Tuesday, August 22, 2006

Smyth and Horcoff at 5V5? WTF!?!?!

Ryan Smyth and Shawn Horcoff took a beating at 5V5 last year. I have no idea what happened to them because honestly they kind of sucked at keeping the puck out of their net. When you compare the numbers to 03/04 it gets kind of scary.

Ryan Smyth: EV+48/EV-38 in 03/04. EV+52/EV-56 in 05/06.

Shawn Horcoff: EV+45/EV-45 in 03/04. EV+56/EV-58 in 05/06.

The hourly rates work out to:
Smyth: +2.5/-2.0 in 03/04 and +3.3/-3.5 last year.
Horcoff: +2.5/-2.5 in 03/04 and +3.2/-3.3 last year.

As a benchmark, the team was EV+148/EV-121 in 03/04 and they plummetted to EV+133/EV-148 last year.

Ouch.

So let's play ask and answer...

Q: What didn't change?

1. Injuries. They both had some nagging injuries and MacT played them anyway. Damn the torpedoes. Some things never change.

2. Opposition. They played tough minutes both years according to Vic's event based stuff.

3. Paired up. They played together a lot. Maybe a little more in 05/06.


Q: What did change?

1. The rules. Yeah the rules changed. I haven't the foggiest idea how that may have affected these two players. I'm taking suggestions.

2. The goaltending. The Oilers' goaltending was pretty sucktastic in 03/04. They set a new low in the 05/06 regular season. The Oilers' league relative SV% in 05/06 was 984 and in 03/04 it was 995. By my rough calculations that's about a 13 goal difference in all situations. The Oilers gave up about 63% of their goals against at 5V5 last year so we'll say goaltending was worth about 8 goals at ES last year. According to Tyler's team numbers and the individual NHL.com numbers, Smyth played about 27.3% of the Oilers' minutes at ES. That means Smyth's share of the goalie hurt at ES was about 2 extra goals against at 5V5 vs. 03/04.

(Note: this is rough - if Tyler was a lifetime student he'd have updated his damn site and I could've used his situational numbers.)

From 03/04 to this past season, it appears to me that Smyth gained +4 GF and added -18 GA. Horcoff gained +11 GF and -13 GA. Do you still think goaltending explains this slide? I don't. Even if my GA due to goaltending estimate is wrong and you triple the number to six goals, they still had some serious issues with GA at 5V5.

3. Hemsky. Smyth and Horcoff didn't play much with Hemsky in 03/04 and they played a heap with him this time. To my knowledge, Hemsky hadn't seen opposition like those two were used to at 5V5 so he had a few more things change going from 03/04 to 05/06. He's also a high risk/high reward player. That's a stark contrast to Smyth and Horcoff's usual 03/04 RWers which were usually Pisani or Dvorak IIRC.

I think it's Hemsky's fault and I think it's because Smyth's strength as an outscorer is gaining possession in the offensive zone and keeping it there. Same really for Horcoff, although he's probably a little better in his own end. Hemsky's turnovers at the blueline hurt the line's ability to keep the other team pinned in their own zone and thus they saw their outscoring drop. Maybe it was also a little contagious as well and they played a riskier game in general. Admittedly I'm reaching there.

Funny thing though:

Hemsky: EV+27/EV-31 in 03/04. EV+40/EV-45 in 05/06.

Crap. WTF does that mean?!?! Sure Hemsky played fewer ES minutes in both seasons but the rates (+2.0/-2.3 in 03/04 and +2.6/-2.9 last year) don't really help much. Hemsky's the worst of the three in GD/hr of course but none of this really makes much sense to me. This was mentioned before on the boards but somehow Hemsky's the lower event player? Does that really jive with what you see on the ice?

Anyone out there have a unifying theory?

The bottom line is that the Oilers saw their 5V5 games take a beating last year and Smyth and Horcoff are two of the team's best forwards. They need to do better next year and play to their strengths in order for this team to have success.

39 Comments:

Anonymous Bank Shot said...

Hemsky played with Moreau, Peca, Stoll, and Torres.

Smyth and Horcoff didn't.

Hemsky's 4 on 4 partner was also Peca.

Different linemates might explain the discrepancy in GF/GA.

