Drance Numbers: The Official PITB Advanced Stat Glossary

Drance Numbers is the silly research wing of PITB. While Messrs. Wagner and Mooney blog nationally and solve mysteries, Drance Numbers will look into the minutiae of quantifiable NHL data and bore you with it every Friday. Today, Drance explains all those made-up words he uses defines some terms.

***

CORSI

What does it tell us?

The simplest way to think about a Corsi number is as an expanded version of your traditional plus/minus numbers. Plus/minus is a limited — and I’d say “totally worthless” — statistic in that it only tracks even-strength goals for and against. Corsi counts goals for and against as well, but includes a greater number of events too, such as: all shots for and against, all missed shots for and against and all blocked shots for and against.

Why is that important?

The league leaders in plus/minus will vary widely from year to year and rarely do they reflect a skater’s overall two-way ability (Marek Malik, anyone?). So how useful can it really be as a metric for judging an individual or a teams’ two-way play?

Corsi is a more stable number, and because it tracks most on-ice events, it’s used as a proxy for puck possession. If a skater is able to limit the amount of time spent in his own team’s zone while maximizing the amount of time spent generating shots at the other end of the rink, he’s probably doing something right (or being deployed exclusively in the offensive end).

That said, there are limits to the application of Corsi and it’s not an all-encompassing, “perfect” number for judging how valuable a player is. There’s a lot of context that Corsi alone will miss (it doesn’t account for usage, or quality of competition, or production efficiency). What it primarily lets us see is what players and teams are controlling the flow of play.

When was it created?

The statistic is named after Jim Corsi, a former NHL goaltender who is currently a goaltenders coach for the Buffalo Sabres. When Jim was a professional goalie, he sported a wicked Afro and mustache combo. He also noticed that the shot totals in a particular game wouldn’t necessarily reflect the amount of work he’d put in that night. As he told David Staples in a fantastic interview:

“When you look at 20 or 25 shots, you say, ‘Well, it’s a light game.’ But if you look at a game, an (attacker) comes down the right side, he’s got the puck, and he turns up the boards, towards the wall, the goalie is not standing on his heels going, ‘Oh, you know, whatever.’ He’s going to bear down, tense up, prepare that this is going to end up as a shot. Now as (the attacker) turns, he (the goalie) has still got to be engaged and wait for the next play.”

“You get a 25-shot game, but with all the blocked shots and the increase in blocked shots and the tightness of the defensive zone coverage, the activity of the goalie could almost be like 75 to 100 actions at the net.”

So Corsi developed his number early in the oughts when he became a goaltenders’ coach in an attempt to more accurately track the workload his goalies were facing on any given night. Innovation often has unintended consequences, and in the case of Corsi numbers, Corsi’s metric ended up being widely adopted by hockey-math nerds as a metric to measure puck possession.

Where can I find it?

Gabe Desjardins’s indispensable site Behindthenet.ca is the easiest resource for checking out a skater’s raw corsi number (his Corsi-ON), or his relative Corsi number (that players’ corsi number minus the the team’s Corsi number when that player is off the ice).

For team numbers, you can use timeonice.com’s “mplayershots” script. Timeonice.com is one of the best advanced stat resources out there, but it is extremely user unfriendly. Here’s the Canucks’ overall team Corsi numbers from this season (that link will take a while to load) and here’s a guide for using Timeonice.com from Red Line Station.

Which Canuck does it flatter?

How much fun is it to watch Chris Higgins do battle along the boards? His ability to maintain puck possession by kicking his puck to his stick in traffic may not show up on the scoreboard or in the boxscore, but fans can see it with their eyes and are well aware of how valuable it is. Corsi, on the other hand, can measure it.

Higgins’ puck-battle prowess will show up in his Corsi number because his ability to maintain possession leads to more offensive zone-time for his line-mates. Higgins’ Corsi-on is middle of the road on the team, but if you adjust it for his usage – he’s right at the top.

Which Canuck doesn’t it flatter?

Corsi was called out and derided as bupkus by Don Cherry a few seasons ago because it didn’t flatter a particular former Canuck. That former Canuck, of course, was Ryan Johnson.

In fairness to Ryan Johnson, that guy was like a precursor to Manny Malhotra — Vigneault used him in some extremely tough minutes. While Malhotra is able to win face-offs, clear the zone and turn play the other way, however, Johnson was significantly less skilled in that regard. As such, when he was on the ice he was usually in the defensive zone, and so was the puck.

FENWICK

What does it tell us?

