the standings and acquired top-end talent we all agree you need to build a contender (the aforementioned NYI, TBL, EDM, FLA, and BUF)?

 

[herman]

Weren't the 2013-14 Avs more a product of a fun PDO run?

Their adj-Corsi has been relatively stable in this (bad) quadrant for the past 5 years, and this was the only season when their result was vastly different than expected.

Here are the Leafs:
Last Carlyle/Horachek:

First Babcock:

This season:

Look at that steady march towards progress!

For funsies:
07-08 Chicago:

Then for the next 7 seasons: 3 Cups!

Until last year/this year: resulting in playoff stumbles

 

[mr grieves]

The blob is moving in the right direction!

What to make of all that? I didn't really think we were close the Colorado end of the spectrum. Even the talentless team last year had better coaching/systems, and now we've got the personnel. Barring serious injuries, I don't expect any huge steps back...

I guess I'm more curious about where we are along that Colorado (or 11-12 Edmonton) to Blackhawks spectrum. I'm sure, a couple months back, I was thinking of the Hawks comparable because they became a dyntasty -- but there are, of course, plenty of other teams with a similarly aged/ pedigreed core that didn't turn into the Hawks.

 

[herman]

For me personally, I haven't really gravitated towards team personnel comparisons, mostly because a) the metagame of the league is a moving target; b) the context of those players and their costs is a moving target; c) results appear to be more luuuuuuuck driven than most people are comfortable understanding this sport to be.

So it comes down to the essentials in team building and winning philosophy that I can only extend back to concepts I'm more familiar with: RPGs and card games. Luck is inherently built into those games (RNG rolls, card draw probabilities), and playing fields are generally equalized. Success depends on who/what you pick for your limited group, where you allocate limited resources, and how you deploy your strategy adjusting for current trends in play. How do you stack the odds to ever be in your favour?

The core of a championship caliber team build:
- metagame dominating or breaking: playing in a way that most other teams cannot easily account or have no answer for
- monopolizing opportunities: this translates to the concept of puck possession; and even if you grab more opportunities, can you make them count consistently?
- cost-efficiency: there's a beauty in cutting corners that doesn't hurt your win condition
- redundancy: can you survive or still win if one or more of your heavily relied upon players is rendered inactive/ineffective?

There's probably more, but it's not something I've thought of super formally. There's a push/pull to each element (that affects other elements) and every team faces those challenges.

 

[mr grieves]

For me personally, I haven't really gravitated towards team personnel comparisons, mostly because a) the metagame of the league is a moving target; b) the context of those players and their costs is a moving target; c) results appear to be more luuuuuuuck driven than most people are comfortable understanding this sport to be.

The results, yes, but we all use different stats to control for the randomness, assess play, and determine the probability that results will go your way. In principle, I think we could look at 2 rosters and compare the more meaningful stats to see how those teams stack up.

The RPG and card game inspired rules are interesting. I think they relate, without too much work, to some ways a lot of folks have been thinking about the game and building a team lately.

The core of a championship caliber team build:

- metagame dominating or breaking: playing in a way that most other teams cannot easily account or have no answer for = play style moves somewhat from championship to championship, but the general rule has been a possession-dominating style.

- monopolizing opportunities: this translates to the concept of puck possession; = that's the meta-game, broadly sketched

- and even if you grab more opportunities, can you make them count consistently? = skill level. Scoring's somewhat random year to year, but folks look at primary point rates, shooting percentage, etc to see past the randomness to get at chance-making and chance-finishing talent

- cost-efficiency: there's a beauty in cutting corners that doesn't hurt your win condition = there are plenty of ways to think of this, but, at its most general level, it's in the premise of the comparison
-- you need a core of young, cost-controlled players that'll hit their peaks at roughly the same time (20 or under 1C, 20 or under 1W, 2C who's not on a UFA contract, etc.)

- redundancy: can you survive or still win if one or more of your heavily relied upon players is rendered inactive/ineffective? = That goes two ways: overall depth, so chances of success are good even when your best players are rendered ineffective, and having enough core pieces that one being rendered inactive isn't the end of your season (that's why I'd think of the core as including more than 1C, 1D, 1G or any of the other "essential recipes to win the Cup"). 

 

[herman]

--snip--

Yeah, those game tenets definitely translate for the most part. Hockey will have additional nuances, obviously; no other team can have your exact player, for example, so merely controlling a premium player is an advantage unto itself as no other team has him.

The metagame is all about the psychology of the copy-cat league. Most will want to emulate the champs. Others want to ride ahead of the curve, or poke through the weaknesses of the championship build (which a large majority of the league will try to copy) with a specific anti-metagame strategy.

