Jump to content

Claimed:Statistically Advanced - Week Two: Toronto Legion


Recommended Posts

Statistically Advanced - Week Two: Toronto Legion


Statistically Advanced is a weekly feature which will give an in-depth look at a VHL team, discussing said team's weaknesses and strengths, using fancy "advanced" stats to predict trends. Each week, I'll analyze a different VHL team. Warning: If you hate numbers, this post will likely not be for you. 


This week will focus on the Toronto Legion.


If you haven't read last week's post, you'll want to. It basically explains what I'm going to be talking about today. And seriously, if you haven't read it, you're a terrible person and I'm extremely offended (kidding). Right then, let's get started on my first ever team analysis!


Wait actually, before we start, you all need to see this. It might be the most adorable thing I've ever seen.







Team Overview

The Legion have been a very strong team this year. As of today, January 13th, they sit in second place in the North American conference, just 2 points behind Quebec CIty Meute. Currently very hot, they're 7-2-1 in the last 10 games.


Second in the league in shots against per game, with 25.9, and the best team in the league in terms of both goals against average with 1.60 (a meh stat for goalies, but usually a solid one for the whole team), and in goals for/against percentage (with a whopping 190.63%). So, it's probably safe to say that the Toronto Legion are one of, if not the, best defensive teams in the entire VHL.


So, it looks like things are all fine and dandy for the Legion. Right? Well, why don't we find out?


Advanced Trends

First off, let me begin by giving you an idea of my sample size. Because the VHL game sheets are long, and I just don't have time to spend a lot of time going through the sheet for every game, I've only calculated the advanced stats for the Legion based on the last 10 games as of January 12th. Not a huge sample, I know, but that should give us at least a bit of an idea. Of course, I was able to calculate PDO for the whole season, and there are certain stats that are available on the league stats page.


Now, over the 10 games I looked at, the Legion were 9-1. That's a wicked hot streak. Seriously, they've been great. Of course, that's with the exception of the one loss, coming at the hands of the (also very strong) HC Dynamo. That game was a little rough. But, just a bump in the road.


Let's get into the stats, and what they mean.


Over the whole season, the Legion have recorded a PDO of 102.7. That's very good. Not entirely unsustainable, but look to see the team come back to Earth a little bit. Goalie Remy LeBeau has a save percentage of .938, which is really good. The shooting percentage is 8.9%, which isn't astronomical. So, it's probably fair to say that the Legion haven't been outrageously lucky this season, but could see a slight drop off in the future.


Now to my absolute favourite stat, Corsi.


First off, something that you may have picked up in reading the articles I linked to last week, I want to mention the two variations of Corsi I'll be employing here.


Picture this. It's the third period, and a team is down by a goal. What are they likely to do, in order to increase their chances of tying the game? That's right, they'll use their offensive players more, and play a more gung-ho, attacking game. So, situations like these tend to distort Corsi stats, since both teams (the one desparately trying to get shots and the one attempting to go into a metaphorical defensive shell) are not playing like they usually would.


To mitigate this effect, we have something called Corsi Close. A game is considered “close” when either:

1. The game is tied, or

2. One team leads by one goal in the 1st or 2nd period.

So, a team's Corsi Close stat is just all their shot attempts for, minus shot attempts against, while the game is in a “close” situation.

Thus, we get a better picture of what the team is like normally, on any given night, such as at the start of the game.


Similarly, we measure a team's Corsi stats for only when the play is 5 on 5, eliminating the factors of powerplays and penalty kills giving teams more chances for shots.


Finally, CF% is basically the ultimate Corsi number. It's just the percentage of the total shot attempts in a game that belonged to one team. (So that's Corsi For / (Corsi For + Corsi Against) )


Okay, enough. Let's get to the actual freaking stats now.


Over our sample, the Toronto Legion have a very solid CF% of 52.5%. On average, the Legion have 5.9 more shot attempts than the other team in a given game (To put this in perspective, the Los Angeles Kings of the NHL, the best possession team in the league, are +12.3).


(For more perspective, here are the top 10 CF% teams in the NHL:)


So that's not bad!


In close situations, the team tends to be even better, posting a CF% of 54.1%. They have a slightly lower Corsi differential with +5.6, however that's mostly due to the fact that the regular Corsi stat measures a whole game, and the close stat tends to measure about 2/3 or so of the game (therefore, less time for shot attempts).


So it's all sunshine and rainbows for the Legion? Well, not quite.


The Legion's powerplay is slightly below league average, at 15.56%, good enough for 6th in the league. However, the team's penalty kill has been absolutely lights out, leading the league by killing 89.74% of their penalties. This appears to have skewed the team's overall Corsi.


In 5v5 situations, the Legion actually have a negative Corsi. A CF% of 49.7%, and an average differential of -0.5. Now, that's not completely awful (i.e., not Toronto Maple Leafs levels of bad), but really not great for a team that's been doing so well.