Reasons for why Smyth and Horcoff might have been less effective 5 on 5 besides Hemsky.

Smyth made very extensive use of the absence of the two-line pass rule. Sitting out around the redline all the time might have been a double-edged sword.

Maybe Horcoff had to relearn how to play defence. He was Dr.Hook for a large part of the season. Leading the team in minors during the regular season. In the playoffs where he was a 5 on 5 crazed maniac, he had one of the lowest penalty totals on the team.
Maybe it took him the better part of the season to learn how to shutdown other players without mauling them.

8/23/2006 5:20 am  
Anonymous lowetide said...

Sometimes shit happens. It just does. I'm not able to safely say how much of this is just randomness and how much is actual measurable falloff but part of the story is there too.

I don't know. Odds are that every team that sends a gifted, high event young winger out against the opposition's best is going to get their hat handed to them time and again until the kid learns how to played a more disciplined game (pick his spots).

Is Hemsky so gifted that a coach who is devoted to logic and matchups like MacTavish will simply say "screw it" and send him over the boards with his two best forwards no matter the matchup?

It would appear so.

8/23/2006 7:07 am  
Blogger Black Dog Hates Skunks said...

What are thoughts on playing someone like Sykora or Pisani with Smyth and Horcoff? I agree with the whole offensive zone possession angle - in my mind this line was constantly creating chances on the rush rather then working it deep, Smyth's obvious strength. Or have Sykora play centre with Hemsky and ??

Or keep Horcoff and Hemsky together and move Smyth, replacing him with Lupul maybe?

Although Lupul and Hemsky together - yeesh.

Never mind.

8/23/2006 9:05 am  
Blogger Vic Ferrari said...

I agree that it is a really important question. Usually the fate of your goaltender and your best players determines the fate of your team. In the case of the Oilers their best 5on5 guys (94 & 10) weren't the same as their best powerplay guys (83 & 44) so they were at least a bit diversified. Still, it hurt.

I think the playoffs are really telling. There was a turn of tactics and Hemsky was mostly paired with Stoll and they were kept away from quality opposition at almost any cost in the postseason. The game in the S.J series (Game 3) where he was with Smyth & Horcoff was the one game that those latter two didn't have the Thornton responsibility (it fell to Peca). And in the Stanley Cup finals when MacTavish reunited them (a crucial error IMO) he was loathe to have them out there against the Brind'Amour/Stillman line, and preferred to avoid the Staal matchup as well.

It's not unusual for young guys with promise to be thrown into deeper water to help them progress. Look at Hamhuis, your whipping boy. I know that there was a monster thread on the subject at HF, it was referred to in Lowetide's blog ... but at 13 pages I'm sure it would have lowered my intelligence and just generally pissed me off if I read the thing, so I didn't :) Apologies if I'm covering old ground.

Just for the helluvit I ran a quick script to find the games that Hamhuis played with Timonen, and there was a stretch of 3 or 4 weeks in the middle of the season where that seemed to be his main D partner. During that time Hamhuis played tough opposition. Now maybe there were injuries or Trotz really thought that Hamhuis was his best option for that role, but more likely Trotz just recognized that they weren't going to catch Detroit anyways, and that 4th spot overall was pretty much a lock, may as well put on his player development hat. The problem with dewveloping players at the NHL level is that you lose a lot of games, and it looks like it wasn't long before 4th spot overall was no longer a certainty, and he reverted back to playing the vets where they should be played.

I haven't checked, but I'd bet that Nick Schultz played much tougher minutes AFTER it became apparent that Minnesota were not going to make the playoffs. And so on and so on.

So for teams with really nothing to lose (either terrible and out of the race or a playoff spot all but locked up) ... it's commonplace. Much rarer for a team that's hanging around the eighth playoff spot all year.

Or maybe it just made it easier for MacTavish to get the Horcoff/Smyth natchup against top players on the road (put Pisani or Dvorak with those two and Wilson aboids them like the plague with Thornton ... put Hemmer with those two and he's gunning for them with big Joe. We all saw that time and again.