I like playing the middle ground, assessing what resources and players are currently in the fold, finding a stable general strategy that works best with them, and then tweaking the margins and outer shell to take advantage of the meta. This is because it costs quite a bit to completely revamp a build, but much less to adjust strategy.

In the Leafs' case, we have a glut of mobile, skilled wingers who appear to have oodles of passing and/or finishing capability coming up the pipeline. We play a structure that utilizes our speed to gain numerical advantages in isolation, and push for opportunities in areas of the ice where saves are harder to make. Our defense is largely mobile as well, used almost like secondary forwards to keep pressure bottled up in the offensive zone, with the trade off being less of the walling capability a traditional hockey defender would have. We have a lot of trouble against heavy possession teams when they get cycles set up, but we can usually win out against counter attack, transition teams. It sounds like the Leafs are developing their players for general completeness, rather than speciality, which makes depth redundancy a great deal easier, and merely spending more time to increase individual caliber.

 

[mr grieves]

Thinking about herman?s elements of a championship-calibre team, I was reminded of a neat article from last spring by Dom Luszczyszyn, "Eight Simple Rules for Building a Stanley Cup Contender." He looks at the cores of teams -- which he defines as a 1C, 1W, 2C, 4F, 1D, 2D, and G -- and uses wins above replacement (WAR) to compare teams that?ve made the final four since 2007-8.

He comes up with the following eight rules. 

1. The most important thing is a No. 1 center (gotta be elite, which is top 10% of the league, which works out to a WAR of 2+)

2. Your No. 1 center needs a wingman (also elite)

3. You definitely need an elite No. 1 D-man, but don?t stress too hard on the No. 2 guy (with caveat: ?WAR is a lot harder to quantify for D-men then it is forwards so that?s important to keep in mind here as some guys tend to get underrated because their skill-set is difficult to quantify, hence 2016-17 finalists' elite defensemen being Letang, Stralman, Vlasic, and Shattenkirk?).

4. He doesn?t have to be elite, but you usually need a strong second line center (that's within .5 of the top 10%, so WAR 1.5 and higher)

5. Goaltending doesn?t matter as much as you think (not for making the semi-finals, anyhow. He says goaltending can, in the end, be important?)

6. The average contender has four elite players and at least one other very good player (so, a really great 1C and 1D won?t do it ? minimally, you need another 2 elite forwards)

7. A solid core is more important than depth, but depth is more important than any one player (he notes: ?combined WAR of the non-core players on each of the final four teams from the past eight seasons was 5.3. Out of the 224 core players, only seven had a higher WAR?).

8. What separates the champions from the maybe-next-years? Centers, depth and goaltending

With the exception of goaltending, that?s a pretty uncontroversial list of things you need to contend. What distinguishes it from other approaches, I think, is that it doesn?t fixate on a few pieces ? 1C, 1D ? and embraces core depth (herman?s ?redundancy?). So, if I wanted to see where the Leafs were in the course of building a contender through the draft, I?d look to see which teams, over the past bunch of years, have tried to build in this way (Chicago, Colorado, and others), and I?d want to see how they fared ? not so much in results but in having the above-listed pieces ? in the season after bottoming out.

So, might use this thread to figure that out.

 

[CarltonTheBear]

Now, that Colorado team was one that a lot of smart folks thought was poised to become perennial contender, though they had some serious reservations about Roy and the team?s system.

They, of course, fell apart? and hit a new low this season.

Herman pointed this out already but "smart folks" seems pretty subjective here. I know a lot of "smart folks" who also completely predicted that crash for a variety of reasons. When you combine those with the losses of Paul Stastny and Ryan O'Reilly for essentially nothing then Colorado's downfall looks extremely predictable and entirely at the hands of their own management.

It still blows my mind that Colorado had one of the best centre depths in the league and essentially said "nah, this is TOO good. Let's destroy it". They let Stastny walk because they still had MacKinnon, Duchene, and O'Reilly. They traded O'Reilly for peanuts because they didn't want to pay him and still had MacKinnon and Duchene. Then they transitioned Duchene to the wing and are left with MacKinnon and scrubs.

 

[mr grieves]

Now, that Colorado team was one that a lot of smart folks thought was poised to become perennial contender, though they had some serious reservations about Roy and the team?s system.

They, of course, fell apart? and hit a new low this season.

Herman pointed this out already but "smart folks" seems pretty subjective here. I know a lot of "smart folks" who also completely predicted that crash for a variety of reasons. When you combine those with the losses of Paul Stastny and Ryan O'Reilly for essentially nothing then Colorado's downfall looks extremely predictable and entirely at the hands of their own management.

