#253: 2012 NHL Playoffs - Round 2
SOGD, we have a problem.
Incidentally, we also have a problem with SOGD + SV%.
For those who have no earthly clue what I'm talking about, I suggest reading this post.
The first round of action in the 2012 Stanley Cup Playoffs is concluded, and, based on the initial predictions, you could say that things, especially in the Eastern Conference, are well off from what I expected.
The skinny for those who like to skim and not read:
The new model for predicting series winners using SOGD (Shots On Goal Differential) offset by SV% (Save Percentage), the record was 4-4 (getting the St. Louis/San Jose series exactly).
Our old model, using SOGD alone, went 3-5 (with the Los Angeles/Vancouver prediction coming out exact).
The old model has now been struggling two years in a row, while the more advanced predictor has already lost one more series than it did all of last postseason.
If things don't go well the rest of the way, the NHL may indeed have found a way to counter SOGD, and a different statistic may have become a better predictor.
Let's not get too far ahead, though.
First, what happened in Round One? Were any of the expected trends bucked for those best-of-sevens?
#1 Vancouver vs. #8 Los Angeles
The picks were:
Official Steeltown Sports pick (using SOGD and SV%): Kings in 7
Old model: LA in 5
The Kings were "unanimous" here, and they delivered. And it was mainly because of the goaltending they received from Jonathan Quick in games 2 and 3, where the Kings were outshot by at more than 20 shots per game (both Kings wins, regardless). Otherwise, this series may very well have gone longer. Vancouver's goaltending was actually slightly better than its season average. Both teams shot the puck during the series at a higher rate than their regular season averages. Quick was the difference in ending the series earlier.
#2 St. Louis vs. #7 San Jose
The picks were:
Official Steeltown Sports pick: St. Louis in 5.
Old Model: San Jose in 6.
Again, goaltending was the main story here, at least in the series' brevity. Brian Elliott came off the bench for an injured Jaroslav Halak in game two and didn't lose a contest. He posted a save percentage of .949 in 3.66 games. For the series, St. Louis had the better of the shot number, too. About 1.6 shots more per games. Not a vast difference, but San Jose was the 2nd-highest shooting team in the league in the regular season (33.8 Shots per game).
Former Stanley Cup champion (and current Shark) goaltender Antti Niemi performed slightly below the season average, helping the Blues take the most advantage of their slight shots sdvantage.
#3 Phoenix vs. #6 Chicago
Official Steeltown Sports pick: Phoenix in 5.
SOGD only: Blackhawks in 5.
It's series like this one that convinced me to use the updated model. By all rights, Chicago should have run away with this series in 5. They pounded the Phoenix net with pucks, but Mike Smith denied the Blackhawks again and again and again.
Even more impressive is how he was able to stay composed as each game (except for #6) went into overtime in 5 straight games (the first time a series had gone to OT in each of the first five games in an NHL series since the 1951 Stanley Cup Final).
But the goaltending for Chicago, a bottom-5 showing throughout the regular season, caved to a 2nd-best-in-the-league Phoenix team.
#4 Nashville vs. #5 Detroit
The picks were:
Official Steeltown Sports pick: Detroit in 7.
SOGD alone: Wings sweep in 4.
While both predictions were wrong, the official one was "closer," if that can be said. Nashville's efficient 5-game elimination of Detroit was pretty much due to the play of its netminder (and recently announced Vezina Trophy nominee...which I expected in the last preview) Pekka Rinne. In both games 3 and 4 of the series - which were in Detroit - Rinne posted outstanding .976 and .953 save percentages, despite facing over 40 shots in each contest. Nashville never managed more than 26 in a game. Wings goaltender Jimmy Howard was apparently not up to the task.
Alexander Radulov, while not invisible, was not the force that some might have expected. This was a patient team effort spearheaded by its goaltender.
Detroit had a shot advantage of nearly 9 more per game, but Rinne's .931 SV% vs. Howard's well-below season average of .883 proved to be Hockeytown's undoing.
#1 NY Rangers vs. #8 Ottawa
The predictions were:
Official Steeltown Sports pick: New York in 5.
SOGD alone: Rangers in 6.
Is it just me, or does it feel like the Rangers lost this series?
After a bold win by New York in Game One, jumping out to a 4-0 lead, neither team scored more than 3 after that (and only once scored the third goal in regulation...Rangers in Game 6).
