Author Topic: So... how goes your rebuild?  (Read 3456 times)

0 Members and 1 Guest are viewing this topic.

Online herman

  • All Star
  • *****
  • Posts: 7816
  • Gender: Male
    • View Profile
Re: So... how goes your rebuild?
« Reply #15 on: May 05, 2017, 12:54:34 PM »
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?

Online Zanzibar Buck-Buck McFate

  • Sittler Status
  • ******
  • Posts: 12200
    • View Profile
Re: So... how goes your rebuild?
« Reply #16 on: May 05, 2017, 02:03:24 PM »
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.

Online herman

  • All Star
  • *****
  • Posts: 7816
  • Gender: Male
    • View Profile
Re: So... how goes your rebuild?
« Reply #17 on: May 05, 2017, 02:19:38 PM »

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.
« Last Edit: May 05, 2017, 02:22:06 PM by herman »

Online Zanzibar Buck-Buck McFate

  • Sittler Status
  • ******
  • Posts: 12200
    • View Profile
Re: So... how goes your rebuild?
« Reply #18 on: May 05, 2017, 02:33:21 PM »

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.

Offline mr grieves

  • Veteran
  • ****
  • Posts: 1457
    • View Profile
Re: So... how goes your rebuild?
« Reply #19 on: May 05, 2017, 03:55:42 PM »
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!

Online Zanzibar Buck-Buck McFate

  • Sittler Status
  • ******
  • Posts: 12200
    • View Profile
Re: So... how goes your rebuild?
« Reply #20 on: May 05, 2017, 05:14:49 PM »
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!

Ha, cheers to you MG.  And good luck on your doctorate.  Mine are all fake.............. :-[ :-X :-\ :-* :'( :P :P :P :P ::) 8)

Online Significantly Insignificant

  • Veteran
  • ****
  • Posts: 2440
    • View Profile
Re: So... how goes your rebuild?
« Reply #21 on: May 05, 2017, 09:22:22 PM »
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!

Ha, cheers to you MG.  And good luck on your doctorate.  Mine are all fake.............. :-[ :-X :-\ :-* :'( :P :P :P :P ::) 8)

You....lied to us?  There is no emoticon for what I am feeling right now.
"Progress lies not in enhancing what is, but in advancing toward what will be.” - Khalil Gibran

Offline mr grieves

  • Veteran
  • ****
  • Posts: 1457
    • View Profile
Re: So... how goes your rebuild?
« Reply #22 on: May 08, 2017, 10:10:10 AM »
1C -- Day one: vitals & boxcar stats

All numbers are all situations.

Below's what each of the rebuilding teams I defined as "comparable" above (1C or 1D at top of draft) had at the 1C slot. Ages are in red if they call outside the 18-19yo definition, and, when players aren't rookies, their D+ year is in red.



This is the easy one.

So far, looks like Matthews is about as good scoring 1C as has been drafted in the last decade. He at least had the best season of any.

He shoots nearly as much as Stamkos, but doesn't have that crazy SH% (which Stamkos has, of course, pretty much sustained). Matthews is also one year behind in his development (in his D+1 year Stamkos scored 23-23-46 in 79 games). Take out the guy with a year's head start, and Matthews leads the way.

Almost. McDavid's not, looked at this way, on an entirely other level. But then he was injured.
« Last Edit: May 08, 2017, 10:13:33 AM by mr grieves »

Offline mr grieves

  • Veteran
  • ****
  • Posts: 1457
    • View Profile
Re: So... how goes your rebuild?
« Reply #23 on: May 09, 2017, 06:56:50 PM »
1C — Day two: 5v5 Production

Same folks as in the previous post, but now isolating scoring at even-strength. The table is more individual at the left than at the right, which offers the on-ice scoring numbers. Gonna move right to left, more or less.



Stamkos is still very impressive, but, take out the PP, and he’s now among mere mortals (with MacDavid and Matthews at ~2 P160). Even that insane on-ice GF60 appears to be inflated by luck (over 1 goal more than expected!). Much the same seems true of Toews, so it might have to do with these centers being paired with established talent (Havlat and St. Louis).

If we privilege results stats like G60, P160, and GF60, none of Tavares, Barkov, or Eichel did much at 5v5. Adding in xGF60 we can see MacKinnon seems to have got pretty lucky, but RNH was both lucky and good. Tavares had an unlucky year, and Matthews is in similar boat, though other numbers are much stronger: the GF60 is middling, but he’s 2nd in xGF60 and leads P160.

McDavid, of course, is the most impressive of the bunch, leading the way across the board.

Or nearly so.

While Matthews doesn’t really stand out among the 1Cs for on-ice GF60, he destroys in goals scored. And his iG isn’t that far off from his ixG, which suggests he didn’t get especially lucky. He is, I think, something the league hasn’t had in a long while — an elite shooting 1C who, beyond just having a great shot, can do it on cluttered ice during regular play. I guess that's why Granato had to go back to, like, Sakic for a comparable.

