Saturday, December 09, 2006

Desjardins' Quality of Opposition Numbers

Gabe Desjardins is a guy that Lowetide keeps referencing with regard to prospect point projections. He has his own site now, and there is some good stuff just posted there as well as links to some other things he has written in the past.

There are now dozens of guys analyzing hockey out there in the blogosphere. And there is some terrific data gathering going on and a tremendous amount of effort applied to analyzing it. And unfortunately almost always some tragic leaps in logic being taken and some really mad conclusions being drawn. But to my eyes anyways, this guy can be lumped in with mudcrutch, lowetide and the IOFers as someone whose stuff pases the BFS (Basic Sense) Test. We'll all disagree on some aspects, not so much on the math side but on the 'how it applies to what really happened on the ice' side, but to my mind anyways; there is a strong thread of reason in all of it.

Gabe lists the quality of opposition level for every player in the league in a post there. I've dumped that into Excel and sorted out the Oilers stuff, this is what is shown in the pick. He doesn't explain how he does it, and his rating is a small number with no specific meaning defined. I've x100ed it and cropped off the decimal points to make it easier on the eyes, so it's a dimensionless number here. [Side note: Perry and Getzlaf play the softest minutes in the entire league by his metric]

From an Oilers point of view, it looks pretty darn rational to me. Hopefully Dennis or any other people who watch a lot of non-Oiler games, and that notice matchups, hopefully they will have a check through and see if Desjardins' stuff meshes with what they have seen.

Also Desjardins has lists of the guys who draw penalties from the opposition for 28 of the teams, I have no idea where he got that stuff but it's really interesting and it matters.

28 Comments:

Blogger Earl Sleek said...

[Side note: Perry and Getzlaf play the softest minutes in the entire league by his metric]

That already passes a sniff test. I'd doubt any system that said otherwise.

12/09/2006 11:50 am  
Blogger Lowetide said...

Terrific stuff, Vic. I'd like to see HOW he gets to where he gets to (there's a sentence) and he did tell us in regard to the Junior to NHL equivalencies in an article a few years back.

But the new stuff on his site is just terrific. Great find.

12/09/2006 11:53 am  
Blogger Asiaoil said...

That we are 6 games above .500 with Shaggy and Smid playing those kind of minutes is remarkable - and a testament to how well Roloson has been playing.

The potential trickle down effect of obtaining a top pair dman is obvious here though - that guy gets Shaggy's minutes - Shaggy gets Smid's miinutes - Smid gets MAB's minutes - MAB gets to play somewhere else - and the team gets waaaaaay better.

12/09/2006 12:51 pm  
Blogger Vic Ferrari said...

Earl:

On Anaheim. He has your boy Pahlsson playing a level of opposition that is miles ahead of anyone on his team. And he shows Marchant and Pahlsson as guys being dragged down by their linemates in the 'quality of teammate' stuff. So as you say it passes the sniff test.

I mean Pahlsson only plays marginally more icetime at evens than Niedermayer-The-Lesser, and it's all bad stuff ... coming onto the ice with the puck going the wrong way in the hands of Ryan Smyth, extra man out there for the own zone draw against Sakic, that sort of thing.

So almost every blogosphere analytical bugger out there is going to mine the data and determine that Rob Niedermayer is clearly the better defensive player and Carlyle is a dumbass for not using him more instead of Sammi. Which is so wrong it hurts me behind the eyes.

For an Oiler perspective ...
same goes for Lupul and Horcoff the last three games at evens. In terms of icetime Horcoff is playing almost exclusively with Lupul in this stretch with one or two short snips here and there, and starting early here and staying out for a bit longer on the next shift there.

Yet Horcoff's EV icetime without Lupul in these last three games is disasterous in terms of shots-for and shots-against, or shots-directed-at-net. And failing divine intervention from the hockey gods, eventually it will be in terms of GA too if MacTavish keeps running these lines. Is Lupul driving the results? Sweet fuck no. And neither was Hemsky last year for the exact same reasons. In fact the exact opposite, but the young players need to learn and they don't need to get destroyed doing it.

