West Coast Buffet: Round 2 Mutual +/- Report, Round 3 Cross-Mutual +/-, and Food Festival
Well that was fun, wasn't it? If not heart attack inducing. So before we move on from genus Orcinus to superorder Selachimorpha (sweaters, not sushi), let's take some time to digest what we've had so far.
![]() Mmmmm... barbeque sauce! |
If you've been following my FanPosts, you should be familiar with my Mutual +/- statistic for tracking line/pairing chemistry:
Of course, we all know how often our lines have been shuffled by Evil Genius Q, and perhaps this time for the better. We got a bonafide chicken checking line that shut down the Sedins to go along with the emergence of Buff McNugget as Vancouver fans now must be having nightmares about.![]()
Without further ado...
Round 2 Mutual +/-
For your enjoyment, this time I've provided a "Mutual TOI" chart if you're obsessed with this enough to care about how much time the players spent on the ice together (you can actually see the lines pretty well with this).
Observations:
- Brent Seabrook really does bring out the Norris qualities of Duncan Keith, doesn't he? But Biscuit had a monster series in his own right.
- Niklas Hjalmarsson experienced a little reversion to the mean this time around. But don't blame Soupy for it.
- In his limited time, it would appear Troy Brouwer fit well on the Sharpie-Panda line and not so much with Daydream Nation.
- Speaking of which, getting united with the Captain and 20 Cent was the best thing that could've happened to Dustin Byfuglien.
- Despite all the good things that happened to Dave Bolland, it didn't show up on this sheet... this statistic is horrible for assessing checking lines.*
- Notice all the zeros? This is a sign that Coach Q actually stuck with his lines, for the most part.**
- Also, I know I left out Cristobal Huet. The only goal that was scored while he was in net was a power play goal (!) so he would've been zeros all the way across.
* One thing I noticed recently (and much too late) was that Behind the Net has a fantastic piece on plus-minus that addresses deficiencies like this... and how to adjust by the quality of opposition and the quality of teammates (the "BtN rating" is +/- adjusted by the rest of the team). Hawerchuk suggests adjusting by what I called Mutual TOI, and the spirit of what he did is similar to what I'm doing, but on a bigger scale -- and it's probably more meaningful.
** The sample size is still small and comparing Rounds 1 and 2 may not be hugely meaningful... but for the curious, I added up the charts for both rounds to get Rounds 1 & 2 Mutual +/-.

Round 3 Cross-Mutual +/-
The question came up on one of the game threads -- which of the Blackhawks lines is going to match up with the Sharks lines, in particularly those belonging to Big and Little Joe?
I promised I'd try to adapt my statistic and mightily morph it into what I'm calling Cross-Mutual +/-, a measure of +/- for each 'Hawk when facing each Shark. I made this from the 'Hawks point of view, so a plus for one of our guys is a minus for one of theirs; here's what I came up with (click to see full names for the Sharks):
...and a Cross-Mutual TOI chart for this, too.
(keep in mind that the 'Hawks took the season series 3-1-0 and outscored the Sharks 17-11 in the process; the playoff version of these numbers are unlikely to end up looking like this)
Observations:
- Based solely on these numbers, we should be matching Toews-Brouwer-Ladd with the Sharks top line. Interesting how it kind of works with the "size" argument that's been thrown around.
- That said, our top line numbers seem to be coming from the DDN+Brouwer line doing well as a whole and Patrick Kane being his usual lazy self.
- If we throw our second line at them instead, Patrick Sharp has been a minus player against that top line this season. Might not mean much, but keep that in mind (also remember he was filling for Weapon centering the second line back then, too).
- As for D men against "HTML" (geek alert! must be a Bay Area thing... oh wait.), you can't go wrong with Marlboro 72, but Brian Campbell has done well against the "team he was with for half a season." His ability to move the puck up is going to be real helpful if we try to fight fire with fire and not get bogged down in the corners (as people on Fear the Fin have said would be the way to play us).