I dunno. The simple answer is to blame it on Hemsky. But I don't think that is fair. We all saw both Horcoff and Smyth taking mad risks as well. But when Dvorak or Pisani was bumped in with them, or Peca on the starboard side late in a close game ... they played a different game. Hell, we heard Peca, Pronger, Moreau and Pisani all make veiled and not-so-veiled references to it during the year. Pisani was the most gracious, talking about how "we weren't cheating for offense in this one". He ain't talkin' about Moreau, Peca, Reasoner and Dvorak. And Stoll maybe got caught deep too much but rarely against quality opposing forwards.

In any case I think we'll see more of it this season. It just doesn't make sense to invest so much in the kid to abandon it now. Especially not with MacTavish on a four year deal, if he turns Hemsky into a Palffyesque type of 5on5 guy (he's already brilliant on the PP) then MacT is going to be looking a lot more intelligent in years 3 and 4 of his contract.

BTW: In fairness to this trio, I checked the faceoffs by opponent and zone; Horcoff did take a huge percentage of the draws against high end forwards both at home and on the road, and the lion's share of the defensive zone draws as well. Peca was second in the pecking order at this, but a clear distance behind, and he also took more defensive zone draws than offensive zone ones. Stoll's line and the 4th line du jour obviously took a high percentage of their faceoffs in the opponents end of the rink, and rarely against impact players.

My point, and I do have one: Maybe I've been a bit too harsh on that trio. They still should have done better, but the context, over the course of the full 82 games ... it is more severe than I had recalled.

8/23/2006 9:12 am  
Anonymous namflashback said...

Is Hemsky so gifted that a coach who is devoted to logic and matchups like MacTavish will simply say "screw it" and send him over the boards with his two best forwards no matter the matchup?

Well, the truly confusing evidence of this is the fact that he returned to the 94-10-83 line in the SCF. The prior 3 rounds rarely saw the threee of them together with a lesser light but much stronger ES player beside 94-10.

Now, moving 83 down to the 3rd ES line didn't have a strong effect on that line's ability to score -- and 83 was certainly more visible versus Carolina.

Was it because Carolina had far more balance on its lines -- and there was really not a suitable place for 83 to be sheltered? That's my guess. Better 83 be out there with 94-10 rather than Sammy-Stoll.

I can understand the desire to have Hemsky get the learning he needed, but I would rather have him pushed to the 4th line for shelter, play a ton of PP, and have the 1 line sub in somebody else. They could have won the SC for christ's sake.

I don't know how the EV+- played out in the Carolina series -- does it match your regular season analysis?

As a memory, I know that 94-10-83 were on the ice for the GA in game 7. Everyone is villifying Staios for the weak kick play on the neutral zone boards that allowed CAR to gain the zone.

Did Hemsky miss an assignment there? I remember seeing it and thinking that both of Smyth and Horcoff could have blocked it.

8/23/2006 9:13 am  
Blogger Vic Ferrari said...

namflashback:

Looks like we posted a similar sentiment at the same time regarding the SCFs.

I don't know what the thinking was. Maybe just too much time to prepare and practice. Hemsky has no peers on this team when it comes to dazzling in practice. I suspect that MacTavish and company were seduced by this, and chose not to dance with the ones that had brung them to the final series.

I think that they outsmarted themselves. And opportunities like that are very rare. And I think that it should have cost MacTavish & Co. their jobs.

8/23/2006 9:36 am  
Blogger mudcrutch79 said...

The Oilers goaltending didn't suck at ES though in 2003-04 it was pretty much league average or better IIRC. This year, it was terrible at ES. I looked at this but didn't write about it vis a vis Horc. I'll look tonight.

As for firing MacT...regardless of whether the decision was right or not, I don't know if you can tie the loss to it. I'd have to look to see who was on the ice for the G1 goals against but G2 had nothing to do with Hemsky, IMO. G4 and G7...I'd have to look. I seem to recall Hemsky's ice getting slashed mid-series though - I think I remember Riv writing after G6 that Hemsky had the least ES ice time on the team.

8/23/2006 10:17 am  
Blogger mudcrutch79 said...

The Oilers goaltending didn't suck at ES though in 2003-04 it was pretty much league average or better IIRC. This year, it was terrible at ES. I looked at this but didn't write about it vis a vis Horc. I'll look tonight.