I'll post a bit on this shortly, because right now the games are fresh in our minds. Just to end-run the Oilogosphere from quoting internet asshats in March and linking to stupid shit that shows, after much math, that Lupul is clearly a EV force (even Jovo probably couldn't shut this guy down!) and Horcoff and Smyth are riding his coattails.

12/09/2006 12:58 pm  
Blogger Slipper said...

I'm torn about this. Those numbers sure speak to the fact that Bergeron is sheltered. On one side of the coin you can say the dude isn't costing Edmonton games because he's producing positive net results in the role he's given, but on the flip side isn't he then indirectly responsible for the negative impact on the other D who have to compensate for his easy MINs with tough MINs?

As a point of reference Vic, how does that swing from tough to easy ice time compare to other teams? Is this common for teams to have guys on polar opposites of the bar (ecspecially for defensemen) or is the Oilers situation an extreme case?

Oh, and is this strictlly even strength?

12/09/2006 1:05 pm  
Blogger Vic Ferrari said...

Asiaoil:

Bergeron, Hemsky and Sykora may not have been playing very tough minutes, but they have been delivering results. The Oilers are more likely to outshoot the opposition when Bergeron is on the ice than any other player, the gap between him and Greene is stunning considering how much they play together. Followed by Hemsky and Sykora in this regard, those two have been delivering with the cherry minutes too. It's representative of zone time I think.

Same goes for Smid without Staios, gongshow. Or Smith with either of the two rookie D. Defencemen shouldn't drive outshooting results like this unless they're name rhymes with Lidstrom (granted Pronger did for the Oilers last year, he's pretty good too). But the Oilers have not one, but TWO players who ARE driving the results (Smid and Greene) ... driving them right off of the cliff.

IMO either Smid and Greene are a lot better by March, or the Oilers add a veteran D at the deadline and only one of Smid or Greene play in any playoff game. And it's 7 or 8 minutes against the other team's depth with a good partner who can exploit that opportunity, like Bergeron.

So far Tjarnqvist has been the Oiler's best defenseman by my eyes, though you could make a case for Staios and Smith as well I suppose.

12/09/2006 1:08 pm  
Blogger Earl Sleek said...

He has your boy Pahlsson playing a level of opposition that is miles ahead of anyone on his team. And he shows Marchant and Pahlsson as guys being dragged down by their linemates in the 'quality of teammate' stuff.

Yeah, I was so intrigued I did a little number-stealing myself, and looked at a combined teammate-opposition ranking (who's playing with bad teammates against strong opposition) and Pahlsson and Marchant rank 6 and 7 in the league by the added metrics. Rob N. falls to 64.

Not bad, I think.

I am surprised a little to see Scott N. getting the "tougher minute" score here than Pronger, but it's close enough, I guess.

Looks like a great site, by the way.

12/09/2006 1:23 pm  
Blogger Vic Ferrari said...

Slipper:

Generally speaking the 3rd D pair doesn't play anybody that's good on any team. Though less so now with the new rules, harder to change D on the fly plus the icing thing.

As for whether this is just evens ... I have no idea, but it's a terrific question. Ask Gabe.

To me Bergeron is a really useful player, a guy who makes a difference. Trades chances a bit too much to be out there against Sakic, but even then I can live with the bad if there is enough good. Last year Bergeron didn't play in the bottom pairing very often. Just after the new D were acquired with Greene (the only stretch where Greene had decent numbers, besides Bergeron-time the guy was sub-Cross and sub-Ulanov by any measure.

It pissed me off with Hemsky up with Horcoff last year and Bergeron up with Pronger last year. The Oilers bottom pairing was getting murdered yet Bergeron wasn't being sent down there to turn them from a liability to an asset. Madness. I remember writing a post called "The Tutelage Of Bergeron" at Oilfans, but I think I decided not to post it, either that or I decided to not read the responses. The love for Bergy from Oiler fans was just too strong then, and I would have been universally bashed by Oiler fans, and that would have brought out the worst in my personality, so I just left well enough alone.

