• Moderators, please send me a PM if you are unable to access mod permissions. Thanks, Habsy.

Advanced Stats Thread

13/14, Fenwick Rankings, Points Pace

1 Chicago Blackhawks: 125 pts
2 St. Louis Blues: 125 pts
3 Minnesota Wild: 117 pts
4 Los Angeles Kings: 113 pts
5 Pittsburgh Penguins: 102 pts
6 Vancouver Canucks: 93 pts
7 San Jose Sharks: 121 pts
8 New Jersey Devils: 78 pts
9 Anaheim Ducks: 114 pts
10 Dallas Stars: 98 pts
11 Boston Bruins: 111 pts
12 New York Rangers: 82 pts
13 Detroit Red Wings: 98 pts
14 Tampa Bay Lightning: 115 pts
15 New York Islanders: 74 pts
16 Montréal Canadiens: 86 pts
17 Nashville Predators: 82 pts
18 Colorado Avalanche: 121 pts
19 Florida Panthers: 55 pts
20 Winnipeg Jets: 82 pts
21 Phoenix Coyotes: 121 pts
22 Calgary Flames: 66 pts
23 Columbus Blue Jackets: 70 pts
24 Ottawa Senators: 82 pts
25 Philadelphia Flyers: 69 pts
26 Carolina Hurricanes: 78 pts
27 Washington Capitals: 98 pts
28 Edmonton Oilers: 45 pts
29 Toronto Maple Leafs: 103 pts
30 Buffalo Sabres: 41 pts
 
Jeff Veillette @Jeffler
"Our statistics are proving we're having success where we shouldn't be" -Randy Carlyle
 
James Mírtle @mirtle
Carlyle: "We want more puck possession time. There's been a lot said our statistics proving out we're having success when we shouldn't be."
 
Starting to like this Chris Boyle guy at PPP. (@ChrisBoyle33).

he actually does his homework.

Very interesting article today: How Toronto Scores

Condensed:

Total: 55 gls

Poiwer Play: 16
Penalty Kill: 3
Empty Net: 3
Extra Man: 1
PP just expired: 2

Leaves 30 goals at "pure even strength". These break down like this:

Dump-In Goals: 1
Cycle Goals: 1
Point Shots: 1
Dirty Goals: 1
Throwing at net: 0
Pinching Dmen: 0
Face-Off Goals: 2
Neutral-Zone TOs: 3

Goals off Rush: 21

Considering that we set a very high bar here - that it was only to include goal-scoring rushes that started from the Leafs end, and which had to be quick hitting in the other team's ends, at 5-10 seconds max - what's shocking to me here is that the Leafs still had... 21. That's right kids, twenty-one.

You might want to go back over that, and tote em up again.

My jaw was on the floor. [Edit: Sorry, those were my teeth. Which is sh*t-loads less painful than a jaw. But still embarrassing.]
Click to expand...

and the conclusion:

If you have to, sit down by yourself - as I did - and walk through the types of plays which the list above shows to have been absolutely USELESS when it comes to actually scoring goals this year. And I repeat, these are not words I wanted to hear myself ever say about a Toronto team. They just happen to be true. But to all intents and porpoises:

The Leafs cycle game has failed to create anything. In 20 games.
Nor has getting into the dirty areas.
Our point shots at Even Strength have been basically worthless.
Our d-men have been pinching in vain.
Putting pucks on net has simply been a way to hand the puck back to the other team.
And dumping the puck in has been ridiculously impotent as an offensive method.
Want to make it worse? [Sure you do. You're Leaf fans.]Ok. Get this.

NONE of those 4 rare, non-lightning-strike goals were in any way relevant to the team winning actual games or getting points in the standings. Two came when we were just piling on to our lead in Edmonton, one was from late in the game during bad loss in Columbus, and the final one was from that Colorado loss.

In sum, the Toronto Maple Leafs, through 20 games, have won every single game and gained every single point they have in the standings off goals scored by their Special Teams, and off of Quick Hit offensive plays.

Let's just face the singular fact. The Leafs are a quick-strike team. Vipers, quick strike artists, predators and poachers. Everything else is just window-dressing. A head fake. Pretence.

Because the way they really score is to see an opening, just a glimpse of an opening, a chance to pounce on that mistake, to get the puck and just one step, hell, one half-step on a defender.

And after that - it's 5 seconds and a puff of smoke. Goal, Toronto Maple Leafs.