As for firing MacT...regardless of whether the decision was right or not, I don't know if you can tie the loss to it. I'd have to look to see who was on the ice for the G1 goals against but G2 had nothing to do with Hemsky, IMO. G4 and G7...I'd have to look. I seem to recall Hemsky's ice getting slashed mid-series though - I think I remember Riv writing after G6 that Hemsky had the least ES ice time on the team.

8/23/2006 10:17 am  
Blogger mudcrutch79 said...

Yes, this comment was THAT important.

8/23/2006 10:17 am  
Anonymous namflashback said...

mc, you are right and I think RivQ and I were overanalyzing the top line.

using simple math

G1 EV GD = 0; PP/PK GD = -1
G2 EV GD = -2; PP/PK GD = -3
G3 EV GD = +1; PP/PK GD = n/a
G4 EV GD = 0; PP/PK GD = -1
G5 EV GD = +2; PP/PK GD = -1
G6 EV GD = +1; PP/PK GD = +3
G7 EV GD = 0; PP/PK GD = -1

Overall = EV GD +2; PP/PK GD = -4

Removing the G2 and G6 routs makes it EV GD +3; PP/PK GD = -4

Overlooking any of the specific plays, it wasn't 10-94-83's EV performance that sunk the ship.

How can you NOT score with so many 5v3 chances!!!^#@$&^#*

8/23/2006 10:51 am  
Anonymous namflashback said...

added -- Craig Simpson should have lost his job. Too bad he was extended BEFORE the playoffs began.

8/23/2006 10:52 am  
Anonymous namflashback said...

oh god. looking at those game numbers makes me ill about the Conklin-Smith circus in G1 and the no-show in G2.

8/23/2006 10:55 am  
Blogger RiversQ said...

bank shot: I actually thought of the linemates thing and checked it last night.

Hemsky w/ 94 or 10: +38/-40
Horcoff w/ 83 or 94: +47/-47
Smyth w/10 or 83: +48/-49

(That includes the +35/-32 when they were all together as well.) I don't think that really helps the case much. It seems that all three players were worse away from each other, but the sample sizes get pretty small because they played together so much - especially Hemsky only 7 5V5 goals while on a line without Horcoff or Smyth.

mc79: I look forward to the real ES numbers for 03/04 and 05/06. If you are able to justify more than 6 extra ES goals against I'd be very surprised and that still probably wouldn't be enough to explain the difference. Still that's the better way to attack the goaltending contribution, so I hope you get a chance to tackle it.

8/23/2006 11:48 am  
Anonymous PDO said...

Tyler beat me to it.

RiversQ, I think you've forgotten how spectacularly terrible this team was on the PK in 03-04. 29th in the league from my memory, and somewhere around 15th in the league this year. That's a pretty big spread...

If I was going to guess, I would say that you've underestimated (severely) how much the goaltending kicked us in the balls at ES this year. I know our SV% between the three headed monster was .880... given that we had an average PK, I'd guess that means that we had average goaltending on the PK as well? Unless it was just a factor of the ridiculous number of blocked shots.

I'd wait for Tyler to come back with our teams ES SV% between 03-04 and 05-06.. but my guess is once we see that, add in that Hemsky spent a lot of time with Horcoff and Smyth, and it'll be crystal clear.... hopefully ;)

8/23/2006 1:17 pm  
Blogger Vic Ferrari said...

MC:

I think it is. I mean if Modano is a 5on5 minus player in a playoff series Dallas would be in a helluva lot of trouble. If he is a 5on5 minus player playing weaker opposition ... then the Stars barely have a snowball's chance. Now we know that Horcoff-Smyth ain't Modano-Lehtinen ... but the gap in player quality is a helluva lot less than the gap in famousness, at least in terms of 5on5 play.

So in the Stanley Cup finals, when 94-10-83 are gifted a shitload of minutes against CAR's bottom six forwards and end up as a EV+2 EV-3 as a trio, EV+2 EV-5 overall ... that just plain sucks. The plan was for the other lines to hold their own against the better players and for that trio to feast on the weaker forwards on the CAR roster ... and it failed. In terms of puck possession, scoring chances and goals for/against. Just terrible.

8/23/2006 1:30 pm  
Blogger RiversQ said...

PDO: I've already addressed the possibility that it has been underestimated. Twice in the same post actually.