- This all came before Bolland came back, but Bolland-Ladd-Versteeg may be worth trying.
- Watch out for Joe Pavelski. Nobody on our roster stands out as having really shut him down.
- Dan Boyle and Douglas Murray haven't fared well against the 'Hawks. It's no secret that the Sharks blue line, are one of the keys to the series.
- Of course, they also haven't faced Antti Niemi before, so how
they handle himhe handles them could be very interesting. - With only four games, this sample size is small... hence why doing a "Mutual Corsi" would be nice (but alas, a lot more work). So take this chart, and whatever you glean from it, with a grain of salt.
That's not all. But the rest is up to you guys -- I hope this post provides some good discussion points far beyond what my newbie hockey eye test can tell me. So, to the comments section! (I'll try my best to update the chart if the lines need to be rearranged)

Free Food... courtesy of the the City of Vancouver
I know that some of you show up to these geekfests just for the food, so lucky for you, Mayor Daley made that bet with Mayor Gregor Robinson of Vancouver at the start of the series.
Correction: What I have here is last year's bet... I put up a new FanPost (Another Mayoral Hockey Bet!) for this year's (and the one with San Jose's mayor), but you guys can chow down on this stuff as your reward for making it through this post.
A similar bet was made last year, and as promised, payment was made:
What was included:
- beer (Molson's, Granville Island Brewery and Storm Brewing)
- smoked salmon (Musqueam First Nations)
- dried morel mushrooms (Pacific Rim Mushrooms)
- chocolates (Bad Girl and Purdy’s)
- hand-cut organic tortilla chips (Que Pasa)
- vegetarian blueberry sauce (Thai Princess)
- fruit juice (Happy Planet)
Since our Vancouver fan friends so largely exceeded our expectations for classiness, I thought I'd include links in case, as their mayor said, we end up giving them "a bit of a beachhead there for local products." Not to mention with how healthy this stuff is, it might be just what our stressed hearts, stomachs, and livers need right now:
(Quick Plug: If you're in the New York metro area on Sunday, I'm trying to get a meetup going... check it out)
10 recs |
43 comments
|
Comments
SO THIS IS WHY YOU NEVER ANSWER MY E-MAILS
I kid. This is great! What’s most comforting going into the WCF is how well certain guys have played against the Sharks this season—Toews, Brouwer, Hossa and Ladd have all been huge for us, and even going back a couple more seasons, Toews in particular has always been sick-nasty when he faces the Sharks. (Scoring his first NHL goal on his first ever shot against Nabby might just have been an early premonition of this, wow.)
It also looks like the Meatball Soup pairing does well against the shutdown pair of the Sharks (Blake and Murray), especially when they played with the (formerly) top line of Light, Magic and Industrial (19-88-22). My head is swimming right now because I had to work 13 hours to day, SO MUCH BULLSHIT, but these numbers make me super excited. It should be an incredibly dynamic series no matter what happens.
Light, Magic and Industrial
Holy shit I love that.
by HungryHungryPanda on May 14, 2010 8:31 AM CDT up reply actions
my nerd-radar
tells me it’s a play on Industrial Light & Magic (ILM), the special effects shop? but I feel like I’m missing the real reference here…
I haven't slept since posting my first comment, and it's still the only one? Sad face.
I’m now obsessed with doing a full set of WoWYs in the offseason. I guess I’ll have to crack open that stats textbook my bff gave me, after all. (sigh)
karma fail
for me not answering your e-mails… never mind job transition / apartment hunting / not sending enough self e-mails (what I use that account for, largely) and general laziness, I obviously had enough time to do this that writing back should’ve been easy…
I think I learned my lesson. The hockey-blog gods have spoken! I repent! I repent!
Good job Verstig
I read it last night night but didn’t comment since I was reading BtN before diving into your cross-mutual +/- (and fell asleep).