The "smart folks" consensus I was remembering was: Colorado is stacked at forward, might have the best center depth in the league, but whatever system they're playing would lead to disaster. I don't remember anyone anticipating the horrific mismanagement...

 

[CarltonTheBear]

... but whatever system they're playing would lead to disaster. I don't remember anyone anticipating the horrific mismanagement...

Well when a management is oblivious to those types of issues it generally says a lot about their level of incompetence.

 

[L K]

... but whatever system they're playing would lead to disaster. I don't remember anyone anticipating the horrific mismanagement...

Well when a management is oblivious to those types of issues it generally says a lot about their level of incompetence.

Maybe I'm misremembering things but I'm pretty sure there was plenty of talk for how the Avs were primed for failure.

 

[mr grieves]

... but whatever system they're playing would lead to disaster. I don't remember anyone anticipating the horrific mismanagement...

Well when a management is oblivious to those types of issues it generally says a lot about their level of incompetence.

That's true. We should've anticipated a group happy with the Corsi Roy's team managed would somehow screw up the talent they had.

 

[mr grieves]

Following the longer post above, the next thing to do is to define what the ?core? is. I?m not looking at fully formed contenders but emergent, wanna-be, might-be contenders, and this has some implications for how I define the core and thus slot players into it.

Those implications are:

First, since 1C and 1D are the most important spots and getting truly elite talent at one or both of those spots is most often done by picking at the top of the draft, I?ll only look at teams that?ve drafted one of these 1-2OA. I?m including CHI for obvious reasons. At some point, we?ll want to use stats to compare, and, since advanced stats start for the 2007-8 season, I?ll start with that season.

That gives me the following list of teams:
2007-08 CHI = Kane drafted 1OA in 2008 (Toews at #3 the previous year)
2007-08 STL = Erik Johnson
2008-09 LAK = Doughty
2009-10 NYI = Tavares
2009-10 TBL = Hedman (Stamkos drafted 1OA in 2008)
2011-12 EDM = RNH (winger Hall drafted 1st year before)
2012-13 CBJ = Ryan Murray (2nd? Yak was #1)
2013-14 COL = MacKinnon
2014-15 FLA = Ekblad (Barkov drafted at #2 previous year)
2015-16 EDM = McDavid arrives
2015-16 BUF = Eichel
2016-17 TML = Matthews

Second, we need to fill Dom?s key roster spots: 1C, 1W, 2C, 4F, 1D, 2D, and G. That?s your core. I?d like to see what candidates each of the above teams have to fill out the rest of the core after they bottomed out and got the coveted 1C or 1D. That?d let us see how the Leafs stack up against teams who?ve (in theory) gone from tanking to ascending the ranks to, if they?re lucky/smart, becoming contenders. Since most of the above teams were bad for a while before they drafted very high, they should have candidates in the system to fill our those core spots. At the outset, then, I'll eliminate STL, since that team didn?t draft very high often and was built differently (trades, etc.) than the Leafs and other comparable teams.

Third, for herman?s cost-efficiency, we?re looking for a core that?ll grow and contend together during their cost-controlled/peak years. So,  we should look for guys who fill these spots out that are within certain age ranges. For forwards, in the first year of building up, they should be mostly between 18 and 22, especially for the positions where the truly elite ought to be (1C and 1W), though we can add a couple guys in their mid/late 20s, as they?d be relatively affordable and not declining with age by the time they contend (when most of the core is 21-24). For defense, let?s say we want em young and elite (Doughty?s and Ekblad?s are ideal), but will include older guys since defensemen take longer to develop and really high picks that are immediately impact defensemen have been relatively rare. I wouldn?t include guys older than 28, because, by the time the young forwards are in their peaks, these old defensemen are getting worse.

Fourth, to Dom?s core list, I?d add a veteran forward spot, since that seems to be thing teams add to put them over the top (e.g. Hossa in Chicago, Justin Williams in LA?). I guess they could be counted as what he calls non-core depth, but I?m including these anyway, with the expectation that some of these guys are upgraded (for the ones who turn into contenders, like Chicago getting Hossa to replace Havlator whoever) and lost/not replaced for the coulda-beens (Colorado).

So, the cores I?d be looking at:
1C = 18-22yo
1W = 18-22yo
2C = 22-26yo (younger and 1C-calibre ideally)
4F = 18-22yo
VF = 25-28yo
1D = 18-26yo
2D = 20-24yo

Fifth, I won?t sweat depth. The way I?m defining the core means teams should have the cap space to assemble cheap, efficient depth, but what they assemble seems to change from year to year as teams retool to maximize their chances of winning the Cup (and here?s where herman?s meta-game most often comes in, I think). No point trying to figure that one out at this point.

