I thought this was real cool. At the start of the second period between the Los Angeles Kings and San Jose Sharks, Chris Cuthbert threw it down to reporter Ryan Rishaug for a quick hit where he actually gave some pretty good info as a result of a mid-game interview:

Chris, I spoke with Kings' assistant Bill Ranford at the break I asked 'which part of the stats page makes you the most upset?' He said 'their number of attempts and our lack thereof'. If you consider that San Jose had 15 shots on net, they missed the net eight times and had 14 blocked. That's 37 attempts on net. The Kings had just seven.

There you see the zone time, not even close. Los Angeles with half the offensive zone time that San Jose had. Ranford says 'we are losing every battle in our own end and we are not competing'.

I've screencapped the accompanying graphic, which is some real hipster stuff because you can't find "zone time" anywhere online. It has to be manually tracked. We use Corsi, basically shot attempt differential, as a proxy for zone time, but only because the zone time numbers aren't available. TSN did a great job at showcasing the Sharks' advantage in the first period.

Kudos to Rishaug and TSN there.

Introduction

You probably don't know me, my name is Josh Weissbock and I am a Masters student at the University of Ottawa up in Canada and I study Machine Learning.  Many of the other authors on the Nations Network, who use advanced stats, typically use statistical models to analyze hockey where as I use algorithmic models. What I can do is use Machine Learning techniques which is an application of these same statistical ideas.  Machine Learning is simply the ability for a computer to take a whole bunch of data, learn from it and be able to make decisions. 

Part of my research is combining Sports Analytics and Machine Learning; can we use these algorithms to learn from hockey data and try and make accurate decisions or predictions?  This area actually has very little academic research; in fact I could find no peer reviewed published works on predicting anything in hockey in any field.  Basketball, American football and soccer all have plenty of research in machine learning but none have been found in hockey. 

This post is the first of many on applying Machine Learning to hockey and the NHL.  In this simple experiment to see if it is possible to make simple predictions on who will win the game, using known stats on each team.  The goal is to beat 50%, or a coin flip.  This type of predicting model is valuable both to team managers and for betting.

Data

I've been collecting data daily on every game that happened over the NHL season.  This experiment used games over a ten week period fro February to April 2013, a total of 386 games.  Before each game I collected advanced stats and traditional stats on the teams.  After the game I collected the result of who won, the score and the shots for and against.  This gives me the flexibility to calculate more stats in the future.  

The traditional stats that were collected were: Home/Away Status, Goals For, Goals Against, Goal Differential, Power Play Success Rate, Power Kill Success Rate, Shot Percentage, Save Percentage, Winning Streak and Conference standings.  For the advanced stats I collected the Fenwick Close %, 5-on-5 Goals For/Against ratio and the season PDO of each team.  These stats have been shown to be much more correlated with wins than traditional stats.

Classification & Results

After pre-processing the data I plugged the vectors of game data into the algorithms to let the algorithms do their learning.  The algorithms I used were Neural Networks (NN) because they are good with noisy data; Support Vector machines (SVM) because it does well in machine learning; Naive Bayes (NB) to learn on past data to predict the future; and Decision trees (J48) as they create a human readable output.  I won't go into the math of how these algorithms work but I have linked them to their Wikipedia page which gives further detail. 

Testing of the algorithms was done using 10-fold cross-validation.  The results of the learning with these algorithms are presented in the table below.  I ran the algorithms on just the traditional stats, just the advanced stats and a mixture of both.

	Traditional	Advanced	Mixed
Baseline	50%	50%	50%
SVM	60.10%	55.05%	60.10%
NB	59.07%	54.53%	57.25%
J48	58.03%	52.85%	58.29%
NN	58.68%	55.83%	58.03%

Results of the J48 Decision Tree on traditional data

Conclusion & Future Work

60% accuracy, that's pretty decent for a very simple experiment.  That would probably make profit in the long run if used in betting.  I can't confirm which algorithm is statistically best without running a t-test.  What I find most interesting is that we see this advanced stats on the Nation Network are the best for forecasting future success but here they didn't seem to make any improvement in the forecasting rate.  That doesn't mean they aren't valuable in machine learning, it most likely stems from the fact I only used three of them.  Also interesting is if you look at the decision tree generated ML finds the most value in the teams Goals For, Power Play % and Conference standing in predicting a game winner.  I don't think PDO is that useful in predicting over the long run as it regresses to 100% but it would be much more useful in predicting the short term. 

