Saturday, May 17, 2014

NHL playoff predictions: Conference Finals edition

What have we learned from our projections of the second round of the playoffs?

1)  The Canadiens pulled off quite an upset. This is probably the first true upset of the playoffs. In the preview, I said, "If the Canadiens are to win this series, Carey Price will have to be amazing, and the Habs will need to win the special teams battle." Well, that is exactly what happened. Even though the Bruins were better at puck possession, Carey Price (.943 5v5 save %)  out-goalied Tuukka Rask (.929), and the Canadiens scored 7 power play goals compared to 3 for the Bruins. The Bruins were clearly the best team in the Eastern Conference, but now Milan Lucic will have to get his kicks by spearing pizza delivery guy in the nuts.

2) The Pens blew a 3-1 lead. Here's the thing: the situation isn't as bad as it seems. The Pens had better possession numbers in this series. Marc-Andre Fleury was good. However, Henrik Lundqvist was brilliant, especially in the final three games of the series. The Pens ran into a hot goaltender and some bad luck, and unfortunately gave up early goals. A lot is being made about Sidney Crosby's scoring woes, but he was a dominant possession player who ran into bad luck and a hot goalie. I fear we're going to see a lot of overreaction from Pens' ownership, but that's for another post. Bottom line: the Rangers had a better score-adjusted fenwick %, 5 on 5 save %, and penalty kill % than the Pens all season. Make no mistake: the Pens had a 3-1 lead and should have finished the series. But the Rangers are probably the better team. In the first round projection, I said, "The Rangers are a darkhorse; don't be surprised to see them go deep in the playoffs.

3) The Kings had more trouble with the Ducks than I thought they would. Losing defensemen Willie Mitchell and  Robin Regehr to injuries hurt the Kings, and the emergence of John Gibson really boosted the Ducks. The Kings are the better team, and likely the best team remaining in the playoffs.

4) The Wild had a solid showing against the Blackhawks, but it wasn't enough. The Hawks are a really solid team.

Speaking of the Blackhawks, I've commented on how the model may be underestimating their chances, Well, I think I figured out why: I use 5 on 5 team save percentage as a predictor in the model as a way to measure goaltending. The numbers are easily available, and they seem to be a good measure. There are some cases where this isn't true. For example, the St. Louis Blues acquired Ryan Miller at the trade deadline this season. The save percentage I used in the model was based on a lot more Jaroslav Halak (40 games) and Brian Elliott (31 games) than Ryan Miller (19 games). So the number in the model probably didn't reflect the true goalie situation. The Blackhawks are similarly affected. They had the worst 5 on 5 save % (.914) of  any playoff team. It was because their backup goalies were awful. Corey Crawford was fine (.925), but the Hawks backups played 30 games and pooped this out: Antti Raanta (25 games, .896), Nikolai Khabibulin (4 games, .833), Kent Simpson (1 game, .714).  So the Hawks goalie numbers are skewed by the terrible performance of their backups. In the future, I may need to use the numbers of the anticipated starting goalie for each team.

Ok, now onto the conference finals predictions. Keep in mind that the third round projections don't take into account how the teams played in the first and second rounds. So, we're assuming that a team's performance  is independent of its performance in rounds 1 and 2.

Montreal Canadiens vs. New York Rangers
Both teams are playing their best hockey right now. Henrik Lundqvist was brilliant toward the end of their series against the Pens. Carey Price was equally good. The Rangers are the better possession team, but will playing two 7 game series, including 4 elimination games, affect them? This should be a great series.
68% chance of a Rangers victory.
32% chance of a Canadiens victory.

Chicago Blackhawks vs. Los Angeles Kings
These were the two best possession teams in the NHL. This is going to be some great hockey. Like the Rangers, the Kings have played two 7 game series. They've faced elimination six times. Will it affect them going forward? The Blackhawks know how to close out a series, so the Kings can't count on making another dramatic comeback.
77% chance of a Kings victory.
23% chance of a Blackhawks victory.

Now, just for fun, I ran the model using the regular season 5v5 save percentages for the four starting goalies for these teams. The Rangers become more modest favorites (60/40) and the Kings also become more modest favorites (62/38). I also ran Cup winning simulations based on both team and starting goalie 5v5 save percentage. The Blackhawks' chances improve more than two-fold using starting goalie numbers.

Updated Stanley Cup probabilities





team 5v5 save %

starting goalie 5v5 save %
Los Angeles Kings
51%
Los Angeles Kings
38%
New York Rangers
31%
New York Rangers
31%
Montreal Canadiens
10%
Chicago Blackhawks
22%
Chicago Blackhawks
8%
Montreal Canadiens
9%









