In any case the player doesn't decide who he plays against, he just makes the most of his opportunity. Bergeron and Hemsky have clearly delivered this season. And in fairness to the coaching staff, having been bumped up to a higher weight class last year is surely part of the reason. Play a stretch against Sakic types and you start feeling like an aging, overweight sparring partner out there. Do it for long enough, then get shifted to a regular gig taking on welterweights like Bret McLean ... suddenly you feel like you're 20 pounds of muscle stronger and you can kick their ass all over the ring.

12/09/2006 1:34 pm  
Blogger Earl Sleek said...

As for whether this is just evens ... I have no idea, but it's a terrific question. Ask Gabe.

I have no idea either, but based on the Duck numbers, I would guess yes?

Pahlsson/Marchant kill all our penalties and see no PP time, and I gotta think that's factored in, given the score spread/league rank.

Then again, Getzlaf kills some penalties; Perry doesn't. Getzlaf gets some top PP time; Perry doesn't do that either. It might be a wash?

12/09/2006 1:45 pm  
Blogger Earl Sleek said...

I mean by "guess yes" that it includes all situations, not just even-strength. Sorry that wasn't clear.

12/09/2006 1:46 pm  
Blogger Asiaoil said...

Here is a bit more meat for the pack - the Desjardins quality of opposition numbers along side an overall on-ice versus off-ice +/- (soory for the ugly formating but you get the point)

Opp. Rating
SMITH 11 -1.92
TJARNQVIST 9 -0.19
SMYTH 8 0.89
PISANI 8 0.02
HORCOFF 8 -0.78
THORESEN 5 0.99
SMID 3 -0.4
REASONER 2 -0.39
SYKORA 0 1.46
TORRES 0 0.91
STAIOS 0 0.87
HEMSKY 0 0.62
LUPUL -3 -1.39
STOLL -4 -0.65
GREENE -6 0.33
BERGERON -9 1.87

Shaggy and Smith clearly struggling - but against very tough opposition
Smyth well what can you say - I think about trading him then numbers like this show up
Lupul and to a lesser degree Stoll have been getting killed playing soft minutes - unacceptable
MAB is a great player in a very well defined role - he beats up weak opposition very effectively but dont let him near and significant opposition or these numbers will change fast

12/09/2006 1:48 pm  
Blogger Vic Ferrari said...

Terrific stuff, AO.

And how about Raffi? Good numbers all things considered in spite of taking our wrath. Now that Nieuwendyk has hung up the skates, that guy is now officially the most consistently lucky fucker in the league.

12/09/2006 2:23 pm  
Blogger Vic Ferrari said...

Earl:

I think you're right to agree with slipper, I suspect he's getting the separation beacuse of PP and PK time. I've been giving Desjardins too much credit here, but it's strange how well it seems to work vs eyesight, with most guys anyhow. Probably really screws some guys unfairly though.

12/09/2006 2:26 pm  
Blogger Andy Grabia said...

Is no one else intrigued by the fact that Thoresen is higher on that list than Reasoner? That one stands out to me.

12/09/2006 3:00 pm  
Blogger Slipper said...

Grabia, I'm sort of a honky in the ghetto when looking at this stuff, but I think the centers might be getting a rough ride here. When you consider that special teams is included and also situational events like defensive zone face-offs (where sometimes you'll have two C's on the ice in case one gets booted from the draw) then the wingers should look a bit better. Still, the idea that Thoreson plays tougher minutes then Reasoner? Is that legitimate, or would that be skewed by seeing soft scorers on the PK?

12/09/2006 5:31 pm  
Blogger Asiaoil said...

Vic - I like the breadth of info on Desjardins site as you can look at multiple indicators and see strength of opposition and on-ice results at the same time - very very handy stuff to get a more rounded look at performance.

Yeah Torres can be called lucky - but he's been doing this for a while so unless he's made a pact with our red friend with a tail he's doing something right and you've got to give the kid credit. He's a crazy bugger and I like him - if he was about 3 inches taller and 20 lbs heavier no one would go anywhere near him.