There is much future work to build off this, first I would start with using a different variation of PDO.  Machine Learning can predict more than just who will win a game, it has been used to predict the scores of games, who will win a tournament, which prospects are likely to work out, and there has been plenty of ML research in baseball on predicting award winners and hall of fame nominees. 

If you are interested in my data for your own work let me know.  If you have any questions or discussion feel free to send me a line.  This experiment is a positive result in making predictions in hockey using Machine Learning and future posts will look at it in more depth.

The Friends with Numbers are back, with a recap of the 1st Round of the NHL Playoffs; we take a look at what interested us, and what the main takeaways were from the opening round. Then, we give our predictions for Round 2, and break down each of the four matchups. 

We would love to hear from you, the listener, for suggestions on future topics of discussion. Rather than us just being in our own little world, we'd like to make this as interactive a show as possible. Feel free to tweet at either Dimitri or Cam.

Click Past the Jump to Listen to the Podcast!

Click here to stream!

Thanks for Listening.

Cam -

Penguins in 7 over the Senators

Bruins in 7 over the Rangers

Blackhawks in 5 over the Red Wings

Sharks in 6 over the Kings.

Dimitri -

Senators in 7 over the Penguins

Bruins in 5 over the Rangers

Blackhawks in 5 over the Red Wings

Kings in 6 over the Sharks.

With Marc-Andre Fleury’s fourth consecutive playoff implosion, the Penguins have finally done the inevitable and decided to use a different goaltender in the post-season. With all due respect to his strong regular season play, the possibility of this was the biggest reason why signing Tomas Vokoun in the summer was a decision that had to be made.

Reputations

There’s a funny thing about goaltender reputations; they often don’t seem to match reality.

For example, things like this were still being said about Fleury as recently as spring 2012:

He’s the most underrated athlete in Pittsburgh, perhaps the NHL’s most underrated player and the best big game goalie in hockey. Fleury is 27, just entering his prime and already has a Stanley Cup ring that’s getting dusty. He is playing the best hockey of his career, finally using his unprecedented athleticism in perfect harmony with a newfound mastery of positioning and poise. For whatever reason, Fleury doesn’t receive the hype of goalies like Henrik Lundqvist and Tim Thomas. But I’d take him against anyone. As long as Fleury is healthy, he always gives the Penguins a chance.

Even at the time it was an odd take. Fleury was legitimately brilliant in 2007-08 when the Penguins lost in the finals; he was just okay in 2008-09 when they won and he was quite bad in the post-season in the two years following. Since then, he had a brutal 2012 playoffs that prompted the Vokoun signing, and he’s getting pulled four games into the series against the Islanders.

Meanwhile, no less a hockey man than current Calgary Flames general manager Jay Feaster had this to say about Tomas Vokoun in October 2009:

While we had goaltending problems in Tampa at the time, we were not interested in Vokoun. Our pro scouts were not sold on Vokoun’s ability to win a championship and thrive under big-game pressure and his contract was too rich for us, both in real dollars and available cap space… [Backup goaltender] Clemmensen has been away from the team dealing with personal/family matters recently, which has resulted in Vokoun carrying the load. Not surprisingly, Vokoun has faltered and the Panthers find themselves desperately needing help from their backup goalie yet again. The only way for Florida to climb back into the hunt is for Clemmensen to do for his current squad what he did last year in Newark, and time is of the essence.

That was written after a season where Vokoun managed a 0.926 save percentage. At the time of the writing, he was 1-4 on the season, with his lone win coming in a game where he made 52 saves for Florida against the Blackhawks. He’d finish the year with a 0.925 save percentage. Vokoun won 26 games for the woeful Panthers that year; in only six of those wins did he face less than 30 shots against.