I just spitballed it by extrapolating the incomplete 05/06 numbers on his website and assuming the league relative ES SV% was 1000 in 03/04. If you do that you get 30 goals difference for the team. That would be 8 and 9 goals for Smyth and Horcoff respectively. Is that enough to explain all those extra GA at 5V5? I don't think so, but it definitely closes the gap from 15-20 of unexplained GA to about 10 GA.

8/23/2006 2:51 pm  
Blogger RiversQ said...

By "his website" I mean mc79hockey.com.

And of course I'm talking about the Oilers' league relative ES SV% in 03/04 being 1000.

8/23/2006 2:54 pm  
Blogger mudcrutch79 said...

given that we had an average PK, I'd guess that means that we had average goaltending on the PK as well?

Christ, I have so much cool stuff that I never put online. If they'd give me an assistant here, I'd have him do it.

Anyway, the Oilers PK goodness is entirely shot prevention. They gave up the fewest shots/60 on the PK by an obscene margin.

8/23/2006 3:21 pm  
Blogger Doogie said...

Anyway, the Oilers PK goodness is entirely shot prevention. They gave up the fewest shots/60 on the PK by an obscene margin.

Given that they gave up the fewest shots period last year (and, again, the goaltending), that's hardly a surprise.

8/23/2006 6:31 pm  
Anonymous jeremy said...

so I've been playing with some stats... here are the giveaways by team... edmonton was bad (hemsky and pronger were both in the top 5 worst in the league in this category - torres, smith, horcoff were also in top 25)

EDM 1322
DAL 1049
N.J 986
STL 968
ATL 932
WSH 930
DET 922
BUF 903
L.A 901
TOR 892
CGY 888
MTL 870
OTT 864
S.J 861
NYR 844
NYI 824
PHI 773
MIN 740
T.B 722
COL 716
FLA 703
PHX 648
PIT 646
NSH 633
VAN 625
ANA 607
BOS 521
CBJ 515
CAR 496
CHI 486

I'm baffled by chicago...

Giveaways we were middle of he pack.

8/23/2006 6:41 pm  
Anonymous jeremy said...

oops I mean TAKEAWAYS we were middle of the pack.... you can see on giveaways we were the worst

8/23/2006 6:42 pm  
Anonymous Anonymous said...

The giveaway stat is a joke because it is so subjective. Clearly the stat guy in Edmonton has an itchy trigger finger because there is NO WAY we gave the puck up 3 times more than Chicago did.

Ian

8/23/2006 7:00 pm  
Blogger mudcrutch79 said...

On the topic of Horcoff.

This year he was EV+ 60 EV- 62. That works out to EV+3.4 EV-3.5 per hour.

On those 62 goals...the Oilers EV save percentage was .898 this year. If that was there save percentage when Horcoff was on the ice (which it probably wasn't), Horc was on the ice for 610 shots against. With a league average save percentage of .915, that works out 52 goals against. EV+3.4 EV-3.0.

That's assuming that a) there was a .915 save percentage when he was on the ice and b) the chances against weren't materially worse. If those are the case though, the goaltending looks highly blameable to me.

8/23/2006 8:12 pm  
Blogger mudcrutch79 said...

Oh and because enquiring minds want to know...Horc was EV+46 EV-47 last year. The Oilers were a .921 ESSV% team in a .922 ESSV% league. Unless Horc somehow caught a lot better goaltending this year, or gave up shitloads of worse chances, I'd argue that 05-06 was a better year, EV wise.

8/23/2006 8:16 pm  
Blogger mudcrutch79 said...

And to clarify, I'm not saying that Vic is wrong, I'm just saying that the numbers don't back this position up, I don't think.

8/23/2006 8:28 pm  
Blogger JavaGeek said...

I have a study that shows that the minus statistic can be +/- 10, in other words players minus statistics include a lot of "luck":

I would contend the minus rating has significant error to the factor of ± 10 for most players, meaning a -10 is potentially a 0, or +10 is potentially +20 or even 0 (big difference).

8/23/2006 9:19 pm  
Anonymous oilerdiehard said...

JavaGeek said...

