17 Responses to Scoring chances for/against: Jan 7th: Van vs Edm

1. 

   David Staples

   January 8, 2009 at

   Yeah, that’s the way I saw it, too, with the Viz and Grebs having their worst game in some time, not only in quantity of scoring chances against, but quality.

   And I also saw Staios, Cole and Horc having good games out there, along with Souray. That’s how I graded it . . .

   I like the new chart Dennis. Easy on the eye.

   And I’m looking forward to the 40 game or 41 games summary.

   What would you say are the chief things you’ve learned about the Oilers (and perhaps about hockey itself) by doing this exercise?

   And, finally, if Garon and Roloson have widely different scoring chances for and against, how do you account for that? I’ve seen this with Corsi numbers and other stats, how the goalies on the same team have widely different plus/minus ratings. … Most likely it’s the teams that they play against, but it’s still struck me as interesting and odd.

   Reply
2. 

   corykg

   January 8, 2009 at

   Further reason to be baffled about Moreau being out there so much in the second half of the third.

   Reply
3. 

   Dennis

   January 8, 2009 at

   Staples: now that Vic had made things so easy to compile, perhaps I can get going with categorizing where these shots or shot attempts are coming from.

   Perhaps I can do one at a time and just highlight slot opp as an example. Which brings me to what I’ve learned about this club.

   1: The D isn’t good at keeping the crease clean
   2: The slot is often open for the opposition.

   What I’ve learned in general – and was something I thought I knew anyway – is that the closer to the middle of the ice you are, the better your chance to score.

   Reply
4. 

   Rick

   January 8, 2009 at

   I was watching to see who was being sent out for the PK faceoffs, and as far as I can remember MacTavish was sending out Horcoff + Brodziak for every own zone draw. Vic’s shift charts confirm this for the start of every Vancouvers PP. Are faceoff times recorded anywhere so I can check who was sent out for faceoffs in the middle of penalties?

   Going from memory, I thought MacTavish was looking switch this up if the Oilers were able to clear, but it turns out that 51 just got stuck out for nearly the entire PP at 38:00.

   I randomly picked vs DET on Nov 20; at the start of PPs MacTavish sent out 10/27, 10/51, 10/85, 10/18, and 10/18.

   I think we can conclude MacTavish trusts 10 to take own zone draws: 100% to start the PK is pretty compelling even only over two games. It also looks like he’s gone back to his comfort zone of sending a righty out with his lefty while he hadn’t been doing that earlier in the year.

   The Detroit game illustrates why 18′s been working on draws, as it looks like MacT would like to run 10-18 as a PK tandem. I also think the episode last night illustrates why 18 shouldn’t be let anywhere the faceoff dot.

   Another thing that stuck out is that 71 got ~0 minutes of PK responsibility over both games. MacTavish has shown a lot more trust in 43 as a PKer. I still don’t think the PK issues are personnel based, but trading away two PKers for non-entity in 71 probably isn’t a positive factor.

   Reply
5. 

   Dennis

   January 8, 2009 at

   Rick: No doubt, the Oil moved a couple of guys who could kill penalties and they never really replaced them. Plus – and this is without looking at what the 16-19 combo did in terms of PK faceoff pct-wise – they did it in a summer when the NHL changed the rules about where all the PP’s would start.

   Reply
6. 

   Rick

   January 8, 2009 at

   I’m beginning to doubt that the Visnovsky trade was the clear win everyone was assuming it to be. Sure the guy is a horse at EV, but before looking into it more in depth there are two options:
   1. He sucks at killing penalties.
   2. He said something nasty about 18′s brother.

   Meanwhile we gave up a bonafide PK/ faceoff stud in Stoll and a D in Greene that may fill this teams needs better than 71.

   I mean 37 is basically a younger version of 71 that MacTavish trusts to kill penalties. They have similar Corsi, 71 has the better GA/60, 37 with the better GF and points/60.

   Plus Greene stylistically fits what the Oilers lack on the back end this year – tough, stay at home, PKs, right handed. I’d like to look into the anecdotal evidence of him taking ‘the next step(tm)’, but one may be able to make the case the Oilers would be better off this year had they not made the 71 trade.

   Reply
7. 

   Rick

   January 8, 2009 at

   I checked a small sample of players PK Corsi on behindthenet:

   Stoll -74, Green -81, Reasoner -96.

   Which looks pretty similar to:
   Brodziak -72, Gilbert -87, Horcoff -99.
   or
   Cole -75, Grebeshkov -93 Cogliano -92

   Given that these are the exact PKers that the Oilers gave up, and they match guys they played with last year and are being replaced by this year, I’m not buying that it’s the players fault.