But that’s the thing about goaltenders. A goalie like Vokoun can play brilliantly for a terrible team and be maligned as a loser by an observer judging by the win/loss record; a goaltender like Fleury can win a Stanley Cup behind a team like Pittsburgh despite a pretty middling run and it takes years for the shine to wear off.

Careers

The two charts above show the save percentage and even-strength save percentage of both goaltenders by year since the 2004-05 NHL lockout. In seven of the last eight seasons, including this one, Vokoun has saved a greater percentage of the total shots he has faced; in six of the last eight, again including this one, he’s saved a greater percentage of the shots he has faced at even-strength.

By career save percentage and career playoff save percentage, Vokoun has been the superior goaltender. This is the best opportunity he’s ever had to show that he can backstop a strong team to great things, and while anything can happen in a short series there’s every reason to believe that he’ll be an upgrade on Fleury in net.

It very well might make the difference; unlike Pittsburgh, the Islanders don’t really have another option to replace their struggling starter.

STREAKCRED

Sign up for StreakCred - the new playoff pool game from the Nation Network. For only $20 you can win solid prizes and a portion of the proceeds go to supporting the MS Bike Tour and the Edmonton Down Syndrome Society. Sign up here.

Recently by Jonathan Willis

• Should the Oilers consider keeping Nikolai Khabibulin?
• Which teams did the Oilers' kid line play well against?
• Timeline
• No Yakupov or Brodin among Calder finalists?
• Edmonton Oilers: Shots vs. scoring chances
• Follow Jonathan Willis on Twitter!

Before we begin breaking down individual scouting reports for the various first round hopefuls I decided to survey the entire field from a high level using NHL equivalencies (NHLE). For those unfamiliar, NHLE is a method that corrects for league quality relative to the NHL, allowing us to compare kids across various leagues.

Of course, like all pure scoring stats, NHLE is blind to various external contributing factors, such as team quality, variance and other circumstances like a players role. We'll attempt to fill in the blanks somewhat with various targets, but for now we'll rank them according to their NHLE to at least get a broad feel for who is available.

That said, there's a reason to pay attention to NHLE. Via Willis at the Cult of Hockey, the correlation between NHLE and future NHL point-per-game pace:

Not perfect, but the relationship is pretty clear.

2013 First round Forwards NHLE

Player	League	PPG	Factor	NHLE	birthdate
Jonathan Drouin	QMJHL	2.14	0.29	50.96	Mar.28/95
Nicolas Petan	WHL	1.69	0.30	41.58	Mar.22/95
Nathan MacKinnon	QMJHL	1.70	0.29	40.53	Sept.1/95
Aleksandr Barkov	Sm-Liiga	0.91	0.54	40.10	Sept.2/95
Elias Lindholm	SEL	0.63	0.78	39.98	Dec.2/94
Max Domi	OHL	1.36	0.30	33.44	Mar.2/95
Sean Monahan	OHL	1.34	0.30	33.08	Oct.12/94
Hunter Shinkaruk	WHL	1.34	0.30	33.06	Oct.13/94
Anthony Mantha	OHL	1.33	0.30	32.68	Sept.16/94
Kerby Rychel	OHL	1.28	0.30	31.47	Oct.4/94
Artturi Lehkonen	SM-Liiga	0.67	0.54	29.52	July 4/95
Valentin Zykov	QMJHL	1.15	0.29	27.44	May 15/95
Ryan Hartman	OHL	1.07	0.30	26.36	Sept.20/94
Morgan Klimchuk	WHL	1.06	0.30	25.97	Mar.2/95
Adam Erne	QMJHL	1.06	0.29	25.18	Apr.20/95
Frederik Gauthier	QMJHL	0.97	0.29	23.01	Apr.26/95
Bo Horvat	OHL	0.91	0.30	22.40	Apr.5/95
Curtis Lazar	WHL	0.85	0.30	20.84	Feb.2/95
Alex Wennberg	SWE-1	0.70	0.36	20.54	Sept.22/94
Valeri Nichushkin	KHL	0.33	0.65	17.77	Mar.4/95
Jason Dickinson	OHL	0.71	0.30	17.52	July 4/1995
Andre Burakovsky	SWE-1	0.26	0.36	7.55	Feb.9/95
JT Compher	USNT	1.14	0.00	0.00	Apr.8/95

(The list of players was more or less based on NHLNumbers consensus rankings. Table is sorted by NHLE).

