Wednesday, May 03, 2006

Wings still better on balance?

I not 100% sure which game it was. Maybe someone can help me here. I'm thinking it was the 2-0 defeat to the Red Wings on April 11. Might have been a Dallas game, I suppose. It was almost surely a game in an American city. I believe this because I was listening to the local crew's colour commentator and play-by-play guy as they gradually, but persistently, lowered the intelligence of the listening audience. It was already the third period, so the game-day beer was well into fulfilling its engineered purpose of numbing my brain. Between the commentary and the beer my senses were dulled sufficiently that I'm not sure exactly what was being said, but it went something like this:

Voice #1: Blah blah blah Oilers blah blah seem to suck blah blah.

Voice #2: Yeah, you know a lot of commentators are saying that this Oilers team is a team they wouldn't like to meet in the first round of the playoffs.

#1: Blah?

#2: Definitely. I..I just don't see it. Unproven goalie, unable to generate offense, and several other shortcomings that are likely to be true.

I know what you're thinking, of course: that there's a lot of #2 in most TV commentators. True. I hope someone can recognize that specific conversation and recall that exact exchange and can report it here. And for the sake of the Irony Gods, I hope it really was the Detroit crew, because we now have a pretty good idea who won the first round series between the Oilers and Wings (even though the stat to prove it hasn't been fully invented yet).

The main reason I think it was the Detroit game is because I seem to remember a strong reaction of indignation to what was being said. Not because of the suggestion that the Oilers were sucking. I believe they were at the time. It was the idea that the commentators were looking down from the top of the standings when their home team was perched there only because they had the luxury of taking easy route up via the Central division. Yes, despite the beer, and ignoring the blood gushing out of my ears, I was thinking about the unbalanced schedule.

So, really, just how good was that Red Wings team that may or may not have lost to the Oilers in 6 occasionally thrilling games?

Balancing the schedule

One way to try to answer this is to weight the actual 82 game regular season schedule in a way that accounts for the differences in strength. Sagarin, for example, is known for his use of uneven tournament models to create team rankings. But since I'm not a journalist or lawyer, I find these heavy math models difficult to work with.

A different approach would be to simply make up my own weighted schedule by subsampling the real 82 games schedule. The results on this schedule can be used as a basis for estimating ranking. The Wings had 8 games against Chicago and the Oilers only 4? Take only 4 Chicago games each. If one just selects the western conference, for example, there was a virtual 56-game season played in a 16 team league: 4 games against 14 opponents (each team in the west played each other a minimum of 4 times). Simple, right?

Actually, there are a lot more virtual seasons that one could define since certain teams played each other 8 times, meaning there are 8 choose 4 = 56 combinations of games one could sample for each pair of such teams. Can something better be done? Well, one could treat each season series between teams A and B as providing an estimator a longer series. If Calgary and Edmonton played each other 82 games, presumably the results would closely resemble the results after 41 games projected to 82, which would (perhaps) somewhat less closely resemble the results after 20 games projected to 82, and so on. Using this idea, one could try to project all head-to-head season series to an 8 game series, making use of more of the head-to-head data. However projecting a 4 game to an 8 game, it might be argued, also magnifies the inaccuracy of using a smaller number of games. One might prefer projecting down to the level of a 4 game series. Or one could instead select a 6 game series and project the 4 games up and the 8 games down.

If one can even out the schedule differences thusly, the remaining question is how to develop a ranking from the virtual results. It seems like everyone since Mullet has been using goal differentials to assess the strength of a team. Given a simulated season and a scaling factor for each head-to-head series, one could then calculate goal differentials and use these to rank the teams.