I have a study that shows that the minus statistic can be +/- 10, in other words players minus statistics include a lot of "luck":

I would contend the minus rating has significant error to the factor of ± 10 for most players, meaning a -10 is potentially a 0, or +10 is potentially +20 or even 0 (big difference).



Using words like "luck" and intangibles (even though you did not use this one) in the middle of stats convention is chucking a grenade in the bunker ;)

Brave man.

8/23/2006 10:58 pm  
Blogger RiversQ said...

Hmmm... Why is there such a discrepancy in Horcoff's EV+/- Mudcrutch? You've got 3 extra goals for last year and 8 more for this past year. I wonder where that comes from. Are you including 4V4 here to make the EV icetime and SV% numbers more applicable? If so, it's an answer but not exactly one for the 5V5 question.

Anyway that's down to 5 extra GA which surely could be explained by luck. Just to add - Horcoff's ES SV% was probably less than .898 last year, but I take it we don't actually know how many SA he gave up in his icetime. True? Of course, the ES SV% behind him in 03/04 was surely lower than .921 as well for the same reasons. It's probably a fair assumption that the difference due to quality of shooters and the number of shots they generate cancels out in his case.

However, Horcoff's only half the question because there's still Smyth's 18 goal difference to account for, which started the whole question in the first place.

javageek: I'm not following how the minus stat can be separated from the plus stat, but I'll wait to see your plus post. I agree with Vic in this case - I don't think defense and offense can be separated from each other beyond accounting for goaltending.

8/23/2006 11:07 pm  
Blogger RiversQ said...

oilerdiehard: You're clearly not familiar with many of the positions presented on this board. You should read the posts in the archives. Luck is no stranger to this blog. Not in the least.

8/23/2006 11:08 pm  
Blogger RiversQ said...

Ah, I checked. The difference is 4V4 stuff.

8/24/2006 1:08 am  
Anonymous Anonymous said...

what do you guys think about re-uniting the Smyth/Horcoff/Winchester line...

8/24/2006 2:16 am  
Blogger Chris said...

Rivers just a thought that has occurred to me but couldn't the discrepency your pointing to have alot to do with quality of opposition. I mean its been raised here and on various blogs and forms that the northwest division is the most competitive division in hockey. Schedualing changes which see us playing against stronger teams more frequently could have an effect upon goals for and against. Gaining more games against Calgary, Colorado and Vancouver and losing games against Washington, Florida and Pittsburg definately would have an impact.

I think you can probably make a fairly strong argument that because of the divisional intensive schedual the Oilers played more games against playoff teams this year than they did in the past.

8/24/2006 3:09 am  
Blogger JavaGeek said...

javageek: I'm not following how the minus stat can be separated from the plus stat, but I'll wait to see your plus post. I agree with Vic in this case - I don't think defense and offense can be separated from each other beyond accounting for goaltending.
I agree to a point, good offense = good defense, but there are ways to seperate the two. In fact it's important to seperate the plusses from the minusses.

The goal of offense is to shoot and score, from where ever possible. People with better shooting percentages are better because it takes fewer shots to score. Plus statistic comes from this.

The goal of defense in hockey is much more complicated, but IMHO easier to quantify. The goal of defense is ultimately to prevent goals. One way to simplify that is to say that the goal of defense is to limit shot quality and quantity.

The error argument comes directly from assuming that every second in hockey is binomial in terms of goals scored (you score or you don't). In 70,000 seconds of hockey most players will have a minus of about 50, which results +/-14. I found if you look at shot quality against and shots against that you get around the same error as the above approxemate, making the minus statistic almost worthless...

8/24/2006 8:54 am  
Blogger JavaGeek said...

What I'm really trying to say is that differences are to be expected even if all things were equal...

8/24/2006 8:55 am  
Blogger Vic Ferrari said...

javageek:

Hrmmm. You've combined shot quality with shots against (presumably for just the last bit of the season when the NHL started recording TOI). And the spread you got is similar to that of roulette wheel spinners, each shot against being a spin. Of course you could delete 'shot quality' and get the same result just using shots, or even Corsi's shots-directed-at-net.

Then again you could add in eye colour as a parameter, together with shots against, and get a similar spread, because the latter is driving it.