Thursday, May 8, 2014

Marc-Andre Fleury is exactly who he is


Guys, we need to talk about Marc-Andre Fleury. Specifically, we need to talk about the unrelentingly stupid conversation that’s persisted around Fleury for the better part of the last five years.
Half of Pens fans think Fleury is the greatest thing since Cinnamon Toast Crunch. They put him in the same class as guys like Henrik Lundqvist, Tukka Rask, Zeus and Kurt Vonnegut.
The other half of Pens fans seem to think that Fleury is the biggest goaltending disaster hockey has seen since the Washington Capitals decided to stick a sumo wrestler back there based on the notion that he’d just take up a ton of space.*
These are the people who, whenever the Pens give up a goal, storm Twitter, sports talk radio and presumably their own family dinners with shouts of “TRADE FLEURY.”
To everyone who’s taken one or both of these positions, I offer the following plea: Please stop. You’re coming off as incredibly dumb and you’re making everyone else dumber. Please, shut the fuck up forever.
The fact of the matter is the Marc-Andre Fleury is the league’s most average goalie; nothing more, nothing less. He’s paid like one of the league’s elite goalies, but he’s not that.  
When the Pens signed Fleury to a seven-year, $35 million contract which kicked in for the 2008-09 season, Fleury was nigh on 23 years old. The Penguins had every reason to think that based on his scouting, physical tools and pedigree, he still had time to evolve into the kind of elite goaltender they likely thought they were getting when they drafted him first overall in 2003.

Now, the Flower is 29 years old and a fully known quantity. All meaningful numbers bare this out. We could come at you all day with charts, graphs and numbers which demonstrate that Fleury oscillates between slightly below average and slightly above average, but let’s just use this one, which charts goalies who’ve played at least 350 games between 2005 and today.
What does this tell us?
One, that Fleury is a completely mediocre goalie. He’s not a world-beater, but he’s not actively hurting you. He’s a guy with whom you can win, so long as your defense is acceptable.
The other thing is that Thomas Vokun, who’s missed the entire season after suffering a blood clot in his brain, has the highest save percentage of any goalie since the 2005 lockout. That’s madness. The Penguins got him for nothing, but he’s still been better than almost any other goalie in the league.
Nobody should be opposed to the Pens bringing Fleury back if he’s willing to take a salary more commensurate with his ability, but if he’s going to command any kind of a raise, the Pens simply shouldn't bother.
This argument ends now. Fleury isn’t god’s gift to goaltending, nor is he its Josef Stalin. Everyone shut the hell up.



*Author’s note: this never actually happened, but was suggested during an episode of “The West Wing.”

Thursday, May 1, 2014

FTC's NHL projections: Round 2

The first round of the NHL playoffs has come and gone. Let's look at a few interesting things from our round one projections:

The Pens got it done. They were pretty mediocre through four games, but really turned it up in Games 5 and 6. They controlled the puck, and they need to keep playing that way.

We really nailed it with the Minnesota-Colorado projection. The Avalance simply weren't as good as their results showed.

So far, the model seems to be performing pretty well overall.

The model may be underestimating both the Blackhawks and the Pens. The Hawks were the 2nd best possession team all year, but the model is holding their goaltending and penalty killing against them, and only gave them a 34% chance to beat the Blues. Corey Crawford wasn't great, but the Hawks killed 27/29 penalties against the Blues. If Crawford can hold it together and the Hawks continue to do well on special teams, they'll be hard to beat. As for the Pens, they may finally have enough healthy forwards so that their 3rd and 4th lines aren't possession sinkholes like they were in the regular season. That's a big deal. Also, Kris Letang started to play much better toward the end of the series. If the Pens can keep playing how they did in Games 5 and 6, they can do some damage.

Now, onto our second round projections. Keep in mind that the second round projections don't take into account how the teams played in the first round. So, we're assuming that a team's performance in round 2 is independent of its performance in round 1.

Boston Bruins vs. Montreal Canadiens 
It's always great to see another installment of this rivalry. The model loves the Bruins; they're an elite team with great puck possession, great goaltending, and great penalty killing. If the Canadiens are to win this series, Carey Price will have to be amazing, and the Habs will need to win the special teams battle.
77% chance of a Bruins victory.
23% chance of a Canadiens victory.

Pittsburgh Penguins vs. New York Rangers
In my first round preview, I said that the Pens have to play a fundamentally different brand of hockey if they want to go deep into the playoffs. They started to do that at the end of the first round. If they can keep that up, they can get back to the eastern conference finals. The Rangers struggled to close out the Flyers despite out-possessing them throughout. The model likes the Rangers; they were a good possession team all year with good goaltending, but the Pens are finally healthy and could be heating up.
74% chance of a Rangers victory.
26% chance of a Penguins victory.

Chicago Blackhawks vs. Minnesota Wild
The Blackhawks are very good. The Wild were able to squeak by the Avs, but they face an uphill climb against the Hawks.
65% chance of a Blackhawks victory.
35% chance of a Wild victory.

Anaheim Ducks vs. Los Angeles Kings
The Kings came back from 3-0 to beat the Sharks. They probably won't have as much difficulty with the Ducks.
80% chance of a Kings victory.
20% chance of a Ducks victory.

Updated conference winning probabilities



Eastern

Western

Boston Bruins
51%
Los Angeles Kings
64%
New York Rangers
33%
Chicago Blackhawks
19%
Montreal Canadiens
9%
Anaheim Ducks
11%
Pittsburgh Penguins
7%
Minnesota Wild
6%



Updated Stanley Cup probabilities



Los Angeles Kings
37%
Boston Bruins
31%
New York Rangers
17%
Chicago Blackhawks
6%
Montreal Canadiens
3%
Anaheim Ducks
3%
Pittsburgh Penguins
2%
Minnesota Wild
1%