It turns out there is program on my computer that does just this. It discards all empty net, OT, shootout, and penalty shot goals. If I run the thing for western conference games using a weighting of 5.9 (just less than the 6 mentioned above), I get the following rankings:

1 DET 81.12 96.00 -14.88
2 DAL 57.53 41.00 16.53
3 COL 32.45 28.00 4.45
4 ANA 28.76 33.00 -4.24
5 NAS 25.81 30.00 -4.19
6 MIN 22.12 13.00 9.12
7 EDM 18.44 3.00 15.44
8 CGY 11.80 23.00 -11.20
9 SJ 5.16 30.00 -24.84
10 LA -8.11 -18.00 9.89
11 VAN -27.29 -5.00 -22.29
12 PHO -30.24 -23.00 -7.24
13 CHI -68.59 -60.00 -8.59
14 CBJ -72.27 -59.00 -13.27
15 STL -76.70 -74.00 -2.70

The 5.9 scale factor effectively projects the goal differential to an 82 game season, with 5.9 head-to-head games against each team. This table is not to be taken very seriously, I think, though it's arguably useful in using numbers to mull over Detroit's luck, or lack thereof. By this model, the goal differentials of Detroit is inflated by the division they're in, but if you treat all other teams fairly too Detroit still outclasses the rest.

I suppose this table just adds another rough voice to the chorus that mudcrutch79 is singing in, namely that some things likely went the Oilers way this year. And just think: just two bounces in the regular season and the Oilers are out of the second season and the savages are hunting for the heads of MacTavish and Lowe. As it is, they're in the second round and Stevie Y is asking "why?"

What a game!


Blogger RiversQ said...

Good stuff oilswell - I agree that goal differential is the best tack to take here as most of the others probably will as well.

The one missing factor here, which you had no choice but to ignore of course, is the impact of the deadline deals. The Samsonov/Reasoner deal has probably made relatively little difference to the GD on this team, but the Roloson acquisition has a palpable impact particularly since the team adjusted to the changes and started playing in front of him. Of course, this has little to do with the Red Wings, but the relative disparity between them and the Oilers is probably further reduced by that factor.

5/03/2006 10:21 pm  
Anonymous iwocpo said...

Couple things: I doubt it was the Wings announcers because Redmond and Daniels are largely complimentary to everyone the Wings play. Not only that, but down the stretch they repeatedly referred to Edmonton as the team we didn't want.

Secondly, I doubt Stevie Y was asking "why". He knows why better than the rest of us.

Thirdly, I hope the Oilers aren't dwelling on the first round like you seem to be. No pun intended.

Fourth: good luck.

5/04/2006 8:45 am  
Blogger Vic Ferrari said...

Good stuff oilswell. A pretty sensible way of looking at it IMO. Your goal differential numbers, loosely applied, probably pegs the rate of scoring chances in that series pretty well, you're maybe flattering the Wings a bit.

I'm not sure how many periods there were in which the Oilers outplayed and outchanced the Wings, but it wasn't many, maybe three or four. Probably 10 or so others pretty close, say outchanced at a 5 to 4 clip or similar. But on this one, maybe it's best if we leave the cadaver in the cooler with the door shut. :-) Accolades go to the winners, uncomfortable questions about desire and heart go to the losers, that's the way it works.

On the "no respect" thing. I heard several times from Babcock that when the Wings started doing advance scouting for the playoffs they watched a couple of Oiler games, decided it was damn unlikely that they would finish eighth and focused on other teams instead, so they hadn't followed them as closely as they wish they had. In mid or late March I also heart a beat writer the Wings on local sports radio here (from a Windsor paper I think) say that the Wings liked their chances against anybody right through the coming playoffs, but that the Oilers were a team they'd rather not finish 8th. Might just be false flattery of course.

5/04/2006 10:23 am  
Anonymous huffpuff said...

In mid or late March I also heart a beat writer the Wings on local sports radio here (from a Windsor paper I think) say that the Wings liked their chances against anybody right through the coming playoffs, but that the Oilers were a team they'd rather not finish 8th. Might just be false flattery of course.

Yeah, as it turned out, it wasn't much of an advantage for the Wings in finishing first. They ended up facing the team they least liked to face (assumed those rumblings out of Detriot are true), a team that arguably did the most retooling in the 2nd half to improve themselves in the West. Both teams ended up playing 3 games at home. So they didn't even get a chance to have that extra home game to earn the extra revenues (here, of course, I'm totally disregarding the discrepancy in each team's take per home game).