Of course some guys just are unlucky. We all see bad goals go in or crazy deflections off of skates go to an open shooter. As Lowetide says above ... shit happens. Hell, in 03/04 the Stars were more likely to get a shot on net when Modano was on the ice at 5on5 than any other player, and less likely to get a shot on their own net ... yet he had crappy EV- and EV+ numbers that year. According to the play by play guy in Dallas he hit a ridiculous number of posts that year too (and very oddly, only hit one post and one crossbar in 05/06). Shit happens. Get enough fuckers spinning the roulette wheel and a few poor bastards are going to have madass losing streaks, and a few lucky buggers are going to have insanely good fortune. Most guys will end up around the holdback of the wheel, but there will be outliers, in the right measure, simply because there has to be. BTW, as you're a VAN fan ... Naslund lead the West in this regard in 03/04, though it was much less impressive because, as you know, Crawford typically worked to keep him and Bertuzzi away from quality opposition as much as he could.

Go through all the teams in the league and it was the usual suspects for each team leading the parade at this: Sundin, Koivu, Elias, Fedorov (surprisingly), Thrornton, Richards ... which is all the more impressive considering how much these guys play against each other. Lidstrom was the only D leading his team at this, mostly because he's both damn good and because he played behind a team with three nearly equally good forward lines in front of him.

Pisani lead the Oilers in this regard, which was a bit of an oddball, though alert Oiler fans probably remember that Horcoff and Smyth both started the season playing injured and were a bag of suck for the first 6 weeks of the season or so. And Pisani was playing with Reasoner and Moreau and doing well. When Horcoff and Smyth were both healthy was the time that Fernando played the most with them. So I think that it's pretty reasonable to think he got zoomed along a bit there, still a helluva 5on5 hockey player.

This is of course derivative of Neilson's scoring chance + and - system, which has been copied by everyone. It's a tool that the Oilers use to evaluate their drafted prospects as well according to Jeff Ward. Except for instead of scoring chances we're using either shots, or in the case of this season, shots directed at net, because that's what is available. And of course if you don't account for the quality of linemate and opposition it doesn't really matter anyways.

8/24/2006 9:48 am  
Anonymous Anonymous said...

Do you think there is a chance that Hemsky was given the tough minutes to keep his numbers down? Kind of like the opposite of the pump and dump....

8/24/2006 3:47 pm  
Blogger RiversQ said...

Some really excellent posts here. Thanks again to mc79 for the ES goaltending data. I'm not completely convinced that it was all goaltending but it surely explains at least half of it. Tyler, I'm guessing you have the Oilers down for about 118 5V5 GA last year if they had league average goaltending? That's a massive difference.

Chris: You might be onto something there. It's probably alot easier to compare two schedules in the same year, but maybe Vic could take a crack at this one with his GD approach. I think this would be harder to do under normal circumstances but the 03/04 to 05/06 change was pretty substantial.

bank shot: I revisited your points about Hemsky and I think you were close there. I don't know the shared icetimes, but Hemsky had just 7 GF+GA away from 94/10, while Smyth and Horcoff and 11 and 20 while away from the 1st line. I would interpret this as meaning Horc still played with high event players when he left the first line, Smyth played with mostly low event players and Hemsky played almost exclusively with low event players. Since Peca was the hockey equivalent of a chloroform soaked rag last year and since we know he played at least some with Hemsky, I think your theory has some legs.

Anonymous: Hmm... I actually think the fact that Hemsky didn't get an extra 120 PP min to put him in the same ballpark as Pronger is probably better evidence of the anti-pump-and-dump. If the Oilers' situational minutes distribution is the same this year, then Hemsky needs 500 PP min IMHO. I think Vic's explanation is probably better for the ES toughness - it was more of a development tool.

8/24/2006 10:47 pm  
Blogger speeds said...

that is one thing that EDM doesn't have to be worried about with Hemsky now, him inflating his stats on the PP.

Having said that, anyone think he'll get anything near the Kovalchuk treatment this year?

He should see, by far, the most PP minutes of any Oiler this year.

8/24/2006 11:24 pm  
Blogger Showerhead said...

Hey guys - I've been gone for a while and will be for some more time yet but I wanted to comment that RiversQ made an excellent post.. and also, these are fantastic commetns. Great job everyone!

8/25/2006 4:06 pm  

Post a Comment

<< Home