A few other thoughts given this info - that Thorsen kid has something going on as well - and Sykora is doing his job although I think we are seeing a lot of his numbers are Hemsky-driven. When everyone is healthy here are the lines I'd try:

94 71 83
14 10 34
18 19 28
15 16 26

That leaves 8, 78, 20 out but I'd ship 20 back down to Iowa without much of an issue. Gotta be a trade coming though as soon as we figure out the Smyth and Sykora dilema. Love Smyth but not at $5 million - and trading him for a serious dman makes a lot of sense to me as long as we can resign Sykora for similar dollars to what he is making now. Stoll is also on my trade list in spite of one good game where he gave up the body on the PK - he is still getting crappy results against soft opposition and MacT seems to have given up on feeding him tougher minutes. He's over-valued IMHO and we would not miss him much if at all.

12/09/2006 6:08 pm  
Blogger RiversQ said...

Asiaoil said...
Smid gets MAB's minutes - MAB gets to play somewhere else - and the team gets waaaaaay better.


There's no way this end of the trickle down ends up making the team better. Say what you want about MAB, he gets results and Smid has shown very little offensive ability. If Smid turns into a player he'll be from the Brewer school for sure. That's not a bad thing though, unless you're expecting offense.

12/09/2006 10:20 pm  
Blogger Asiaoil said...

Riv - I know what you are saying and freely admit that MAB does very well with 3rd pair minutes. BUT........MAB is 26 and given the makeup of the Oilers blueline with Smid and Greene already on the big team roster and Gilbert just about ready - I don't see a future for him. We just can't develop all these kids AND protect Bergeron. This was the year MAB was supposed to take the next step and start looking more like Rafalski - aint happening. It's the beef I always had with Laraque (another useful role player) tying up a forward slot we needed to use for developing wingers like Winny and Thoresen. If we still had Pronger on the roster then MAB would be a fantastic fit - but with Greene and Smid on the roster he is luxury we can't afford. Better to use him as part of trade package for a real top pair dman and get on with developing our young guys in slots outside the top 4

12/09/2006 11:09 pm  
Blogger Dennis-IOF said...

Good lord that is good stuff!!! This is practically invaluable info for GM's who are looking to make trades. I'm just gonna throw out some stuff right now without even addressing the Oilers on this list, ie we know which of our boys was playing the tough min anyway and I'm just gonna address the toughest min list and look at some of those guys plus/minus numbers.

Heading into tonight the Caps were -13 in GD yet Pothier's playing over 26 min and Morrison playing just three seconds shy of 21 min...they're 8 and 12 in playing tough min respectively and Pothier's +1 and Morrisson is even. That's just mind blowingly impressive. The B's can't catch a break can they? Morrison was highly touted and they dealt him for a Gonchar rental and now it looks like he's gonna pay off big time. And Pothier's come out of nowhere!!!

Let's see...what else...from a rental perspective Brew's -9 but he's "earning" that by playing the tough min. Boynton was -7 before his injury but he was carrying the mail for Phx. So..is it a coincidence that these two guys were the ones mentioned on HC as fellows Lowe was interested in? I highly doubt it. I don't know how many tough min Spacek was playing last year but Lowe snatched him early and he played a tonne of min for the Oilers last going off. Looks like Lowe's on the right path again this year,.

Hartnell's UFA at season's end and right now he has 11 goals and is +9 and these stats say he's playing tough min. Here's a guy about to get paid. I will say thought that I saw Trotz match Erat-Arnott-Kariya against Joe T one night so I assumed he was doing the same very game but these numbers say Hartnell's seeing the tough stuff and I can see that Arnott's down at the bottom so I was way off there. My Habs buddy told me that Johnson and Bonk were doing an absolute job for the Habs and he was dead on. Looks like that fucking knob Carbonneau figured out early on who could carry the mail. Back to the Preds..Shea Weber's #22 on the tough min list and he's plus 8 and only 21 years old. In terms of another guy the Oilers might pick up for their D..Vishnovsky's 22nd among tough min for D and he has 22 points and is -5 and is playing in front of terrible netminding so that's impressive.

Down in the bottom...no big surprise that Perry and Getzlaf are getting the absolute softest min. I watch quite a few Ducks game and Caryle has it figured out and has guys that will obey him. Fellows like Pahlsson and Marchant get the absolute shit min and then your Mcdonald line is getting out there against the softies./ It's a great system he has working there for sure.