Ottawa actually had the better of the shooting, which surprised me. Both teams had better than average goaltending (and when you're including Vezina candidate Henrik Lundqvist in that, that's saying something). Total save percentage was .937, and a little more than 60 total shots per game came to the nets.
The Rangers dug down when they had to, but most folks would look at this series and think that New York's stay in the playoffs might be shorter than most fans of the Blue hope.
#2 Boston vs. #7 Washington
Official Steeltown Sports pick: Boston in 5.
SOGD only - Boston in 5.
Both predictions were equally wrong, as the defending champions bowed out.
And, really, the only reason they lost was due to an unexpected, and, admittedly, incalculable factor: 22-year old Capitals goaltender Braden Holtby.
With injuries to the top two netminders, Washington was forced to play Holtby - who responded by posting a .940 SV%, including a mind-blowing .958 SV% on the road!
The return of Mike Green and Nicklas Backstrom helped to a degree, but in a series where each game was decided by a single goal, Holtby was the true deciding factor.
As for it being predictable, his NHL career's sample size was too small. He'd played in just 21 NHL games. While he posted a none-too-shabby .924 SV% in those contests, that would hardly be enough to go by. His AHL SV% is a little below that. And playing on the other end of the ice from the reigning Stanley Cup-, Conn Smythe- and Vezina-trophy-winning goaltender? Well, as some Spanish speakers I know would say:
#3 Florida vs. #6 New Jersey
Official Steeltown Sports pick: Florida in 6 (in overtime).
SOGD by itself: Devils in 6.
Of all the series I missed, and speaking strictly as a stats guy instead of a Penguins fan, this is the one that makes me cry inside the most.
Travis Zajac of the Devils ruined what was surely going to make me look like a prophet to some attractive puck bunny and I'd be all set.
And then, despite trailing 2-0 in the 3rd period in the deciding game, Florida teases me again by tying it, then losing it in DOUBLE OT.
I referenced in the last preview post about how the Devils were series favorites by a 1-to-2 margin. I think that anyone who took the Devils may have spent their winnings on antacids. Seriously, this was as even as a matchup as anyone could have wanted.
The Devils outshot the Panthers by an average of 4 shots per game, but Florida's save percentage was .05 higher. In the end, the Devils were just one goal (and a very late goal it was) better, ultimately outscoring Florida 18-17.
#4 Pittsburgh vs #5 Philadelphia
The predictions were:
Official Steeltown Sports pick: Penguins in 7.
SOGD-only: Pens in 6.
Despite my disappointment that, for the second year in a row, the Pittsburgh Penguins did not play up to their statistics, particularly on the goaltending end, I am proud of some of my homer Penguin buddies.
These are the folks who get mad when anyone levels criticism at goaltender Marc-Andre Fleury...ever...then ignorantly spout off that those people should just stop watching hockey. I have to laugh, because some of those angry folk used to play hockey on some level.
The irony is thick.
But these same folks have found it acceptable, in this series, to say the Fleury was largely awful (outside a few good periods in games 4 and 5) - but so was everyone else.
Penguin fans will be haunted all off-season by the "ones that got away," namely Games 1 and 2 at home, where first period leads of 3-0 and 3-1 vanished and turned into losses. Then there was the embarrassment in Game 3 where the team came unglued and worried more about fighting than playing hockey (and notice how the fighting did not bring them a victory...). Then, when players got suspended, the team decided they would play hockey and win a couple of games before finally running completely out of gas in Game 6.
Still, I'm gonna give myself a couple of cookies here for some other predictions made about this series:
I predict three fights, but no brawls in a series that is destined to go the distance.
Ok. A third of a cookie here. According to the penalty minutes assigned, there were only three fights (all in Game 3). And there was certainly a brawl. And the series did not go the distance. Why did I bother writing this paragraph?
[T]here should be more than a couple high-scoring affairs. It will likely be the highest-scoring series in Round 1.
Ding ding ding ding!
For the Penguins part, a lot of bad defense (which had been their staple through most of the regular season...seemingly starting to come unglued when Sidney Crosby came back...don't know if those two events are connected), atrocious goaltending, and the inability to skate with the Flyers in the 3rd period led to their humiliating 2012 doom.