Having figured out what defines a ?core,? or the beginnings of the same. the next task is to determine what stats are worth looking at to compare what each team has in each position. Ideally, I can find some stats that are predictive of future success, not subject to the randomness of the game. That'll be the concern of the next post...

 

[mr grieves]

Having figured out what defines a ?core,? or the beginnings of the same, the next task is to determine what stats are worth looking at to compare what each team has in each position. Most of the players being compared are young and not yet in their prime, so the goal is to have some stats that are able to tell us not only whether Matthews? rookie season was ?better? in its outcome than MacKinnon?s but also are relatively reliable for projecting future success. 

These are the ones that I think are most useful?

If we want to see how promising the cores are, we will eventually need some stats to compare.

= Basic stats -- Goals, points, TOI ? all seem pretty obvious.

= ?monopolizing chances? ? the consensus these days is that goal and points are somewhat random and possession stats are the most important for assessing overall play. So CF% and rel.CF would be nice to have. I?d also include xGF% to capture how many scoring chances are generated and given up.

= ?making chances count? ? We can maybe see *how* random scoring is by including SH% and P160. The shooting percentage can tell us if a player is under/over-performing or if his scoring is sustainable. The latter number weeds out folks who got lucky piling up secondary assists. 

= Contextualizing performance ? some players are given responsibilities that others aren?t. To see how much a coach trusts players, would look not only to TOI but also adjustments of Corsi that account for quality of competition (CF.QoC) and offensive zone starts (OZS%), which can tell us when players are being sheltered.

= Superstats ? these single-number measures of performance that Dom uses in the piece linked above and elsewhere include a bunch of stuff, including goals and assists. Dom's site and Athletic work uses Game Score (see image of Leafs roster below), but that (from what I've read) tends to overrate results/production somewhat. Goals Above Replacement (GAR) and Wins Above Replacement (WAR = GAR/6) are more sensitive to play and can be broken in measures of Special Teams play, Even Strength offense and defense (see Nylander card below). One or more of those would give a decent, overall sense of play. Also, some work has been done on how, on average, a stat like WAR changes over time.

So? that?d be my method of comparing players who are pieces in builds toward contention. Would be happy to hear which other stats folks think are useful for assessing individual players (and which of the above are not and can be cut).

Next up is looking at the different core positions listed above and identifying what each team has in that position.

 

[ZBBM]

I gotta say, the amount of work you guys put into this site is astounding.  I didn't bother to put together this much information when I submitted the dissertation for my third PhD (the one in Theoretical Astrophysics).

 

[herman]

mr. grieves, those would be the stats I'm currently looking at to assess our own team (I don't have the bandwidth to go league-wide other than potential trade targets). The ixGF% vs iGF% differential is an interesting measure to me in finding out which players have shooting talent (I.e. Can expect to have higher than average Sh%).

I gotta say, the amount of work you guys put into this site is astounding.  I didn't bother to put together this much information when I submitted the dissertation for my third PhD (the one in Theoretical Astrophysics).

Which branch of Theoretical Astrophysics?

 

[ZBBM]

mr. grieves, those would be the stats I'm currently looking at to assess our own team (I don't have the bandwidth to go league-wide other than potential trade targets). The ixGF% vs iGF% differential is an interesting measure to me in finding out which players have shooting talent (I.e. Can expect to have higher than average Sh%).

I gotta say, the amount of work you guys put into this site is astounding.  I didn't bother to put together this much information when I submitted the dissertation for my third PhD (the one in Theoretical Astrophysics).

Which branch of Theoretical Astrophysics?

The formation of Dark Matter -- within the post-Guthian Inflationary Paradigm, of course.  I felt it was an excellent complement to my second doctorate, in Hotel Management.

 

[herman]

The formation of Dark Matter -- within the post-Guthian Inflationary Paradigm, of course.  I felt it was an excellent complement to my second doctorate, in Hotel Management.

WIMPs? Sounds like it covers the spread nicely with some transferable skills.

 

[ZBBM]

The formation of Dark Matter -- within the post-Guthian Inflationary Paradigm, of course.  I felt it was an excellent complement to my second doctorate, in Hotel Management.

WIMPs? Sounds like it covers the spread nicely with some transferable skills.

Admittedly, my first one in Elizabethan Literature has been a bit of a white elephant.

 

[mr grieves]

I gotta say, the amount of work you guys put into this site is astounding.  I didn't bother to put together this much information when I submitted the dissertation for my third PhD (the one in Theoretical Astrophysics).

Well, sometimes a fellow needs little fun research assignments to give a break from the bigger one (a dissertation, as it happens).

So, next week... #1 centers!