Some stuff that surprises me:

Maurice was talking about how Peca was playing tough min but these numbers say otherwise. Guys like Staal and Jagr are cleaning up against junk as well. I noticed the other day taht Shanny was scoring fuckall goals at ES and now this makes sense. I guess Renney isn't exactly running out the Jagr line in a PVP matchup at MSG. They're five or six games over so you can't argue with his plan but that second line of Purcha, shanahan and cullen are taking the shit.

Playfair's doing a great job of getting Iggy out against the lesser lights and Hossa, Thornton, Kovalchuk and Crosby aren't doing their work PVP either.

And by this metric...just how bloody valuable does Lidstrom look?

12/10/2006 12:12 am  
Blogger Asiaoil said...

If this stuff is for real - and the Oilers numbers seem to pass the "smells right" test - then I am massively impressed by the two CHI kids Seabrook and Keith - holy crap are they putting up good results against tough opposition on a bad team at a young age

12/10/2006 5:51 am  
Blogger Bank Shot said...

Bergeron, Hemsky and Sykora may not have been playing very tough minutes, but they have been delivering results. The Oilers are more likely to outshoot the opposition when Bergeron is on the ice than any other player, the gap between him and Greene is stunning considering how much they play together

The fact that Greene is likely to be on the ice to end a penalty kill and Bergeron is never to be seen could make a pretty big difference in shots directed at net.

Bergeron gets more PP time then any other defenceman and next to no PK time. His numbers should be pretty damn good.

If Bergeron, Staios, and Tjarnqvist all shine at outshooting opposition when away from their partners it may mean a powerplay just expired.

12/10/2006 11:03 pm  
Blogger Jeff J said...

Yes, Desjardins' stuff is awesome.

It looks to me that the Quality of Opposition rating is simply the weighted average of the player's opposition rating, which Desjardins compiled here:
http://www.behindthenet.ca/player_stats.html

It's sort of a player-isolated +/-.

And it's very wrong to multiply the number by anything to make it easier on the eyes. The smallness of Gabe's numbers is significant. It says that the guys facing the toughest opposition are seeing only slightly tougher opposition than the guys playing the so-called 'soft minutes.' One of those other unmentioned numbers guys said this a while ago, but I guess he must have 'failed' the BFS Test.

12/11/2006 9:51 am  
Blogger Vic Ferrari said...

bank shot:

I'd suspect that the effect is the opposite, that the D who finish the PK as the D pair end up with flattering numbers at 5on5 afterwards. Just because that's one of the very few times in the game that a guy like Smyth is playing against players you've never heard of.

They DO end up with better EV+/- numbers as a result, I know that. Is that reflected in the shots-for/shots-against thing? I dunno, maybe not. I'd think so though, that's a long change for the D and to boot that's a time that the coach is probably happy to have them playing, because Sakic and Hejduk are sure to be resting on the bench for the other guys for a while yet.

They usually seem to play a full D shift then by my eyes, granted I don't notice the D changes without trying, I don't see them unless I'm trying to. I'll look for it noe though if I think of it.

12/11/2006 12:50 pm  
Blogger Vic Ferrari said...

JeffJ:

The sisu hockey guy? I think you're a cool guy, if my memory is right you have a bigass backyard rink for your kids. Mine is going for a shit this year btw, the combination of relocating it to high ground and building a "Mystery, Alaska" style skating runway to it, combined with my pathetic icemaking skills and really odd weather has resulted in mayhem. I'm shit as an icemaker, and I've been meaning to email you on that.

Back to point: Do you REALLY want me to perform a BFS test of your stuff? I rarely read it simply because I'm not a Habs fan. But I will if pushed, and I'd guess that you won't like it.

Just basic sense stuff is probably going to come down like a hammer, less so than most of the numbers nutters out there, because you're smarter than them. Still, I don't think you'll like it. And it's probably a huge waste of my time anyways, but somebody has to be made an example of I suppose.

Desjardins emailed me and asked me to have a look at his stuff a few days ago. I emailed him back asking about the basic premise for his numbers, his response was not clear. In any case I suspect you're right, he has used event numbers (which explains the strong general trend) but hasn't parlayed them through, which explains the anamolies.