Fenwick numbers are a modified version of Corsi numbers. A Fenwick number is found by counting up all of the events a Corsi number does with the exception of blocked shots. The idea is that blocked shots, usually the fault of the shooter, shouldn’t be counted — since they have no chance of ending up in the back of the net.

Why is that important?

It’s important because Fenwick is a more accurate measurement of defensive play than Corsi is. In fact, Fenwick is particularly interesting where the Canucks are concerned since it seems like it may have a strong correlation with the Canucks’ in-house tracking of scoring chances. For teams, the “Fenwick Tied” measurement (which counts all “Fenwick events” for and against while the score is tied) was an extraordinarily accurate predictor of how last season’s playoffs unfolded. In my mind, it’s the gold standard for predictive team stats. At least, it was the best predictive team metric until someone realized that the “number of sexual predators who live close to an NHL arena” is comparably accurate (seriously).

When was it created?

Fenwick is a new stat. As far as I can tell it’s at least five years old, and was created by Matt Fenwick of the Battle of Alberta.

Where can I find it?

Fenwick requires some heavy lifting to calculate. For Fenwick Tied your need to use timeonice.com. For individual Fenwick numbers or you can go to behindthenet.ca’s Corsi-On numbers, and subtract the difference between blocks for and against from a player’s Corsi-On number. Alternatively, you can use timeonice’s “mplayershots” script.

Which Canuck does it flatter?

Chris Tanev, and Dan Hamhuis.

Which Canuck doesn’t it flatter?

Alex Edler.

OFFENSIVE-ZONE START PERCENTAGE

What does it tell us?

Zone-Starts are really simple: they count up where every player’s shift starts. A skater’s offensive zone-start percentage (Ozone%) is calculated by dividing a player’s total number of offensive zone starts, by the sum total of that player’s offensive and defensive zone starts.

Why is that important?

It lets you know where a coach deploys a player. Where a shift begins also has a huge impact on a players’ corsi/fenwick numbers, and while some dispute this, it probably has a significant impact on production as well.

When was it created?

The “Oilogosophere” started talking about zone-starts (though they didn’t call them that) on a message board called “Oilfans.net” before the lockout, and Gabriel Desjardins’ started tracking zone-starts beginning in the 2007-08 season. While zone-starts have been around for a while, this is their breakout season. They’ve gone mainstream and have been referenced in the Vancouver Province and on Hockey Night in Canada. That is largely thanks to the radical specialization of the Canucks roster under head coach Alain Vigneault.

Where can I find it?

You can find it at Behindthenet, or you can use the “teamfaceoffs” script at Timeonice.com.

Which Canuck does it flatter?

The Sedins start the highest rate of shifts in the offensive-zone of anyone in the entire league. In fact, they probably start the highest rate of shifts in the offensive-zone of anyone in NHL history. So, you know, them.

Which Canuck doesn’t it flatter?

The other side of that coin is Manny Malhotra. Poor uncle Manny starts the lowest percentage of shifts in the offensive-zone of anyone in the entire league. In fact, he probably starts the lowest percentage of shifts in the offensive-zone of anyone in NHL history.

SCORING CHANCES

What does it tell us?

Scoring chances are tracked for the Canucks by Cam Charron and I over at Canucks Army. The “scoring chance project” was first launched by Dennis King of Mc79hockey.com about five years ago, and now at least eleven teams are tracked by various blogs using the same parameters.

Here’s the description that we run with every post-game chance count:

“A chance is counted any time a team directs a shot cleanly on-net from within home-plate. Shots on goal and misses are counted, but blocked shots are not (unless the player who blocks the shot is “acting like a goaltender”). Generally speaking, we are more generous with the boundaries of home-plate if there is dangerous puck movement immediately preceding the scoring chance, or if the scoring chance is screened. If you want to get a visual handle on home-plate, check this image.”

Why is that important?

The goal is to isolate the impact of variables like “luck” from the results on the ice. Scoring chances are a specific and important event that when tracked over the course of a full season gives us a good idea of which skaters are helping the team control play.

When was it created?

Scoring chances were pioneered in the late 1970s by the man whose statue is out front of Rogers Arena: Roger Neilson. Neilson counted scoring chances as a head coach, and most NHL teams count some type of in-house variation, the Canucks included.

Where can I find it?

For the Canucks you can find these numbers at CanucksArmy.com after every game. We occasionally update our result totals. Other sites that track scoring chances for various teams including Defending Big D (Dallas Stars), Mc79hockey (Oilers), FlamesNation (Flames), Blueshirt Banter (Rangers), The Washington Post (Capitals), Fear the Fin (Sharks), JetsNation (Zombie Thrashers) and Broadstreet Hockey (Flyers).