As Willis recently pointed out, Jonathan Drouin is out on an Island all by himself. In fact, his NHLE is the best we've seen out of a prospect from the QMJHL since Sidney Crosby.

Anyways, there is very clearly a top tier featuring Drouin, Nate MacKinnon, Aleksandr Barkov and...Nic Petan? the undersized center scored 120 points for the juggernaut Portland Winterhawks this year, but he doesn't show up until the 20's or 30's on most draft sheets. His size is an issue as is the fact he jumped from about 30-points last season to 120 this year. In addition, it's assumed he's not necessarily the guy driving the bus in Portland. We'll look at that down the road.

Absent is Valery Nichushkin, who is considered by many as a top-5 pick talent wise. He only played a handful of games in the KHL this year, no doubt in a limited role so his output probably doesn't do his talent level justice.

Elias Lindholm can perhaps be included in that top group before a steep drop off to the next tier, featuring guys lie Max Domi, Sean Monahan, Hunter Shinkaruk, Anthony Mantha and Kerby Rychel. JT Compher goes without an NHLE because we don't have a translation factor for the USNT currently.

Things begin to fall away rapidly once we're outside the 30-something NHLE's. At that point in the draft, you're hoping a guy takes a really big step forward next year or will eventually be capable of filling a support type role in the NHL someday. Most guys below the 30 NHLE mark don't end up scoring a lot of points in the NHL down the line unless there's some rapid development down the line (particularly in their draft+1 season).

2013 Defense first round NHLE

Player	League	PPG	Factor	NHLE	Birthdate
Seth Jones	WHL	0.92	0.30	22.58	Oct.3/94
Ryan Pulock	WHL	0.74	0.30	18.15	Oct/6/94
Josh Morrissey	WHL	0.67	0.30	16.52	Mar.28/95
Darnell Nurse	OHL	0.60	0.30	14.83	Feb.4/95
Rasmus Ristolainen	Sm-Liiga	0.29	0.54	12.77	Oct.24/94
Nikita Zadorov	OHL	0.40	0.30	9.76	Apr.16/95
Robert Hagg	SEL	0.04	0.78	2.37	Feb.8/95
Steve Santini	USNT	0.20	0.00	0.00	Mar.7/95

Obviously it's difficult to grade defenders by output alone, but it's worth a look anyways.

Not a big first round for blueliners, at least not in terms of high-end producers. Seth Jones is the only guy to crack the 20 NHLE threshold. Darnell Nurse gets a lot of attention on a few draft lists as a probably the second best defender available, but he's well back of Jones by this metric. Of course, he also didn't play on the Portland Winterhawks, so there's that to consider.

Other Stuff

I included birthdates for interests sake since some guys are almost a year older than their peers. Sean Monahan and Hunter Shinkaruk, for instance, are almost a full 11 months older than guys like Alex Barkov and Nate McKinnon, which makes the younger guys particular feats all the more impressive. A year-gap isn't all that important once guys are pros, but when comparing kids maturing in various leagues, it can make a very real difference.

StreakCred!

Alright, so you probably know about Streakcred by now. You should also be playing it. If not, here's why you should:

It's fun and easy...

Every night you pick the winners from all the games. Or some of the games. Or every other night. Or once a week. And because it's about building a streak and not cumulative points like many other fantasy games, you can more or less start at any time and become a leader (assuming you can pick winners, unlike me).

Pick only the games you're confident in. Or pick all of them. Just build a winning streak.

The prizes are dope

Seriously. The grand prize is a trip for two to Germany for Octoberfest, including a $1k prepaid debit card/travel voucher. Holy crap.

Second prize is a big ol 52" flatscreen TV from Samsung.