So if you're Sundin, when you're in Detroit on Saturday you take the lion's share of defensive zone draws. Essentially every one with either Lidstrom or Zetterburg on the ice, usually both. Nasty icetime, a shitload of overall 5v5 icetime vs Z and/or Lidstrom as well, more than all the other Leaf centers combined I'd guess.

Just another Saturday night for Sundin. The heavyweights fight the heavyweights. Problem is that the heavyweights are so often throwing punches against each other that they make each other seem damn near average. At first blush icetime against Winchester and Thoresen seems more difficult, because they've been spending most of the year fighting unknown welterweights and (to their credit) holding their own. And if you're right (and I haven't checked, so I won't bash him unfairly) this is exactly what Desjardins has done.

I could go on and on on the subject until nobody was reading any more. :) It matters though, the coaches in this league aren't quite the complete morons being supposed by people who ignore this aspect, and CBJ and PHX aren't quite the powerhouses these kids projected them to be.

And if you traded the type of opposition icetimes of Smyth/Pisani/Horcoff with Winchester/Jacques/Pouliot ... there really would be a serious impact on the way the game played out, really. I know you don't believe that, but hopefully the head-to-head numbers stuff I've been erratically put up on the sidebar has a few people paying attention to it. Though I suspect that most people mad enough to read here already saw that stuff in the first plave without even trying.

12/11/2006 1:51 pm  
Blogger Slipper said...

The quality of opposition number has ot be driven by the Dejardins +/- number for the opponent a defender is playing against, right? So playing against a real star like Jagr at his +3.43 has to drive up the defender's "qaulity of opposition" ratio. The weird thing for me is by Jagr's standard of opposition, by the numbers on that site, it looks to me that either his coach is getting every match-up he wants at will, or else the other team is throwing jobbers out at him.

But I'm not a math wizard, but being driven to look at alot of stats by sites like this one, there's a bunch that doesn't seem sensible. The teammate rating,for instance, in a couple of cases seems spot on. Yet on further review, it doesn't seem to tell an authentic story in many cases. I beleive, of course, that there's a relationship between defensive pairings and forward units, but something about this doesn't seem to express it properly. Also, does anyone beleive that Sykora is the one driving the results with Hemsky?

A question for a guy like Rivers or Vic, is Ribeiro relally the guy tearing it up against the highest competition with below averagew linemates?

12/11/2006 2:28 pm  
Blogger Bank Shot said...

They DO end up with better EV+/- numbers as a result, I know that. Is that reflected in the shots-for/shots-against thing? I dunno, maybe not. I'd think so though, that's a long change for the D and to boot that's a time that the coach is probably happy to have them playing, because Sakic and Hejduk are sure to be resting on the bench for the other guys for a while yet.

Vic: I didn't mean the first shift after the PP or PK. I was refering to the ten to thirty seconds after a PP/PK expires when the players that were on for the special teams are still on the ice.

Like for example Smith and Greene just finish killing off a penalty and are dog tired, but Sakic still has possession in the offensive zone and Tjarnqvist rushes out of the box and ends up playing leftwing. In the resulting scramble the Avs get another shot or two before Roloson is able to freeze it and allow the lines to get back to normal.

Or Bergeron is on to end the powerplay and the Oilers see that the opposition player is on his way back into the play from the box so they take a quick shot and a rebound which count as ES shots.

Situations like that happen every game and I am sure that there are times when Greene and Bergeron have a 5+ gap in their shot differentials at the end of a game because of it.

12/11/2006 3:23 pm  
Blogger Vic Ferrari said...

Bank shot:

You may very well be right, obviously some powerplays end that way. I really find it hard to believe that most do though.

I guess a guy could check the way a bunch of powerplay's ended just by looking at the shift charts and then checking the play by play sheets for shots on goal, just to prove your point. Let me know how that goes ;)

12/11/2006 7:49 pm  
Blogger Slipper said...

Does anyone knoow the trick to the shift charts and such? Whenever I'm looking at those things, Like the TOI shit, the page resets and takes me back to the top. What is that? Some sort of auto-refresh?

It's frustrating to say the least.

12/12/2006 1:45 pm  

Post a Comment

<< Home