What does that mean? Well, it could mean several things. It could mean that the team just gets outrageous amounts of shots on the powerplay – which, as I've pointed out, is unlikely, since the Legion don't have huge amounts of success with a man advantage. It could also mean that they completely shut down the other team on the PK. This is much more plausible. Toronto's penalty kill is very strong, so it's very possible that they don't give their opponents anything on the PK, causing their total Corsi Against to be way low.


For the record, the Legion do take more penalties than some of the other top teams, with 9.15 PIM per game (the league average is 9.96, Toronto are 5th in the VHL). So, I guess they have more opportunities than some other teams to shut down the opposing powerplay, and pad their Corsi.


In short, should Legion fans be worried? Well, that remains unclear. The 5v5 CF% is bad, but it's not unsalvageable. If the team can pick up their play with even strength, they'll go far. If not, well, I doubt they'll fall off a cliff, but they might drop a few more points in the standings than perhaps they should.


Other Notes

  • This Legion team has actually had more scoring from its second line, it looks like. Centre Niklas Lindberg has 24 points in 20 games – 10 of them goals. That's more than a point a game, also good for 11th in the VHL. He does all that with a shooting percentage of 11.49% (in the same ballpark as NHL players Phil Kessel, Evgeni Malkin and Marion Hossa). So that's good, but not ridiculous. He could probably keep that up, I think.


  • The Legion actually have 4 players scoring at a clip above a point-a-game (Lindberg, Lebedev, Lehtinen and Kjellberg). Jack Sound is ever so slightly below, with 19 points in 20 games. None of those players have particularly nasty shooting percentages, Lindberg being the only one above 10%. Lebedev and Sound are defencemen, so their points totals are made up mostly of assists, of course.


  • Something I haven't mentioned once yet is faceoff percentage. I should mention it. Holy crap. Jakub Kjellberg has won 61.58% of his faceoffs. SIXTY-ONE PERCENT. Seriously, that is really good. Patrice Bergeron, probably the best faceoff guy in the NHL, usually has a slightly lower winning percentage of 59 or 60%. And while I'm raving about Legion players and faceoffs, Niklas Lindberg also has a nasty faceoff percentage of 57.9%. Jeesus. These guys take the majority of the faceoffs for the team. So, it's fair to say that the Toronto Legion start with the puck a lot. (The team's overall FO% is 55.71%, which, although only 4th in the league, is really good.)


  • Finally, netminder Remy LeBeau has been great. He has a really solid save percentage of .936, and has carried his team in a couple games they may have had no business winning.


Prediction for the Season

The Legion will definitely be an interesting team for the rest of the season. They're currently a bit of a powerhouse, and most of their numbers support them. However, the 5v5 Corsi is a nagging issue, as is the powerplay. I think this team, if they can pick up their game in those two areas, will absolutely be a contender this year. If not, then I wouldn't rule out a bit of a regression. Not overly awful, but perhaps a slightly lower league finish that they could achieve. My pick? I think they'll just edge out the Bears for 2nd in the North American Conference.


That about wraps it up for this week's installment of Statistically Advanced. Next week I'll be looking at the Vasteras Iron Eagles. Because shit, that's a really cool name. 


This was really long post, so I won't be too offended if you don't read all of it (actually, I'll hate you forever but that doesn't matter.) As always, please tell me if I screwed up, or left any glaring holes in the analysis of your team. Please, call me out on it if I say something bullshit.


I'll leave you guys with this awesome Vine of Claude Giroux chirping James van Riemsdyk:





Link to comment
Share on other sites

Excellent article, extremely well done. I love the idea the Corsi Close, although I'm not sure whether our sim engine designates a difference in team strategy for down-heavy or up-heavy situation situations? That would be a question for those more familiar with the sim engine than I am, and if there's no difference then that's just sample size/randomness there (and I imagine would save you a shit ton of data mining time).


And yeah, having those two excellent centers is a huge help for Toronto. Other than Toronto and Davos, don't believe you'll find a team with two centers with that skill set. Also doesn't hurt Lindberg's scoring that not many teams have two active defensive lines other than Seattle.


Curious for my own stats-minded self (who doesn't know as much about hockey advanced stats as other sports): You have/know where to find NHL Z-scores for the CF%? Would be interesting to see where Toronto's 52.5% sits on the scale.

Link to comment
Share on other sites

Well from my observations the trailing team in non-close situations actually did tend to be getting more chances. I may well be wrong though.

Also, check out extraskater.com, it has all the advanced NHL stats, great site!

Link to comment
Share on other sites

  • Admin

Fyi, faceoff percentage is high for the best centers because there's only a handful of them and they're always superior to other centers in faceoffs. So there's always 4-5 over 60%.

Cowboy, I think there's an option to change strategy when behind and when ahead.

Anyway, awesome. Vasteras will be interesting because everyone thinks they're overachieving

Link to comment
Share on other sites

This topic is now closed to further replies.
  • Recently Browsing   0 members

    • No registered users viewing this page.
  • Create New...