I wish there was a way to increase the sample size on that c-m+/- data because I would believe that not only is it skewed by the fact we won the series 3-1 (although I believe 2 were in OT), but much more so by that 7-2 curb-stomping we gave them upon Hossa’s return. I know you mentioned to take it with a grain of salt, but I’m curious if you looked at the data with that game pulled out before making your proceeding observations to see if they still held somewhat true? Just curious.
Well, folks, I want to thank you for being here for the recording of my live comedy album. Funny material and laughter will be dubbed in later.
by ChicagoNativeSon on May 14, 2010 9:20 AM CDT up reply actions
Great point
You’re absolutely right; that game does make a huge difference. I redid it with just the 3 other games…
In comparison it looks like Kaner and Sharpie got manhandled, and Marlboro 72 weren’t even that good. Looks like Meatball Soup (gmh’s term?) may have to carry a lot of weight this series. Will have to look out for Mr. Pickles.
Of course, all I did here was take a tiny sample size and make it tinier… so, you might want to take a salt lick with you as you look at this.
Hehe true, we may be going in the wrong direction here!
What is very nice to see from that chart and the one above is that Campbell apparently performs well against the Sharks. I really hopes he gives SJ plenty of reasons to boo him. Thanks Stig!
Well, folks, I want to thank you for being here for the recording of my live comedy album. Funny material and laughter will be dubbed in later.
by ChicagoNativeSon on May 15, 2010 10:08 AM CDT up reply actions
re: sample size
Yeah, there’s really no way, short of moving to shots or Corsi instead of goals, to increase the resolution here. I can’t use TMFF’s bootstrapping technique here because trying to use a tiny sample size to simulate a larger data set would just give you more of the same, and would pretty much make it meaningless. You’d have to make too many assumptions.
The only other way would be to go the structural route, and map out all the other stats that came into play, and predict how the game would physically play out. But that would probably be rendered useless because the lines would be different.
That is why I meged data
Used regular and post season for my WCF simulations, weighing them based on percentage of total games, and weighing home ice 5:3 to try to give that a little more edge, since both team were stellar during the regular season on home ice.
If the Hawks do the deed, I will be switching again, but I will deal with that if needed, if the time comes.
Get off my Land!
ART.I§8-11; AM I-XXVII
James Madison is my Hero!
by Toews-makes-funny-faces on May 16, 2010 6:51 AM CDT up reply actions
What I'm saying is
That works fine for what you’re doing, which is giving probability distributions for goal scoring and games won, but for something as fine grained as +/-, all I’d be doing is weighing the data I have (3, maybe 4 games) by some other factors (say, home vs road), which doesn’t really change the descriptiveness of my stats.
Sorry
I was agreeing with you, I just did not make that clear though. The small playoff sample size is why I went to merged data, to use the sample (playoffs), but not let it be the only sample. 11-12 games against only 2 teams is not enough to work with for a 4-7 game series projection.
Get off my Land!
ART.I§8-11; AM I-XXVII
James Madison is my Hero!
by Toews-makes-funny-faces on May 16, 2010 10:21 AM CDT up reply actions
Yeah
Using regular season stats to help weight your playoff ones is a reasonable thing to do… but seeing as all I have is regular season stats to show regular season stats, there’s nothing else left, haha.
That’s why I said you need something more structurally predictive (say, “transition probabilities” between certain plays happening) to simulate games if you want to do something greater with the sample size I was working with (4, or even 3, games).
Not sure exactly how much impact it might have...
Would there be any validity in doing it up for both teams vs Canucks, and both teams vs Scum and then doing a comparitive match-up that way?
It would increase the sample size, against (lets face it) playoff caliber, common foes. Sharks played Nucks throughout the regular season, and Scum in Playoffs. Hawks played Scum throughout the regular season and Nucks in Playoffs.
How badly would something like that a) skew the numbers, and b) be a legitimate basis for comparison?
<<< FAR from a statistician (although I love looking at them and comparing them).
Everyone dies. It is the only true and lasting justice in life.
I really like that idea
To try to compare two data sets against a common one would definitely be a legitimate thing to do, though it’d introduce its own errors.