Which Canuck does it flatter?

Chris Tanev, possession demon, does extremely well by the chance data. So do the likes of David Booth, Chris Higgins, Ryan Kesler and oddly enough Mason Raymond.

Which Canuck doesn’t it flatter?

The likes of Lapierre, Weise and Malhotra have struggled by the chance data. To some extent, that’s because they play very difficult minutes for a fourth line.

QUALITY OF COMPETITION/QUALITY OF TEAMMATES

What does it tell us?

Quality of Competition (QoC) and Quality of Teammates (QoT)  are inventions of Gabe Desjardins, whose name is coming up a lot in this glossary. These stats measure the quality of players a particular skater spends the most time matched up against, or skating with. A players’ quality of competition numbers are determined by the “Average Relative Plus-Minus of opposing players, weighted by head-to-head ice time,” and Qual-Team is basically the same, except sub-out the words “opposing players” and replace them with “teammates.”

Why is that important?

Match-ups matter. A skater who can dominate against the opposition’s fourth-line and third-pairing may post excellent Corsi numbers and produce points by the bucketload, but that’s not quite as valuable as turning play the other way against the opposition’s stars, like Pavel Datsyuk routinely does.

When was it created?

Gabe Desjardins is the brainchild of quality of competition and quality of teammates, and his site tracks them dating back to 2007-2008.

Where can I find it?

You’ll find it under the “player breakdown” section at Behindthenet.ca.

Which Canuck does it flatter?

Hamhuis and Bieksa play against the toughest competition among the team’s defenseman, while the Sedins face the toughest competition among Canucks forwards. Appropriately, Alex Burrows leads the entire team in QualTeam.

Which Canuck doesn’t it flatter?

Dale Weise skates with the “worst quality of teammates” of any Canucks skaters. Not so coincidentally, he also skates against the softest competition of any Canucks skater.

PDO

What does it tell us?

PDO is the sum of a team or player’s on-ice shooting percentage and on-ice save percentage. Those numbers should add up to one-hundred, and if it doesn’t – it’ll regress towards that mean over a large enough sample.

Team’s can usually beat the odds over the course of a full season, or a handful of seasons, if their goaltending is good enough.

Why is that important?

During the early parts of the regular season – it allows us to qualify whether or not a team has been bad or whether or not they’ve just been unlucky. Conversely, PDO will help us identify whether or not a team is actually good, or if their being on top of the standings is a mirage, and that they’re likely to plummet over the balance of the season. In the past, PDO has predicted the demise of the 09/10 Colorado Avalanche, the 10/11 Dallas Stars and the 11/12 Minnesota Wild.

When was it created?

It was created in the comments section of Irreverent Oilers Fans by Brian King, roughly in 2006. Brian King’s username on the blog was “PDO” and that’s where the stat takes its name. I interviewed Brian in this space in the fall.

Where can I find it?

You can find it at behindthenet.ca. Behind the net also offers this handy PDO graphing tool.

Which Canuck does it flatter?

Somehow — and this defies all logic — Mason Raymond has been the team’s luckiest player this season. Usually, PDO can be used to qualify the impressions of a rabid fan-base, and a guy who is a frequent scape-goat usually has a sub-100 PDO. Raymond has a sky-high PDO, and yet has accrued absolutely none of the usual benefits. So maybe it doesn’t flatter him after all.

Which Canuck doesn’t it flatter?

David Booth has been excruciatingly unlucky for years now.

Tags: , , , , , ,

13 comments

  1. JaspreetGill
    March 17, 2012

    Thanks so much for making this! I knew a few of these, but was confused a lot. Nice work! :)

    VA:F [1.9.16_1159]
    Rating: +3 (from 3 votes)
  2. sarah
    March 17, 2012

    I’m so glad you did this breakdown! I’ve always wondered about these advanced stats but never wanted to ask for clarification and appear stupid.

    VA:F [1.9.16_1159]
    Rating: +2 (from 2 votes)
  3. sarah
    March 17, 2012

    Also, for a guy who is apparently incredibly unlucky, David Booth is remarkably positive and upbeat…all the time.

    VA:F [1.9.16_1159]
    Rating: +8 (from 8 votes)
    • Harrison Mooney
      March 17, 2012

      No kidding.

      VN:F [1.9.16_1159]
      Rating: +1 (from 1 vote)
    • J21
      March 19, 2012

      Booth’s demeanor no doubt has to do with a certain long-haired, bearded Jewish man with some crazy ideas that turned the world upside down.