Third place gets an iPad4s. There's also Amazon gift cards and prepaid Visa credit cards up for grabs.

So lots and lots of great swag.

Helps out Worthy Charities

Some of the proceeds from Streakcred will go to the MS Society and Edmonton Down Syndrome Society.

So get 'er going Nation. It's a great game with great prizes and will help out great causes as well.

The leader in streakcred - Rmo2000 - has now crashed and burned and his streak is over at 8 wins in a row. In fact the entire top 20 players in the game have all lost their streaks and everyone is starting again from zero. This is the time for the second wave of people looking to play to get involved.

We need you Nation, team entries are lagging way behind what we thought they would be by now. Keep reading after the jump if you want to see your ol' pal Wanye beg you to support deserving charities.

At the moment we are currently up shit creek in terms of entries. A few hundred entries in the can, we need several hundred more folks to sign up ASAP to square off prizing and be able to make donations to the MS Bike Tour and the Edmonton Down Syndrome Society.

Maybe it's because the playoffs are such a foreign concept here in Edmonton that we can't wrap our minds around what it would be like to participate in a playoff game of any sort. Maybe it's because the game is new and learning is tough.

Either way this game is still wide open for the winning and we know you are smart enough to take it down.

Sign up for streakcred here it's only $20.

LOOK AT YOU - NOW LOOK AT US

We are currently helping a buddy move back to Edmonton from Chicago. Spending almost 3 straight days in a truck with the boys has led to many arguments and fist fights and our collective inability to win at this game has been at the centre of the excitement.  

The Road Trip Leaderboard currently looks like this:

1. @wanyegretz: Team Name: The Ministry of Squee Best streak: 3

2. @thesquireyeg: Team Name: Harry Paratestes Best streak: 3

3. @17samb: Team Name: Samwise17 Best streak: 2

Come on. The playoffs must be rigged or something. Three self proclaimed cool guys like us can't even crack three wins in a row. You can probably do way better than that can't you?

But sitting on the sidelines won't prove that you are smarter than someone who asked a security guard in Chicago when the NBA playoffs start.* We need you to get involved.

Think of what you can win! First place is a trip for two to Oktoberfest in Germany. Say what now? An airplane ride and booze and hotels and a G in pre-paid cash? Out of sight. And this time around entries are open to anyone anywhere in the world and your prizes will work just fine.

Look at all the goodies up for grabs here.

BUT HOW DO I PLAY?

StreakCred is straight forward and is designed so that you can join up at any time. There are still tons of games left in these here 2013 playoffs and it is still anyone's game. 

All you do is pick any team to win in any game on any given day. You can pick as many games as often or as rarely as you like. Don't feel good about LA and Saint Lou? Don't even give them the time of day and skip the entire circus. If your streak goes the longest you win. If you have the most wins in total you win too. It's a win-win-win situation really.

Sign up for streakcred here if you like winning.

REGULAR SEASON WINNERS HAVE BEEN CONTACTED

Speaking of paying out like a slot machine that just came up bells all of the Regular Season winners are going to be contacted by the fine folks at Nation HQ next week. First and foremost let's all give a hearty congrats to Bernie Nicholls' Blue Skate Blades who won the first place of a $5,000 shopping spree at United Cycle.

Think about that for a moment.

That's like 11 autographed Oilers jersies, 500 ref whistles or 2,500 sets of skate laces. Anyway you slice it ol Bernie Nicholls himself is probably envious of the riches his name sake has just won using his immense brain pan.

Our ol pal on the twitterz Uncle Baresnake and The Rippers locked down second place finishing a mere 9 points back. This will get him a $1,000 prepaid visa card for his troubles. This could be used for anything really - internet dating credits for a beloved household pet, a remote control yogurt, some sort of singing clock perhaps. 

All in all we are sending out over $10,000 to several dozen nice folks in the coming days. A massive thank you to everyone who played and an even massiver thank you goes out to United Cycle, LubeX, the Pint and Oodle Noodle. We were able to raise over $3,800 for the MS Bike Tour and the Inner City Children's Fund and now we are looking to send out even more.