No matter if you like physical play (as I do), the Leafs - while laying out some nice hits this year - have plain and simple NOT been a team which creates many goals by using their body in a hard-checking, physical way. Not in scoring, not very often in retrieving the puck following shoot-ins or on rebounds, and rarely even in capturing the puck in the first place back in their own end.

jtdVKws_medium.png
Click to expand...


and some bonus data:

Alright. A few more surprises. Things that, if you were like me [which, if you were lucky, you aren't] you were kindof sortof feeling were maaaaaybe the case. But looked at up close, are kindof mind-blowing.

* 15 of the 21 "Rush Goals" came from the Top 2 lines (comprising 7 individuals.) Two more came from Dion jumping up. And 4 from the Bottom lines - 2 were by Smith, 1 by Kulie and 1 by Leivo. But get this. Rush goals were the ONLY kind of goals scored by the Bottomline players. A long lead pass to Smith; Leivo on a 2-on-1; Kulie off on a give-and-go; etc. In short, we've all noted for some time that our Bottom lines don't score much. But the surprise is.... when they do, it's off a Rush. i.e Not in the style or manner we expect from Bottom 6 players.

* Almost no goals at all were scored from more than 30 feet out. Our furthest out shots were basically Kessel's snipe from the goal-line, Lupul's wrister against Boston from the dot, and Smith's nice shot that got deflected against Buffalo, which came from the top of the circle. But that's about it. An incredible 27 of 30 of the Leafs goals here were scored from the crease or the slot or very close range. This is also pretty stunning.
Click to expand...

and:

1. The 25 Special Teams Goals. Even on the special teams, we see basically the same patterns, the same things working - and not working.

For example, of the 25 goals scored in non-Even-Strength situations, a full 10 have come off the Rush anyway. The 3 short-handed goals and the 3 Empty Netters were pretty obviously going to come off the Rush - but 4 of the PP goals have actually come off the rush as well. Like Kessel circling back into his own end and winding up against Jersey the other night - a PP goal off a Rush.

Once in the zone and having established our PP, another 5 goals have come off of slick, cross-crease, set piece plays. Which is nice, zone time and all, and not off the Rush. But they're certainly not jamming-the-crease or "just putting the puck on net" goals. These are goals off about 5 passes and then Kessel or JVR or Lupul or Kadri shows up on your doorstep with a "thank you very much" smile on his face.

And the remaining 10? Well, 3 come from pure open-ice shots and snipes.... and a final 7 come from tips and rebounds off shots from the D-men. Which is great, that at least on the PP, our D-men are shooting and getting some end-results.
And which is also interesting because it means that, when on the powerplay, either our D-men are better able to get useful shots through; or, our forwards are better able to get positioned so they can make tips or pounce on any rebounds.

So, in sum, even our special teams scoring is heavily-reliant on the Leafs rush game, with support from some strong in-zone passing plays, and improved point-shooting combined with shot-tipping and pouncing on rebounds.

2. A Question. Can you think of any team in recent memory - say, the last 5 or 6 years - that has been nearly so concentrated on just one or two avenues to scoring success?
I can't think of any team that has gotten even close to 93% of its scoring from Quick Hit attacks and Special Teams. And just 7% from all the forechecking and the cycling and the pounding the crease and the point shots. Have there been any teams which were as utterly reliant on some other singular method, such as, perhaps, Boston off the cycle? I can't think of any.

Which would make the Leafs, once again, a team that plays using extreme methods. We've already seen the extreme line-matching, the use of two fighters, the extremely low shot totals form our Bottom lines, the mammoth disparities in the levels of competition and the zone starts doled out by Carlyle. But this is an extreme use of a particular method, not just an extreme use of an individual player. And it's not to achieve a marginal end like fighting, this is how the Leafs score goals.

3. What should the Leafs do, moving forward:

A. Build On Their Strengths. Fact the facts. That the Leafs have an extraordinary group of players with an extraordinary skillset, and that when you have a baseball team of home-run hitters, or a BB team that lives off the fast-break, or a football team that can throw bombs all day long, you might just as well work on enhancing and refining your strengths, and make sure that you have them improved to such a degree that other teams can't adjust and shut you down.

B. Diversify Their Attack. Work to add some new arrows to the quiver, diversify a bit, balance things out a bit. You know, work on developing a cycle-game to try and crank up at least some zone time. Work on making plays off the face-offs more. Work on getting shots through from the point, and having guys hold their ground during 5v5 shots on net, not just during the powerplay. That sort of thing.

C. Break It Up. Recognize that the team is over-built in one direction, over-laden with a particular type of Quick Hit, high-end offensive talent both upfront and at the back, accept that you can't force round pegs into square holes, and that therefore, we probably need to make a trade or two. Say, unloading a Gardiner or a Lupul or a Franson to pick up a different kind of player.