      Yes, I’m talking about Karl Marx.

      VA:F [1.9.16_1159]
      Rating: +2 (from 2 votes)
  4. Anonymous
    March 17, 2012

    This is so useful! Thanks for taking the time.

    VA:F [1.9.16_1159]
    Rating: 0 (from 0 votes)
  5. Arnold Jamtart
    March 17, 2012

    Great article! Question: if +/- is worthless, why use it in QoC/QoT? Why not use Corsi? Or will using an advanced stat inside another advanced stat unwittingly fire us into a ditch near the New Jersey Turnpike?

    VA:F [1.9.16_1159]
    Rating: +1 (from 1 vote)
    • ArtemChubarov
      March 17, 2012

      That’s a legitimate problem with qualcomp/team. Which is why, hockey math nerds tend to prefer “Corsi relative Qual Comp” which is less reliant on +/-. For the sake of keeping this simple, and introductory – I just explained basic qualcomp (which is still useful, if flawed).

      VA:F [1.9.16_1159]
      Rating: +2 (from 2 votes)
      • Lenny
        March 18, 2012

        Does the ‘Relative Corsi Quality of Competition’ on behindthenet.com express this? Or does it simple express ‘Average Corsi of opposing players, weighted by head-to-head ice time’ as the explanation says?

        VA:F [1.9.16_1159]
        Rating: 0 (from 0 votes)
  6. akidd
    March 17, 2012

    thanks so much for the glossary. very clear and concise and soon to be favourited. this info helps to cement my admiration for tanev and hamhuis who have long appeared to my eye as such solid,smart dmen( hamhuis for longer but tanev as the up-and-comer who has yet to disappoint.)conversely, it also feeds my sneaking suspicions about edler’s defensive prowess.

    now booth is quite the stat buster it seems. i’ve admitted that i’ve been warming to him, mostly because of the energy which he seems to bring each and every shift(and his sadsack sandwich options.) but i’m also dismayed at his lack of vision and his selfishness with the puck. that he continues to shoot from anywhere with that muffin of a shot leaves me dumbfounded. how does he get away with not being reprimanded by the coaching staff(and teammates besides raymond) for ‘killing’ so many chances?

    why do i think booth a stat-buster? pdo- i haven’t checked his shooting % but i thas to be low. he has such a weak and inaccurate shot and he lets it go from anywhere. he often doesn’t wait for a screen and telegraphs too. his best real chances come crashing the net but there too the puck often rolls off his stick. ‘unlucky for years’ is not how i would describe it.

    corsi/fenwick- again he shoots the puck from anywhere and kills play after play either by fluttering the puck at the goalie’s logo or ignoring that a defender’s stick or body is obviously in a postiton to block his shot, shooting it anyway and having it deflected over the glass for an o-zone start. and he plays almost every shift with the defending selke winner as his centre taking care of the defensive side.(ergo lots of shots for and lots of shifts ending in o-zone starts; and not so many against thanks to kesler.)

    chances- yes, he shoots from inside homeplate which technically means a scoring chance but as mentioned above, it’s not really a true scoring chance with that shot. and often he makes the anti-scoring choice by ignoring his open teammates or delaying until they are no longer open(and he has no option but shoot the puck or lose it.)

    i like his energy and for a couple million less a year booth would be a steal. but i do find it funny how he seems ,to me, to epitomize the disconnect between observation and advanced stats. if this advanced stat stuff eventually goes the way of the dodo bird you’ll have blissed-out loopholers like booth to blame.

    VA:F [1.9.16_1159]
    Rating: 0 (from 0 votes)
  7. Brent
    March 17, 2012

    Thanks for this BUT had a visual image of Fenwick as a quirky, bald English statistician. Now I find out it is some guy from Alberta. Totally crushed! Is he pudgy and bald?

    VA:F [1.9.16_1159]
    Rating: +4 (from 4 votes)
  8. SteveB
    March 17, 2012

    Thank you.
    I’ve often puzzled over the meaning/significance of these arcane terms.
    This article helps a lot.

    VA:F [1.9.16_1159]
    Rating: 0 (from 0 votes)
  9. Aderam
    March 18, 2012

    This is cool! Thanks for laying it out in a way that’s easy to understand. I admit that I’m an artsy and will never be a true stats nerd, but it’s good to see the numbers every now and then, and this will make that easier!

    VA:F [1.9.16_1159]
    Rating: 0 (from 0 votes)