The only problem is, what I did here was compare matchups between Hawks and Sharks players… I can’t think of a way you could use a third team as a proxy (like, say Toews is a +1 against Kesler, and Thornton is a 3 against Kesler - what would the +/- for Toews against Thornton be?).
I’ll keep that in mind (and all you young statisticians out there too, I suppose — thanks for the insight!) but we might have to get even more creative here.
Okay, so... We need to know the question that we're trying to answer.
Yeah, its tough to find the bottom line in this with the proxy team 3/4…
Would using CORSI or FENWICK against common foes over a larger sample be valid?
You could compare each players ratings under those two systems against the common foes (Scum and Nucks) to create a baseline. i.e. Based upon their games played against common foes, the performance baseline for Jonathan Toews vs Joe Thornton is a +3 (Toews’ ratings directly relative to Thornton’s).
That’s a possiblity, but I don’t know if it provides the end result comparison you were originally trying to reach.
Everyone dies. It is the only true and lasting justice in life.
I suppose it doesn't at all address in any way, a better formula for
putting lines on the ice. It isn’t a narrow enough focus that way.
Perhaps Thornton vs each player compared to Toews vs each player, with the Scum and Nucks as a larger test group.
(Using imaginary numbers here, completely pulled out of my ass, to visualize the idea)
Dammit.. I was trying to abbreviate, but they’re both #19 with initials JT… Okay, anyway:
Dammit.. I was trying to abbreviate, but they’re both #19 with initials JT… Okay, anyway:Toews Vs Datsyuk over the season (to include Playoffs) was a +2
Thornton Vs Datsyuk over the season (to include Playoffs) was a +4
Toews Vs Burrows over the season (to include Playoffs) was a +5
Thornton Vs Burrows over the season (to include Playoffs) was a +1
Dammit.. I was trying to abbreviate, but they’re both #19 with initials JT… Okay, anyway:Toews Vs Datsyuk over the season (to include Playoffs) was a +2
Thornton Vs Datsyuk over the season (to include Playoffs) was a +4
Toews Vs Burrows over the season (to include Playoffs) was a +5
Thornton Vs Burrows over the season (to include Playoffs) was a +1This gives Toews an advantage adjustment Vs common foes of +2
Dammit.. I was trying to abbreviate, but they’re both #19 with initials JT… Okay, anyway:Toews Vs Datsyuk over the season (to include Playoffs) was a +2
Thornton Vs Datsyuk over the season (to include Playoffs) was a +4
Toews Vs Burrows over the season (to include Playoffs) was a +5
Thornton Vs Burrows over the season (to include Playoffs) was a +1This gives Toews an advantage adjustment Vs common foes of +2Head to Head, Toews is a +3 on Thornton In their own games, throw in the adjustment, and Toews is a +5 Vs Thornton?
Dammit.. I was trying to abbreviate, but they’re both #19 with initials JT… Okay, anyway:Toews Vs Datsyuk over the season (to include Playoffs) was a +2
Thornton Vs Datsyuk over the season (to include Playoffs) was a +4
Toews Vs Burrows over the season (to include Playoffs) was a +5
Thornton Vs Burrows over the season (to include Playoffs) was a +1This gives Toews an advantage adjustment Vs common foes of +2Head to Head, Toews is a +3 on Thornton In their own games, throw in the adjustment, and Toews is a +5 Vs Thornton?Any validity at all to it, or is the net cast too wide for a narrow focused question?
Everyone dies. It is the only true and lasting justice in life.
Sigh... . Why does it do that?
It should read like this:
Okay, anyway:Toews Vs Datsyuk over the season (to include Playoffs) was a +2
Thornton Vs Datsyuk over the season (to include Playoffs) was a +4
Toews Vs Burrows over the season (to include Playoffs) was a +5
Thornton Vs Burrows over the season (to include Playoffs) was a +1
This gives Toews an advantage adjustment Vs common foes of +2
Head to Head, Toews is a +3 on Thornton In their own games, throw in the adjustment, and Toews is a +5 Vs Thornton?