Your call.
Click to expand...
 
Wow. Fascinating. Would like to see other teams #s just as a reference point. I don't know how I feel about those stats...
 
Toronto Maple Leafs are not afraid of advanced statistics

TORONTO — Earlier this month, Toronto Maple Leafs general manager Dave Nonis spoke at a sports management conference in the city, and he offered his views on the evolving role advanced statistics play across the National Hockey League: “The biggest thing we use is going to watch a player play.”

They had been looking for reliable applications, he said, but remained convinced the best ways were the old ways. As relayed by Sportsnet, Nonis seems less likely to trust what is on a spreadsheet than what is seen on a sheet of ice: “I haven’t seen anything that’s going to stop that from being the primary source of our decisions.”

In the days since he made those remarks, the Leafs have continued to win games; not as many as they did to start the season, but enough to stay out of trouble for the time being, and enough to stay ahead of the numbers.

The team concedes a lot of shots on net, more than it takes at the other end of the ice, and not by a little. In terms of shot differential, Toronto was second worst in the league as of Wednesday, allowing an average of 9.4 more shots on net than what goes the other way.

Only the Buffalo Sabres (-10.5) were worse.

Many of the teams on that end of the spectrum are not headed to the playoffs, at least not without a reversal of fortune. The fact the Leafs have been able to grab hold of a playoff spot, as they did in a lockout-shortened season last year, has made them a case study in the eyes of those who advocate for advanced analytics.

Being outshot is only one facet, with the general theory that taking shots is an indicator of puck possession, and that puck possession has a strong correlation with winning. The shot clock has not been a friend of the Leafs.

As it was again on Tuesday night, when the Leafs rolled to a 5-2 win over the New York Islanders despite being badly outshot. Toronto allowed 37 shots, while taking 24.

“Our coaches track scoring chances, not shots,” Leafs winger Joffrey Lupul said. “A guy crosses centre ice and shoots the puck at the net, it’s a shot.

“After the game, when they put up the scoring chances, if there’s ever a big disparity in that, it’s going to be an issue with the coaches and with the players,” he said. “But again, if they get 40 shots and five scoring chances, that doesn’t really matter to us.”

Scoring chances, though, are subjective: “Some games we argue with their stat-taking,” Lupul said with a smile.

What has helped the Leafs so far is the work they have received from the men who play in their nets. James Reimer and Jonathan Bernier are among the statistical elite, second (.936) only to the Boston Bruins (.941) in save percentage as of Wednesday, despite the higher volume of work.

“The goal is to win the game, however it happens,” Lupul said. “We’d love to have a game where they only have to face 15 shots. But right now, that’s not happening for us.”

Bernier stopped 35 of the 37 shots he faced against the Islanders on Tuesday, including 16 of the 17 he faced in the third period.

“It’s tough to expect them to do that every night,” Lupul said. “There’s going to be a game where it goes the other way, where Bernie or Reims has a bad game, and maybe we’ve got to score six goals to win him the game. That’s what being part of a team is about, really.”

Do the goaltenders ever notice the shot total during games?

“I think what it comes down to is chances, that’s the biggest thing,” Reimer said. “The shot total sometimes gives you an indication of the chances. But in a game like [Tuesday] night, I think the chances were pretty even.”

Toronto has outshot the opposition in only three of its 21 games this season.

“We want more puck possession time,” Leafs coach Randy Carlyle said earlier in the week. “There’s been a lot said of our statistics proving out that we’re having success when we shouldn’t be.”
Click to expand...

http://sports.nationalpost.com/2013/11/20/toronto-maple-leafs-are-not-afraid-of-advanced-statistics/
 
The thing no one seems to be mentioning in this is that this is a young team that has had lots of turnover. This offseason they added 8 new players (Clarkson, Bolland, Raymond, Ashton, Gardiner, Rielly, Ranger, Bernier) which is almost half of the roster. Not to mention that JVR, McClement, Fraser and McClaren were all new last year. So that is over half the team that has only been with the Leafs for ~60 games. This is exacerbated by the injuries and suspensions which have seen another 7 players (Broll, Leivo, Bodie, Holland, Smith, Devane, and Abbott) make their Leafs debut this season. So that is 19 guys who have made their Leafs debut within the last 60 odd games.

No shit the team isn't perfect, and has things to work on. Having a talented team on paper is one thing. Getting that talent to come together as good team takes a bit of work.

All this talk about regression. What about improvement that comes with stability and experience.
 
Last edited:
I put together some numbers a few nights ago and then promptly forgot all about it. This is a bit of a rush job, and I wanted to do more stats than just this, but this is essentially what I wanted to do anyways.