*True story, the guy looked at us with such scorn we almost died.

The Friends with Numbers are back, and are gearing up for the playoffs. In this edition of the podcast we break down all 8 of the 1st round playoff matchups, and give our predictions for each.

We would love to hear from you, the listener, for suggestions on future topics of discussion. Rather than us just being in our own little world, we'd like to make this as interactive a show as possible. Feel free to tweet at either Dimitri or Cam.

Click Past the Jump to Listen to the Podcast!

Click here to stream!

Thanks for Listening.

Cam -

Penguins in 6 over the Islanders

Canadiens in 7 over the Senators

Rangers in 6 over the Capitals

Bruins in 5 over the Maple Leafs

Blackhawks in 4 over the Wild

Red Wings in 6 over the Ducks

Canucks in 6 over the Sharks

Kings in 5 over the Blues

Dimitri -

Penguins in 5 over the Islanders

Senators in 7 over the Canadiens

Rangers in 6 over the Capitals

Bruins in 4 over the Maple Leafs 

Blackhawks in 4 over the Wild

Red Wings in 6 over the Ducks

Canucks in 7 over the Sharks

Kings in 5 over the Blues

Have you received positive reviews at work lately? Perhaps a supervisor or boss commended you for "excellent use of time around the office" or a "focus on efficiency?" Well throw that attitude out the window friends. It's time to waste some time during the playoffs. Some serious time. I'm talkin' bout StreakCred - the Nation Network Playoff Game.

Curious? Of course you are. Who wants to climb the corporate ladder anyway? AMIRITE?*

WHATS STREAKCRED?

Since 2009 the Nation Network has been operating hockey drafts for some of Alberta's leading pub and restaurant chains - running 7 drafts and processing over 100,000 team entries. In January 2013 we opened this game up to a pay-to-play online version with sweet prizes and we raised over $3,800 for Edmonton Charities.

But something was missing. Buddies who played the game reported back "yeah Wanye it was fun to make my team. But now all I do is sit back and watch Brooks Laich be injured and week by week the thing gets less exciting**"

So we went back to the drawing board and the boys at Nation HQ have been working feverishly to develop StreakCred. It's a playoff picking game where you select the winner in as many or as few games per night for the entire 2013 playoffs.

The longest streak of correct games picked in a row wins prizes. Big prizes. And there are additional prizes for correctly predicting the most correct games as well. 

We tested it out with the boys the last week of the season and we already curse the very air around us for our stupidity. It's SUPER fun.

WHY YOU SHOULD PLAY STREAKCRED

We can hear people now. "But Wanye, why should I play StreakCred? I have a perfectly good Sega Genesis at home I rarely use. If anything I should play that thing."

Wrong. There are two big reasons - other than "having fun" and "avoiding real work" - that should tip the scales and have you running to the StreakCred website.

PRIZES

The prizes you can win are awesome. They are global prizes first of all meaning we can fulfill them anywhere around the world and so entries are open to the entirety of planet Earth. Secondly, the prizes themselves are sick. How does a first place prize of a trip for two to Oktoberfest in Germany sound? I KNOW RITE? How about a "big screen tele-vision?"

And we have lots more including piles of Visa Prepaid Credit Cards, iTunes Cards and all sorts of other goodies that we will confirm once the first week of entries have closed. The aim of the game is to ship as many prizes as we can all around this blue marble we call home.

CHARITY

As part of our Nation Network annual goal of donating $200,000 for charities annually, we are expecting to donate a substantial sum of money to 2 deserving organizations from the 2013 Playoff StreakCred Game.

In this instance it is the MS Bike Tour and the Edmonton Down Syndrome Society. Both are excellent causes who need love and the idea of hockey fans around the world supporting this Edmonton cause is sweet.

Super sweet.

IM SOLD WHAT DO I DO NEXT?

Kick on over to the StreakCred website, fork over $20 of your hard earned dollars and get ready to Facebook your Grade 8 girlfriend who told you that you weren't cool enough. The cool train is about to depart.