The Flyers, though, showed determination and the innate ability to accurately use a timeout. Overall, the battle may have been won with Peter Laviolette outcoaching Dan Bylsma.
Now, let's review the SOGD of the playoff teams who remain (and I have to hurry because, as I type this, the second round begins in about two hours):
1. NY Rangers +0.7 (.920)
5. Philadelphia +3.8 (.903)
6. New Jersey +0.7 (.907)
7. Washington -2.2 (.909)
2. St. Louis +3.9 (.929)
3. Phoenix -2.0 (.925)
4. Nashville -3.2 (.919)
8. Los Angeles +3.2 (.924)
Three of the eight quarter-finalists are in the negative SOGD department, and two play each other, so there will be at least one in the final four who will try to be the 3rd team since 1990 to win a championship in the negative.
Also, on a trivial note, one of each seed has survived: 1, 5, 6, and 7 in the East, 2, 3, 4, and 8 in the West.
#2 St. Louis vs. #8 Los Angeles
This should be a fantastic series.
The Kings have already slain the President's Trophy Goliath in Vancouver, they have a couple of Stanley Cup Final veterans, and they have a Vezina Trophy nominee in net.
The Blues have methodically come from obscurity into the limelight by way of solid SOGD approach, and a couple of solid goaltenders that they acquired from previous playoff teams.
Neither team has ever won a Stanley Cup, so that should add a little "heart" into this series.
Official Steeltown Sports pick (using SOGD and SV%): Blues in 7
Old model: Blues in 7.
#3 Phoenix vs. #4 Nashville
This is our battle of SOGD-negative teams (sounds like a blood type).
On the surface, it looks like Nashville has a couple more goal scorers and the better goaltender. Not only that, but they made a couple of trade deadline acquisitions.
We probably wouldn't be talking about that, though, if not for Pekka Rinne.
Phoenix spreads the scoring around, and folks seem to forget that they won their very competitive division.
Goalie to blink first probably loses, because of these teams is used to being outshot. When both teams are in the negative, the results aren't quite as predictable. That said, this postseason hasn't been predictable, either.
Official Steeltown Sports pick: Phoenix in 6.
SOGD only indicates Coyotes in 6.
#1 NY Rangers vs. #7 Washington
Ovechkin vs. Lundqvist is probably the selling point of this series. And maybe a bit of "rematch" factor here as the last two times the Rangers made the playoffs, it was the Capitals, as a higher seed, who eliminated them.
Still, my eyes will be on young Holtby. If he struggles, it's possible that we could see a return to the net for, most likely Michel Neuwirth first, and possible the more injured Tomas Vokoun.
The Playoffs are a different animal. Some goalies rise, some fall. Holtby has begun his rise against the defending Stanley Cup Champions. Now let's see how he does vs. the Conference's top seed.
Official Steeltown Sports pick: New York in 6.
SOGD alone has Rangers in 6.
(I'm not doing this on purpose)
#5 Philadelphia vs. #6 New Jersey
This should be the most dynamite of this round's four series. They're the only divisional rivals this go-around, their home arenas are less than two hours apart (traffic notwithstanding). Philadelphia has a slight edge in the season series, going 3-2-1 against the Devils (getting 7 points vs. New Jersey's 6).
It will also be a testament to whether a well-rested team (Philly) with a week between games has the advantage over a team on just two days of rest.
As usual, I break it down in numbers.
Devils goaltender Martin Brodeur is on his way out of the league (probably just has one more fair year left in a Hall of Fame career) and his stats have been subpar for him. The key to a New Jersey win will be maintaining what was a conference-best lowest shots per game against average.
However, there's too much firepower on this very motivated Flyers team. The Devils may find themselves on the wrong side of a series that they just won themselves.
Official Steeltown Sports pick: Flyers in 7 (in overtime).
SOGD by itself: Flyers in 6.
Interestingly enough, both systems agree on a winner in all 4 cases, and they have agreed 9 out of 12 times on a winner through two rounds.
Unfortunately, the track record hasn't been so good. At least, not so far.
Now, I have to revise the picture going all the way to the end.
Last time, through all the projected winners, I had the Bruins repeating at the expense of the Blues.
Now, according to the SOGD+SV% model, the Rangers will win the Cup over the Blues in 6.
Come back again in a couple weeks the second round post-mortem/round three preview.