Verdict? McDavid is terrifyingly good. The more seasoned 1Cs (Kopitar, Stamkos, Johansen) all look solid. Some of MacKinnon's performance looks lucky, and subsequent seasons have born that out. The only guys that I'd worry don't have the scoring ability to be plausible 1Cs are Eichel and Barkov. As for the Leafs' 1C? Matthews at 5v5 is as close as any current post-Crosby/Malkin center is likely to get to McDavid.

The xGF% measures chances generated and given up, so they’re properly discussed in the next post…

Tomorrow, 5v5 play/ defense.

Offline mr grieves

  • Veteran
  • ****
  • Posts: 1457
    • View Profile
Re: So... how goes your rebuild?
« Reply #24 on: May 10, 2017, 08:32:42 PM »
1C Day three: 5v5 Play/Defense

Same folks as in previous post, but now isolating play at evens, using advanced stats. All the stats are pretty familiar, and they’ve got their problems. Nik’s had convincing arguments against CF.QoC and zone starts, IIRC. I wonder how reliable Corsica’s numbers for scoring chances are for games nearly ten years ago, and that’d call xGF and KPO into question…

One less familiar stats were included to try to capture defensive play. KPO, or Keeping Pucks Outside, represents what proportion of Corsi against were NOT scoring chances. Some have said it can actually capture defensive ability, particularly for those stay-at-home types. According to Original Six Analytics, 85% is about average, 83.5 is a Std Dev below, 87% is a Std Dev above, 82% is bad, 89% about as good as you can get (his distribution chart is for defensemen, but he says the pattern is the same for forwards).

Table:



Thoughts:

Standing out in a good way are Kopitar, Johansen, Barkov, and McDavid.

The latter two were teenaged rookies. That McDavid was so good without the puck so early isn’t a surprise. I did not know, however, that Barkov was such a natural defensive phenom.   

So, that leaves Matthews in the company of rookies Toews, Tavares, Nugent-Hopkins, MacKinnon, and Eichel. Matthews has better xGF% and CF% than any of them. In fact, he’s the only one above 50% in both…

The first two positive standouts  — Kopitar and Johansen — were in their 3rd seasons, and I’d suspect gameplay stats improve as talented high picks develop. All that about skill guys being so good that they don’t need to learn to play without the puck until the NHL. Corsica's stats don’t go back to Kopitar’s rookie year, but they do for Johansen. His rookie numbers were in line with Tavares’s above — and Tavares’s were, by his 3rd season (11-12), nearly Kopitar’s. Toews, in 09-10, was up to a 60% CF/xGF in his third season...

Dare we hope Matthews follows Tavares’s path and turns in a Kopitar-ish two-way season at some point? Matthews’s impressive two-way performance is undercut, to some degree, by his only having to play one way for most of the season: his ZSR indicates how sheltered he was. On the other hand, he faced the same quality of opponents as Toews and Tavares, and the difference between Matthews’s CF% and theirs is about what folks say the ZSR difference would produce.

So… verdict here? He’s probably almost certainly a better 1C defensively than Stamkos, RNH, and MacKinnon but probably not as good as Barkov. About as good, defensively, as Tavares would make for a 1C I'd be happy with.

Last day at 1C tomorrow. Topic? Those single-number metrics (Game Score and GAR) and percentiles. 

Offline freer

  • Veteran
  • ****
  • Posts: 2328
  • Gender: Male
    • View Profile
Re: So... how goes your rebuild?
« Reply #25 on: May 11, 2017, 09:14:13 PM »
IMO

I think that we are only one good D man and back up goalie, from being a real threat.

Offline mr grieves

  • Veteran
  • ****
  • Posts: 1457
    • View Profile
Re: So... how goes your rebuild?
« Reply #26 on: May 12, 2017, 08:31:58 PM »
1C: Day four: Single-number performance stats

Lastly… a comparison of Goals Above Replacement (GAR) and Game Score. Both factor in actual results (goals for and against) more than the other stats considered so far, so I’m not sure if I’d trust them more than numbers that isolate gameplay (Corsi) and scoring chances (xGF/GA). On the other hand, they both factor in things that the other tables haven’t, most notably penalties drawn and taken.

The main reason I’m curious about these is because they can give a quick sense of how performances compare to the rest of the league. Game Score has cut-offs that correspond to different lines (see that Maple Leafs line-up a few post up.. and here’s the piece from The Athletic’s Dom L on the stat). GAR separates out Even Strength Offense and Defense, PP Offense, and more, and those who use the stat look at percentiles that players fall into (see that Nylander chart in the post above). 