Entries don't close off till the start of the Finals, but the sooner you join the sooner you can start winning. We are already leading despite the fact the playoffs haven't started yet.***

*I am indeed "rite"

** Game experience as reported by an anonymous player we will call W. Gretz. No wait, WG. Yeah, that works.

*** Totally.

PDO doesn't stand for anything, but that doesn't mean we can't learn anything from adding up the overall shooting and save percentages for a team at even strength. A layman's explanation for 'PDO' and why we use it can be found here over at the Backhand Shelf. Basically, if a team is playing with a PDO number way higher than 1.000, they're producing above their expected output. If a team is playing with a PDO number below 1.000, they're producing below their expected output. Over the course of a long season, the number will generally correct itself.

For any daily updates you may be so inclined to find, Hockey Analysis compiles these numbers. behindthenet.ca has a page that offers a team's shooting percentage and a team's save percentage numbers, and we'll use those for weekly rankings here at NHLNumbers. Shooting percentage is the 17th column from the left on BTN's team shots page—the first one to say SPCT. It's cousin, team save percentage, is three columns to the right also saying SPCT. The team shooting percentage needs to be subtracted from 1000 to get the actual number.

I've also included team Fenwick Close % numbers, pulled from behindthenet.ca as well. Treat it as a proxy for the amount of time each team spends with the puck. Any team with a rate over 50% is pretty good and should make the playoffs in an 82-game season unless something goes horribly wrong.    

Here are the team PDO standings through games played Sunday night:  

TEAM	Fenwick Close %	Team Shot %	Team Save %	PDO
Toronto	43.80%	11.00%	0.922	1.032
Pittsburgh	50.28%	9.80%	0.929	1.027
Anaheim	47.56%	9.80%	0.926	1.024
Chicago	56.12%	9.40%	0.925	1.019
Washington	47.24%	9.00%	0.926	1.016
Tampa Bay	44.75%	10.50%	0.910	1.015
Vancouver	51.85%	8.90%	0.926	1.015
Lumbus	44.86%	9.00%	0.924	1.014
Dallas	47.95%	10.00%	0.912	1.012
Boston	54.27%	8.00%	0.928	1.008
Montreal	53.35%	8.90%	0.917	1.006
Buffalo	43.44%	8.40%	0.921	1.005
NY Rangers	53.25%	7.40%	0.930	1.004
Phoenix	50.00%	7.90%	0.925	1.004
Ottawa	51.92%	6.70%	0.935	1.002
Edmonton	44.35%	7.70%	0.921	0.998
Detroit	52.87%	6.90%	0.928	0.997
Winnipeg	50.19%	8.80%	0.908	0.996
Minnesota	47.85%	8.20%	0.913	0.995
Nashville	45.81%	7.90%	0.916	0.995
San Jose	52.90%	6.70%	0.928	0.995
Los Angeles	58.20%	8.00%	0.913	0.993
St. Louis	53.75%	7.80%	0.912	0.990
NY Islanders	52.07%	8.60%	0.902	0.988
Carolina	51.60%	7.60%	0.908	0.984
Philadelphia	48.58%	8.10%	0.899	0.980
Colorado	47.09%	7.30%	0.906	0.979
New Jersey	54.96%	6.90%	0.906	0.975
Calgary	48.94%	8.50%	0.889	0.974
Florida	49.58%	6.50%	0.899	0.964

Last week's numbers found here.

Thoughts:

• It's all about #Lumbus this week. Nobody is pulling for #Lumbus more than me. Their shot up the PDO standings is nearly identical to their week-to-week change in points percentage (lucky I graph everything now):

• On February 25, a .981 PDO and a 6-12-4 record. It may have been curtains for them there. Their team save percentage, mostly Steve Mason at that point, was a mere .900! Sergei Bobrovsky came in and they've gone 16-5-3 since. #Lumbus are proof that even a 48-game season can be long. It may be the right length for them. With just a 44.86% Fenwick Close percentage, #Lumbus's record is an illusion. But it's still fun as hell.
   
• Fun fact: Sergei Bobrovsky was once a Philadelphia Flyer.
   