   Incidentally 71 has the best PK Corsi on the Oilers and plays less than 1 min on the PK. File this under more fodder for the fire MacT crowd.

   Reply
8. 

   Bruce

   January 8, 2009 at

   Dennis: Loving the new look. I really like the addition of TOI, puts things in context on a per-minute or per-60 basis and allows one to identify the high-event guys in a hurry. One other idea for your and Vic’s consideration: in the player summaries show GF/GA in adjacent columns to SCF/SCA in each situation. Useful info, esp. to somebody who is tabulating the game to game totals. (Has Scott done an update recently? If so I’ve missed it.)

   While I’m in suggestion (NOT: request) mode, TOI is the one column that I really would love to see in Vic’s otherwise fantastic shots data at the ironically-named TimeonIce.com . I use that data all the time and always wind up cross-referencing TOI from elsewhere. Just a suggestion for your consideration, Vic, if you’re reading.

   And, finally, if Garon and Roloson have widely different scoring chances for and against, how do you account for that? I’ve seen this with Corsi numbers and other stats, how the goalies on the same team have widely different plus/minus ratings. … Most likely it’s the teams that they play against, but it’s still struck me as interesting and odd.
   

   You’re right David, it is both of those things. I don’t know about scoring chances but the shots for/against numbers are very different between the two. Garon +23, Roloson -107. Particularly interesting given their respective W-L records; more grist for the outshooting mill where I note the Oilers as a team are once again posting a superior points percentage when outshot (.537) than when outshooting (.500).

   This is one area where individual TOI would be particularly valuable, since NHL.com does not seem to keep such data for goalies by game situation (despite ample room on their “special teams” goalie pages). As a crude guess I have taken total minutes played and assumed that 75% of that is played at ES for both goalies. Accepting that assumption, here’s the breakdown on a per-60 basis:
   
   Roli / Garon
   ———–
   ~979 MP ~536
   25.6 SF 31.1
   32.1 SA 28.5
   ————-
   1.88 GF 1.68
   1.47 GA 1.76
   ————-
   .918 Sv% .939
   7.2% Sh% 9.8%
   .990 PDO 1.037
   ————–
   
   The bottom set of numbers are straight from Timeonice with PDO numbers (which I’ve been following for years under a different name) simply being the sum of the two percentages. Note NHL.com has Roli’s ES Sv% at .941 and Garon’s at .922, very close to ToI’s numbers. The difference can possibly be explained by NHL.com including 4v4 play.

   It’s a funny set of numbers. The team takes fewer shots in front of Roloson, but is far, far more efficient with them. The team allows quite a few more shots against Roloson, but judging from his superior GAA he stops more than all of them. How much of that is Roli/Garon, how much is team play in front of them, and how much is game score effects are questions I am capable of asking but not answering at this moment.

   On this latter subject, the Contrarian Goaltender currently has a very interesting post on Playing to the Score that is well worth the read.

   Reply
9. 

   Bruce

   January 8, 2009 at

   Ah fuck, look at that, I screwed up again. Every time I post stats on one of these no-preview sites I seem to mess something up. In this case, the bottom set of stats for the two goalies is reversed.

   This should make more sense:
   
   Roli / Garon
   —————-
   ~979 MP ~536
   25.6 SF 31.1
   32.1 SA 28.5
   —————-
   1.88 GF 1.68
   1.47 GA 1.76
   —————-
   .939 Sv% .918
   9.8% Sh% 7.2%
   1.037 PDO .990
   —————-

   Sorry about that.

   I’m loving your new site banner, MC.

   Reply
10. 

    Tyler

    January 8, 2009 at

    Did you see that Bruce? The room I was in erupted as CBC gave us the timely closeup.

    Reply
11. 

    Bruce

    January 8, 2009 at

    Yeah, somebody mentioned it on LT’s GDT, so when I watched the game on my PVR later I made a point of watching for it. What the baseball types call a “no doubter”. That’s why I paid the big bucks for HDTV, right there.

    Reply
12. 

    Doogie2K

    January 9, 2009 at

    No doubt, the Oil moved a couple of guys who could kill penalties and they never really replaced them. Plus – and this is without looking at what the 16-19 combo did in terms of PK faceoff pct-wise – they did it in a summer when the NHL changed the rules about where all the PP’s would start.

    Ask and ye shall receive. Turns out the Oilers’ faceoff prowess, at least on the PK, may have been somewhat overrated. I didn’t go through and compare their PK% to their EV% or Total%, though it might be an interesting project for another day, but it is odd that of all our centres or faceoff “specialists,” only Mike Peca could string together multiple 50%+ seasons of any real sample size. Note also that in two years in Anaheim, Dustin Penner didn’t take a single SHFO.