So, the last table:



A few things:

For Game Score, “Elite” seasons don’t often happen for rookies. And RNH and MacKinnon’s 1L seasons are pretty borderline (under .8). Before McDavid and Matthews, the best rookie season by game score was Toews’s in 07-08. — both McDavid (no surprise) and Matthews have had rookie seasons in the same ballpark as Kopitar’s and Stamkos’s second+ seasons.

Similarly, the Overall GAR scores for rookies seem pretty inflated by PP performance, and it looks hard to hit 1L-er level in even-strength play early in one’s first year. Only McDavid’s done it. But Matthews was pretty close to the 76th percentile cut off.

No GAR from before the 2008-9 season, unfortunately, so who knows what the deal with Toews is.


1C Day five: Conclusion

So… to sum up the week, on what the Leafs have got at 1C vs what other teams have had:

The method of choosing teams to compare — high pick C or D — more or less ensured that everyone here would be decent. So, there aren’t really any bad centers, and there really isn't any team here where I'd say "well, they don't have a first-line center," though most have some holes in their games or drawbacks. MacKinnon doesn’t look very good defensively, while Barkov doesn’t seem to have much scoring talent. The worst that can be said for Matthews is that he wasn’t really tested (those zone starts), but otherwise there isn’t a center here, besides McDavid, I’d take over him.

So, to go back to the 8 Simple Rules:

1. The most important thing is an elite No. 1 center

We have a center that’s well on his way to becoming that. In his rookie year, he got closer to it than Tavares, Stamkos, and Toews.

Next up, the 1W.

Offline mr grieves

  • Veteran
  • ****
  • Posts: 1457
    • View Profile
Re: So... how goes your rebuild?
« Reply #27 on: May 16, 2017, 03:09:19 PM »
1W — Day one: Boxcar stats, core ages, first impressions

Below is a table of the same teams from last week, but now comparing 1W, the second thing Dom says your Contending Core needs.

Rookie seasons have the D+year highlighted in green.

While 8 of the 11 centers considered last week were rookies, only 4 wingers — Kane, Jenner, Reinhart, and Marner — are rookies, of these only Kane and Marner were teenagers.

Also, while all the centers considered were given top-6 minutes, Jenner played behind Gaborik (31), Umberger (31), Foligno (26), Horton (28).





Ages and Comparable Cores

First thing to note here is that there a few outliers from the parameters I’d set earlier. So, probably need to reassess which teams are really comparable — high picks for a 1C and/or 1D probably isn’t enough.

Most glaringly, TBL poses a problem. St. Louis is indisputably the best winger on the team, but he’s in his mid-30s. The next youngest wingers on the team were Teddy Purcell and Steve Downie, neither whom played the minutes or have the pedigree to qualify as potential 1Ws. St. Louis played top-line minutes and was really productive, both in this season (2009-10) and the next, when they went to the ECF.

But his being unavoidable here is a reminder that TBL didn’t rebuild so much as it retooled: they didn’t really turnover their last core (jettisoning St. Louis, LeCavalier, Malone, etc) to get bad in order to get good (Stamkos and Hedman) and their present core is largely drawn from successful later picks who took a few extra years to make it to the NHL. As a result, those players don’t appear in this comparison. And, with respect to what this comparison wants to assess, TBL has greater age distribution among core pieces than is ideal, which probably has contributed to their salary cap troubles.

LAK’s Dustin Brown is a bit older than the 1W we’ve been targeting. That too has problem led to some of LAK’s recent troubles — big contract for older and declining Brown hasn’t helped them stay contending — but they did manage a window where they won two Cups, so I’m inclined to keep them.

The upshot is I should’ve looked at the overall roster ages to pick cores to compare to that of the 2016-17 Leafs. Below chart is ages: the team’s average and (to keep the Jagr’s and Weight’s from throwing things off) the median, difference between these and league average.



So, I’m keeping teams with a median age over a year younger than the league average will count and dropping TBL. No point including LeCavalier and Ryan Malone in the upcoming weeks.


First Impressions

From this distance, the three 1Ws that stand out are Kane, Hall, and Landeskog. Marner slots in one tier down. But it remains to be seen whether Marner’s apparent mediocrity — or mere ‘goodness’ — has to do with inflated PP scoring for those top 3.

So, tomorrow… 5v5 production.

Online herman

  • All Star
  • *****
  • Posts: 7816
  • Gender: Male
    • View Profile
Re: So... how goes your rebuild?
« Reply #28 on: May 16, 2017, 04:27:37 PM »
Nylander?

Offline mr grieves

  • Veteran
  • ****
  • Posts: 1457
    • View Profile
Re: So... how goes your rebuild?
« Reply #29 on: May 16, 2017, 07:11:07 PM »
Nylander?

4th forward?

1C - Matthews
1W - Marner
2C - Kadri
4F - Nylander
1D - Rielly
2D - Gardiner

Hopefully he's not traded before I get there (July?)

TMLfans.ca

Re: So... how goes your rebuild?
« Reply #29 on: May 16, 2017, 07:11:07 PM »