• Last week we had seven teams outside the bounds we can expect by chance of about 1.025 and .975. This week we're down to four, so math is fightin' really hard to win this war against statistical anomalies Calgary and Anaheim. The teams' respective hot and cold streaks are helping out there.
   
• The Toronto Maple Leafs make the playoffs and their PDO hardly moved an inch all season long. That's also fun to watch because other than their coach's insistence on having two goons in the lineup any given night, they have a lot of players easy to root for like James Reimer and Phil Kessel and Nazem Kadri and a cast of players virtually written off before the start of the season. They're going to make the playoffs and their fans are going to have so much fun and be so obnoxious and it will make the public broadcaster tonnes of money so they can write another sitcom about non-white families in Saskatchewan, I guess.
   
• Scary thought: the New York Islanders are in 6th place and their PDO is just .988. I wrote some words about that team here. They're also fun to watch. #Lumbus-East, maybe, but without the cute Twitter account.
   
• San Jose, St. Louis and Los Angeles are grouped neatly together in the PDO standings. With their good shot differential numbers and lower-than-expected records this season, one of the three is going to surprise in the playoffs.
   
• The New York Rangers could be the team in the Eastern Conference to do that. They've been out-performing their record all season and have started to score some goals. Low shot percentage all year...
   
• Lou Lamoriello TP'd the Mount Olympus of the Hockey Gods. Their Corsi Tied is 54.9%. They'd be the team with the highest Corsi Tied rate to ever miss the postseason, beating out the record set by the 52.9% 2010 St. Louis Blues.
   
• Ergo, time to hang 'em up, Marty.

I'm hearing it everywhere this season, well, mostly on Coach's Corner and on the comment section of my website. "The Leafs are winning because they're tougher and fighting more."

In fact, the New York Rangers lost their fighter Brandon Prust in the offseason. They've gone from 1st in the league in fighting majors to 24th (through Saturday, according to HockeyFights.net) and have gone from Presidents' Trophy contenders to bubble team.

Prust went from New York to Montreal in the offseason. Montreal went from 22nd in fighting majors last season to seventh, and have since gone from a lottery team to a team competing for the title in the tough Northeast Division with the likes of the Boston Bruins! The other team in that division, Toronto, lead the league in fights and are going to make the playoffs for the first time in nine years!

Clearly, fighting means something right?

Well, no, not really.

I went back and looked at all the seasons since 2001 when the league expanded to 30 teams. What I did was look at see whether drastic changes in the number of fights a team had really changed much in the way of win totals.

The standard deviation for change in fighting majors in a season is between 17 and 18. I checked the average win total for teams that had eclipsed that particular standard deviation, plus or minus, from year-to-year and checked the average win total:

Standard Deviations	# of teams	Average Wins Increase
2	3	0.3
1	33	0.6
-1	34	2.5
-2	11	6.7

Pretty standard stuff.

The main takeaway is that teams that increased by more than 35 fights (or two standard deviations) got about a third extra win. Those that decreased their fight total by more than 35, however, increased their win total by 6.7.

Of course, not *every* team wins more games year-to-year. Teams that didn't increase or decrease their fight total by a standard deviation tend to lose a little more (about 1.3 games) so I think the general idea is that a general shift in philosophy is positive, but it's more positive if you take a progressive approach.

The other thing isn't necessarily that fighting less equals wins. Fights generally come from replacement-level players. The 2006 Buffalo Sabres and New York Rangers, respectively, came out of the last lockout replacing thugs like Eric Boulton and Chris Simon with skilled players Tim Connolly, Thomas Vanek, Petr Prucha and Martin Straka.

Still, it's important to note that an emphasis on toughness, while it worked for the Maple Leafs this season, is not a universal truth, no matter what Don Cherry says. The Leafs relied on good goaltending, good shooting luck and a Hart Trophy nomination-calibre season from Phillip J. Kessel. Given that teams in the past that have historically increased their fight totals by a wide margin haven't done better than those that took a more progressive turn, I wouldn't be so quick to say that the fortunes of the Maple Leafs, Canadiens and Rangers indicates that more toughness is good.