Basically, since guys like Mirtle have defended possession stats' inability to predict the following year's standings very well (especially with roster changeover) by arguing that it's more valuable as an "in-season" indicator, I just wanted to run a little test of those numbers up against a more basic over/under achiever indicator like goal differential, and just see which one does a better job of predicting the risers and fallers by the end of the year.

So I just ranked all the teams according to Pts%, GoalDiff, and Fenwick Close, and picked out the largest discrepancies.


(and of course the overachievers are the teams that we should expect to fall in the standings, and the underachievers to rise).

Goal Differential:

Overachievers: 1.PHX +9, 2.NSH +7, 3.DET +4, 3.CAR +4, 3.MIN +4
Underachievers: 1.MTL -8, 2.PIT -7, 3.TOR -5, 4.BOS -4, 4.OTT -4

Fenwick Close:

Overachievers: 1.TOR +18, 1.PHX +18, 3.COL +16, 4.WSH +12, 5.TBL +5
Underachievers: 1.NYI -12, 2.NYR -11, 2.NJD -11, 4.VAN -10, 5.CAL/PIT -8

Notice that Phoenix is at the top of both overachiever lists, so they won't be a useful case study (unless somehow they defy both). None of the other overachievers appear on both lists.

Pittsburgh appears on both underachiever lists, but rank higher on the goal differential list than the Fenwick list, so they still might be a useful case study.

Of course, there's only one team that appears on opposite lists for each stat, and that is none other than your Toronto Maple Leafs. Of course. They just have to be as unique as possible. Somehow goal differential actually predicts they'll improve going forward, while Fenwick predicts they'll crash harder than any team in hockey. So, as always, the Leafs will be the best case study of all.


It'll be interesting to see which number does a better job of predicting future regression in the standings this season.

And this could easily end up showing that the advanced stats are better.
 
I put together some numbers a few nights ago and then promptly forgot all about it. This is a bit of a rush job, and I wanted to do more stats than just this, but this is essentially what I wanted to do anyways.

Basically, since guys like Mirtle have defended possession stats' inability to predict the following year's standings very well (especially with roster changeover) by arguing that it's more valuable as an "in-season" indicator, I just wanted to run a little test of those numbers up against a more basic over/under achiever indicator like goal differential, and just see which one does a better job of predicting the risers and fallers by the end of the year.

So I just ranked all the teams according to Pts%, GoalDiff, and Fenwick Close, and picked out the largest discrepancies.


(and of course the overachievers are the teams that we should expect to fall in the standings, and the underachievers to rise).

Goal Differential:

Overachievers: 1.PHX +9, 2.NSH +7, 3.DET +4, 3.CAR +4, 3.MIN +4
Underachievers: 1.MTL -8, 2.PIT -7, 3.TOR -5, 4.BOS -4, 4.OTT -4

Fenwick Close:

Overachievers: 1.TOR +18, 1.PHX +18, 3.COL +16, 4.WSH +12, 5.TBL +5
Underachievers: 1.NYI -12, 2.NYR -11, 2.NJD -11, 4.VAN -10, 5.CAL/PIT -8

Notice that Phoenix is at the top of both overachiever lists, so they won't be a useful case study (unless somehow they defy both). None of the other overachievers appear on both lists.

Pittsburgh appears on both underachiever lists, but rank higher on the goal differential list than the Fenwick list, so they still might be a useful case study.

Of course, there's only one team that appears on opposite lists for each stat, and that is none other than your Toronto Maple Leafs. Of course. They just have to be as unique as possible. Somehow goal differential actually predicts they'll improve going forward, while Fenwick predicts they'll crash harder than any team in hockey. So, as always, the Leafs will be the best case study of all.


It'll be interesting to see which number does a better job of predicting future regression in the standings this season.

And this could easily end up showing that the advanced stats are better.
 
credit to Mirtle, he's actually trying to figure out what's going on with the leafs' shots instead of just screaming "luck".

http://www.theglobeandmail.com/spor...back-plan-keeps-earning-wins/article15579211/

According to behindthenet.ca, which tracks analytics for the NHL, the Leafs are close to a league average team in terms of puck possession when they are trailing by a goal. In those situations, they create 51.2 per cent of the attempts on net, compared to 53.4 per cent for the average team.

But when they lead by a goal, their possession rating is less than 41 per cent, well off the league average of 47.4 per cent.
Click to expand...

It’s a trend you can see in the Leafs record when you look at how they fare after scoring the first goal in games. Including last season, Toronto has scored the first goal in 47 of their last 71 regular-season games, or a league-leading 66 per cent of the time.