Any validity at all to it, or is the net cast too wide for a narrow focused question?
Everyone dies. It is the only true and lasting justice in life.
Phew, I'm glad I scrolled down
To answer your first comment, yes, Corsi and/or Fenwick would be completely valid. I think I mentioned that I wish I could use it, but that I was too lazy to write a play-by-play parser, probably not until the offseason. It’s probably not that much work since the format isn’t all that different from the game summary, but I EM LAZEY.
As for using opposition as a proxy — yes, there’s some validity in this, but it’s really a different question. Different people will play opponents differently, which in turn could be very different from how they play each other — to say nothing about linemates and other factors, like goalies. So, you could do this, but you can’t really compare it to what I’m doing, at least not directly.
That said, Hawerchuk at BtN does a more generalized case of what you’re saying (not comparing the opponents directly, but just measuring up against opposition/competition itself). So he actually assigns a couple of stats: QUALCOMP (quality of competition) and RATING (BehindTheNet Rating, which is +/- adjusted for QUALCOMP and QUALTEAM).
Got it. I guess that's already done then.
Next idea??
LoL
(Great stuff to begin with, by the way – since I don’t believe I mentioned that anywhere – it’s exciting stuff)
Everyone dies. It is the only true and lasting justice in life.
Real good work, Verstig
I’m sort of wondering if, for this series, Q might try moving Bolland up to 2nd line center again with his improved play in the VAN series. Then Madden can move back into his checking line center role, where he absolutely made Jumbo Joe invisible this season.
If Bolland can handle it, and there’s been no consistent indications that he can, I would really like the layout of our lines a whole lot more. I have a feeling 4th liners aren’t going to see a lot of ice time in this series on either side, but we need Madden out there on Jumbo.
by HungryHungryPanda on May 14, 2010 8:35 AM CDT reply actions
Tough call
It basically seems to come down to “does Q focus on offense or defense?” If you move Bolland up, you potentially hurt the offense.
I agree the 4th line won’t be seeing much time, as is typical in the playoffs. IMO this somewhat hurts the Hawks because our 4th line is typically stronger than our opponent’s. We commented all year about the Hawks having such good balance, but that’ll only get you far in the regular season. In the playoffs, it doesn’t hold nearly as much weight.
Well, folks, I want to thank you for being here for the recording of my live comedy album. Funny material and laughter will be dubbed in later.
by ChicagoNativeSon on May 14, 2010 9:30 AM CDT up reply actions
I can see some defensive zone draws
where Q puts both Bolland and Madden out there against Thornton. He’s been doing that a lot this season, and it may be a possibility if Thornton proves to be more of a handful against us than he was in the reg. season.
I wish they kept track of when somebody gets thrown out
I’d say you could tell if two centers on the ice and the lesser one takes the draw, but I’m pretty sure I’ve seen a few times where the lesser FO center tries to take it first and gets thrown out (maybe they can “cheat” a little knowing somebody will cover for them?).
I wish I could understand the rules of F/Os to begin with
I have trouble telling when someone wins (unless it’s obvious/clean win) or why people get thrown out.
centers get tossed for failing to line up right, jumping the gun or teammates jumping the gun - mostly
win/loss can be hard to decide (when I tracked players – I had the stat watchers track win/lose/draw)
My centers were coached with alternative “go for the win” or “don’t lose it clean” strategies – and we had some signs from the bench as to which we wanted in any given draw.
Confusion will be my epitaph.
Its kinda surpring the MLB track EVERYTHING!
I mean EVERYTHING. Plays to the left, to the right, hits off what type of pitch and pitch location etc, etc, etc……
I’m kinda surprised the NHL does not track more with tracking these things. But then again baseball is more of a stat sport the then Hockey, since baseball is one on one battles, and individual efforts that can be broken down, where Hockey is certainly more team orientated and takes on a whole different set of dynamics. Rarely is baseball a truly multiple player effort where one players ability handcuffs or heightens anothers players ability of the same team, like in Hockey.