    Legend: Blue represents split seasons (I couldn’t tease out the numbers on NHL.com easily; they only represent around 60 GP, so it’s probably not too bad), yellow seasons with other teams, black seasons outside the NHL. I tried to include every centre or role player reasonably expected to take a faceoff, but left guys like Sykora and Thoresen off because (a) their numbers were small, and (b) frankly, they weren’t on the team long enough for it to be worth my while. I wasn’t even going to include Nedved until I saw he took 173 PKFO in two years, even if maybe a quarter of them were in Edmonton.

    Reply
13. 

    Doogie2K

    January 9, 2009 at

    Of course, it still bears pointing out that faceoff numbers are down after three straight years of small increases, right when they became more important. So we still have a bit of a problem here, it’s just been amplified by the rule change.

    Reply
14. 

    Doogie2K

    January 9, 2009 at

    Sorry to triple-comment here, but eyeballing the numbers going back to 2003-04, the Oates year, it seems that everyone’s PKFO% is lower than their total FO%, which makes sense when you go back and think about it: they haven’t gotten any better or worse on the dot, in terms of clean wins, but without a second winger to help tie up or win the loose puck, you’re going to lose a ton of scrambled draws on the PK, and similarly, win a ton of scrambled draws on the PP.

    Reply
15. 

    Bruce

    January 9, 2009 at

    Good stuff, Doogie. re: your last comment, that fits exactly with an observation I have been making for some time but never had the stats to back it up. I estimated 55/45 for the PP team, and that appears to be in the ballpark.

    A related question which maybe you can do quickly now that you got the structure in place (or which is the same amount of work again, in which case don’t feel obliged): how have our draws been going on the PP? I am frustrated in many games when it seems like both special teams are getting their asses kicked in the circle, but perhaps over the fullness of time our PP is 5% above the team mean and the PK 5% below and the main thing we need to work on is just winning the damn faceoff in the first place.

    Reply
16. 

    Doogie2K

    January 9, 2009 at

    I decided to approach it a bit differently, and just looked at the team as a whole, season by season, because accounting for who took a lot of faceoffs in what situation started to make my head hurt and my spreadsheet all messy. Without accounting for players who changed teams partway through the season due to trades/waivers (Reasoner in ’06, Smyth and Nedved in ’07, Thoresen in ’08), since I don’t have a script to scrape faceoff data like that, here’s the rough picture through three and a half years. The EVFO% has been on a steady decline year over year, but look at those ST numbers: they stepped into an elevator shaft this year, and based on my perusal of the data, it does look like losing 16-19 is a big part of it. Prior to that, though, it looks like the Oilers won about 57-58% on PP and 47% on the PK; if there’s a PDO number for STFO%, the Oilers had a massive one 2005-08, and a pathetic one (92.9%) this year. Make of that what you will.

    I uploaded the spreadsheet here, if you want to take a look. I’ve never used Google Docs, so I don’t know how, exactly, it works, but there you have it.

    Reply
17. 

    Scott

    January 20, 2009 at

    Totals (Player,GP,EV,PP,SH)

    05 – 23 – 073/085 – 000/000 – 002/010
    10 – 36 – 167/140 – 093/008 – 007/079
    12 – 31 – 116/115 – 034/010 – 000/000
    13 – 36 – 126/123 – 035/012 – 002/019
    18 – 35 – 111/165 – 002/000 – 007/061
    24 – 34 – 108/138 – 002/000 – 004/095
    26 – 36 – 132/127 – 043/007 – 007/016
    27 – 34 – 129/119 – 064/005 – 002/028
    28 – 02 – 002/002 – 000/000 – 000/000
    33 – 10 – 005/006 – 000/000 – 000/000
    34 – 16 – 048/062 – 000/001 – 001/023
    37 – 33 – 163/150 – 033/008 – 006/032
    43 – 31 – 065/090 – 000/000 – 004/037
    44 – 35 – 166/170 – 104/011 – 008/077
    46 – 24 – 030/038 – 000/000 – 000/000
    49 – 01 – 004/000 – 000/000 – 000/001
    51 – 34 – 074/095 – 001/000 – 008/061
    62 – 01 – 001/000 – 000/000 – 000/003
    67 – 05 – 008/016 – 001/000 – 000/000
    71 – 36 – 202/172 – 101/014 – 003/012
    77 – 36 – 171/180 – 044/009 – 006/064
    78 – 33 – 096/090 – 017/000 – 000/003
    83 – 31 – 132/115 – 082/008 – 000/002
    85 – 16 – 028/041 – 000/000 – 000/013
    88 – 04 – 010/015 – 002/002 – 000/000
    89 – 33 – 132/115 – 045/010 – 002/006

    Reply