The Leafs, however, have gone on to lose 18 of those 47 games, putting them on the low end in that stat leaguewide.

In contrast, they’re a pretty strong team coming from behind, using offensive weapons like Phil Kessel and Dion Phaneuf to win nearly 50 per cent of the time when they give up the game’s first goal.
Click to expand...

On Saturday, Toronto got its only goal in regulation from David Clarkson at the game’s halfway point and was on the verge of winning 1-0 before Caps star Alex Ovechkin broke through with his 20th of the season with only four minutes left.

Between the two goals, the Leafs were outshot 2-to-1 and didn’t carry a lot of the play, allowing Washington to press for as many potential chances as they could. And, as can be the case in the NHL, the one Ovechkin scored was a weird one, as he hammered a bouncing puck that teammate Mike Green had fluttered into the offensive zone off the post and in.

Once the game was tied, the balance of play was once again basically equal, with both teams trading chances late in regulation and into overtime before Leafs winger Joffrey Lupul scored the shootout winner.
Click to expand...
 
Mirtle brings up some very good points.

When the game is tied or the Leafs are trailing, the Leafs play a very aggressive offensive style. They forecheck and counter very aggressively and their D pinches a lot. But when they lead, especially going into the last period, they play a very conservative trap style, they send almost no forecheckers in and play everyone behind the puck.
 
PlayerToBeNamedLater said:
Mirtle brings up some very good points.

When the game is tied or the Leafs are trailing, the Leafs play a very aggressive offensive style. They forecheck and counter very aggressively and their D pinches a lot. But when they lead, especially going into the last period, they play a very conservative trap style, they send almost no forecheckers in and play everyone behind the puck.
Click to expand...


They also play basically an umbrella style defensively. They're more than pleased to allow 10+ shots per period from outside of the high leverage scoring areas. Looking at the shot charts this season, I could see if it was an isolated phenomena, that you could simply discount it. But when you track it game over game it becomes really clear that the Leafs get as many shots from high leverage ice as their competition, and typically from much higher skill players than the opposition are getting their higher percentage chances from. It's clearly a strategic imperative and not this inherent flaw that some of the media portrays it as. Whether it's a successful strategy over the long term will be determined (though we're going on 70 games worth of data now).
 
Montana said:
HL beat me to it.....but my response was dripping with considerably more snark and sarcasm.
Click to expand...

Heh. I considered going with something along the lines of "No wonder Habs fans hate advanced stats when they can't figure out where to find the traditional ones"... but decided against it.

:couch
 
I think that all the attention on hockey analytics is starting to produce some better investigations and analyses. These seem to be cropping up every day now.

http://puckprediction.com/2013/11/26/is-fenwick-close-a-consistent-predictor-of-winning/

To the growing hockey analytics movement, the notion that puck possession is a critical predictor of winning is largely taken as an article of faith. Earlier this year, Chris Boyle published a fascinating infographic illustrating the linkage between strong possession play and playoff success, and some work I’ve done has shown that Fenwick Close differential (i.e., net shots on goal and missed shots in close-game even-strength situations) is a significant predictor of regular-season performance.

]Nevertheless, there are inconsistencies in the association between possession and winning that bother me. Across the eight seasons of single-game data I compiled to create the Puck Prediction model, shot differential (which correlates strongly with Fenwick Close) was only predictive of 54.7% of game results. To put this in perspective, you’re more likely to pick the winner of a game correctly by simply choosing the home team than you are by choosing the team with the better shot differential entering the game. (A t-test comparing shot-differential predictions to random predictions is statistically significant, but the correct interpretation of this analysis is less “this relationship is really meaningful” than “a sample of 10,000+ games allows us to estimate a weak association with a high degree of precision”.).....
Click to expand...
 
Not very advanced, but I will put this here.

I wanted to look at shot distribution per period.




13/14, Avg. Leafs Shots For/Against Per Period, Overall:


1st Period:
For: 8.6
Against: 11.4

2nd Period:
For: 8.5
Against: 12.4

3rd Period:
For: 8.3
Against: 12.0



13/14, Avg. Leafs Shots For/Against Per Period, In Wins:

1st Period:
For: 9.1
Against: 11.4

2nd Period:
For: 8.1
Against: 12.6

3rd Period:
For: 8.5
Against: 12.4


13/14, Avg. Leafs Shots For/Against Per Period, In Losses:

1st Period:
For: 8.0
Against: 11.5

2nd Period:
For: 9.0
Against: 12.2

3rd Period:
For: 8.0
Against: 11.3
 
You must log in or register to reply here.
Back
Top