I think perhaps it's time for the NHL to shake things up a bit to give the higher seeds more of an advantage. I've always liked the idea whereby the top 3 seeds get to pick their opponents from the bottom 4. i.e. the 1st place team gets the first pick, then the 2nd place team, then the 3rd, then the 4th gets whoever is left over.

If nothing else this could make for better drama. Imagine the 5th place team gets picked by the 1st place team. That's a blatant lack of repect right there and I'm sure the media would drum it for all it's worth.

Having said that, I'm not sure if it would have helped the Wings much this season. After all, the weakling is seeded at #4. And the unbalanced sched unfairly pumped their point total. But still I think this system might be fun to try going forward.

5/04/2006 1:42 pm  
Blogger Dennis said...

The disappointing thing is that Bettman's on the record as saying the format won't change in time for the '07 season. Six games against a divisional rival are enough, IMO. And it didn't serve them well but we could see that teams like Det and Nsh benefitted from playing in the Comedy Central.

5/04/2006 2:09 pm  
Anonymous huffpuff said...

Six games against a divisional rival are enough

I'm with you here. Actually, I wouldn't mind seeing the whole concept of "divisions" be done away with once and for all. Perhaps the NHL marketing dept has evidence to the contrary but, personally, I find it much more interesting to see variety in matchups between teams & players. I was pretty happy to see the Flames lose last night so we don't have to play them. And it's not b/c I don't think we can beat them...but after all those pre-season & regular season games, I felt it's just overkill (esp with how boring they play).

Don't get me wrong, those BOA back in the 80's were great, but that's b/c they were just 2 great teams that really should have matched up in the finals instead. Looking back, I was probably the happiest guy when they went to the 1 vs 8 playoff format. I mean, year after year of MTL vs BOS in that old Adams div just made me wanna puke.

5/04/2006 3:39 pm  
Anonymous oilswell said...

If I were Yzerman and I knew the likely odds, I'd still ask "why, why, why". Probably just me. And I'm not trying to parade the cadaver, I'm just slow. I may get to analyzing the preseason sometime during the next lockout. Plus, maybe someone noticed that I put other teams in the table besides the Wings?

I understand the urge to tweak the playoffs, but for whatever reason I can't figure out a clean argument for why top teams should get an "advantage". Is it to make the playoffs more predictable? Is that good for fans? Is it because better teams don't already have enough advantages over poorer teams? I wonder if adding MORE risk wouldn't force GMs to reconsider "going for it" at season's start or trade deadline.

5/04/2006 4:16 pm  
Blogger Vic Ferrari said...


I think the comments about the announcers near the end of your post just gravitated me towards the Detroit issue. And I started to fear that this was a precursor to tearing down that series, and while normally I love that sort of stuff ... right now I'm riding a wave of Oiler optimism and I'd like to stay in that zone. :-)

I'm surprised at how much San Jose gets knocked back by this way of looking at it. I'm guessing that this is because they did really well in their own division, no? I make San Jose a goal diff of +37. Just by summing the total goal diffs of the opponents for every game (overall nearly +100 for S.J opponents this season) ... then the program applies the corrections to all the teams and runs them again, ad nauseum, until it stops making a difference. Obviously that wouldn't work as well as Sagarin if there were gross differences between division qualities and only a handful of interdivisional games ... but I think it's fair here, seems so to me anyways.

If you just look at the outscoring numbers since Thornton got there, works out to about +60 for S.J ... so about a 60% chance of winning at SJ and about evens or 49% at Rexall by pure points. I'd improve the Oilers an extra 3 or 4% for the addition of Roloson, but my guess is that the oddsmakers won't. In any case, pretty close to a coin toss here methinks, between two teams that play a pretty fun game to watch. Wish it started sooner. :-)

5/04/2006 10:16 pm  
Anonymous oilswell said...

vf: I don't know how to explain SJS's drop here, as I haven't looked into it. It could be they spent most of their early schedule in tough west divisions and stormed through the east after heisting Thornton.

5/04/2006 11:38 pm  
Anonymous oilswell said...

Also, the third column gives the real goal diff using the described goal selection method (+30). 7 goals off your number, so what's the difference?

5/04/2006 11:40 pm  

Post a Comment

<< Home