Get off my Land!
ART.I§8-11; AM I-XXVII
James Madison is my Hero!
by Toews-makes-funny-faces on May 16, 2010 6:56 AM CDT up reply actions
Thank god you do this VerStig
I dont even have to bother with stat analysis anymore thanks to you and gmh…yeah, Im lazy
Hossa's Attorney says:
TAKE OFF THOSE CLOWN SHOES OR I WILL SHOOT THIS DOG!
HUGE thanks
for these stats breakdowns. The breakdown against the Sharks looks sort of promising but I’m not letting it get my hopes up too much.
it's entirely possible this isn't sarcasm
one error though
Sorry to piss in your cherrios:
The reference to the Vancouver mayor paying up is from 2009.
The quotes in the story that read “Into the bag also went some 2010 Olympic gear and a couple of umbrellas. "Chicago is bidding on the 2016 games and we wish them well," said the mayor.” are a dead giveaway. That and the publish date of May 14, 2009.
angry bird is angry .
Wait, they did this last year too?! Lol!
Good catch though, thanks! Here’s to proofreading, ha.
Pity…that stuff looks yummy. I’ll pull this year’s list from that Puck Daddy article later in the afternoon when things lighten up (and I can go from distracting myself every 5 minutes to distracting myself every 30 seconds), and hopefully somebody writes something about Mr. Robinson paying up so I can add that here too. Looks like it took them 4 days last year so we’re just about due.
You make me and my nubers feel like such a tool!
I need to get hold of your number crunchers to the stuff I do. Excel is nice, but the way I use it, locks up allot or takes forever when I want to change even a few calculations and copy them
Awesome work here!
Get off my Land!
ART.I§8-11; AM I-XXVII
James Madison is my Hero!
by Toews-makes-funny-faces on May 14, 2010 1:56 PM CDT reply actions
sorry to be late to the party
but i’m a HUGE fan of this. And i think it actually gives more concrete ansers to “hmmm, what should we do against the Sharks?” than many of the other stats being crunched…..even though it’s all speculative anyway b/c who the hell knows what’ll happen??
Well done.
Just added the Cross-Mutual TOI for Hawks-Sharks
that I posted in a different thread…
I’m glad you guys (or, just you?) like these, because I think they still look a little too overwhelming to just look at casually. If Q didn’t change his lines so darn much I could just block it out into how each line played (though the purpose of the 1-team mutual stat is to see how to make good lines).
Seriously though, I think creativity is important to statistical analysis (which is why I really appreciate what gmh and TMFF have done)… sometimes it’s not so much how much number crunching you do but rather trying to find a different way to look at the numbers in a useful way.
Question is, what’s next? My super-long stats post was basically me getting disillusioned by the idea of using traditional stats.
Minor update
Just noticed I had the defensive pairings jumbled, so I fixed them. Also, the highlighting has been “enhanced” for clarity.
Wow
This is great stuff! Nice job…and thanks for doing this. In addition to the data and the analysis, I like all the little extras you put in there too.
Question for you VerStig
Been thinking about this and was wondering if your programs can handle it.
I was trying to figure out the matchups that were going on between a team that was playing three lines and one that was playing four. What I wanted to see was the flow of the time on ice. Is it possible to show gant style charts from the Time on Ice stats?
Instead of the totals that Time On Ice shows I was thinking about a graph with the players from both teams on the top and Times below. That way you could have a gant chart for each player when they were on the ice and see at any time who was on and who wasn’t.
Does that make sense and is it possible?
You mean this?
I think Sam (and the Matts?) used to link to this, but for some reason they’ve been linking to the Head-to-Head TOI instead, as of late.
You mean this?
Thank you, I didn’t know Vic did one of those…

by 





























