clock menu more-arrow no yes mobile

Filed under:

Explanations of Used Terms & Abbreviations

On In Lou We Trust, we tend to use several terms and abbreviations that may be common online for hockey analysis but isn't really all that common in general use or most hockey discussion. For the sake of reference, I've decided to put together a glossary of these terms and abbreviations. I'll update this as necessary to either add new terms or clarify explanations.

Corsi - Corsi is a plus-minus value for shooting attempts. All shooting attempts are either goals, shots on net, missed shots, and shots blocked. The sum of all attempts for minus the sum of all attempts against provides a Corsi value. For example, if Ilya Kovalchuk was on the ice for 2 goals scored, 6 shots for, 4 shots against, 1 blocked shot for, 2 blocked against, and 3 missed shots for and against, his Corsi would be +3. 12 (# of attempts for) - 9 (# of attempts against) = 3.

These values are usually collected for even strength, non-empty net situations so they reflect how the player done at even strength beyond whether or not they scored. Over a longer period time, if a player keeps racking up positive Corsi values, then we say he's likely driving the play forward. It's very difficult in hockey to just be on the ice and not have some contribution to offensive success. This can also tell us if a player has been doing poorly. If a player keeps getting negative values, that means the other team has been getting more attempts while they're on the ice. We can say he's getting pinned back, he's drowning, he's getting wrecked, etc. depending on the severity of the value.

We care a lot about Corsi because it approximates possession. Possession is everything. The NHL doesn't record time of possession, so this is the closest way of measuring it using information the NHL already collects in each game. Players will almost always attempt to take a shot on net if they're in the other team's end of the rink. Even strength is the most common situation in hockey, so if a team or a player that out-attempts their opposition in this situation, we know that they're in the other team's end. Generally, the team that manages to do this over and over tends to controls the game. If they can do it regularly, they're usually one of the best teams in the NHL.

Time on Ice has a script that calculates this for each player in a game; here's an example. Look at the goaltender's Corsi value for the team's Corsi value in a game as they are (usually) on the ice for the whole game. Behind the Net collects on-ice Corsi rates for a player, here's an example for the 2011-12 season.

Fenwick - Fenwick is a variation of Corsi. It's calculated the same way except shot blocks are not included. This means it's a sum total of shooting attempts for and against that didn't getting stopped along the way. In a single game, this may not mean as much; but in the long term, it's preferred as a measure of possession as JLikens shown that this post at Objective NHL.

Fenwick% is a common stat at Behind the Net kept for teams. It's simply the sum total of Fenwick for events (shots on goal for, goals for, missed shots for) over all Fenwick events. Anything over 50% is preferable. Like Fenwick (and Corsi) values, higher is better.

Score Effects - As seen in most hockey games, a losing team is going to be more aggressive and attack more often. A winning team tends to lower their aggression or hang back either by choice or by force from the losing team. Therefore, both Corsi and Fenwick are affected by the score in a game.

JLikens found that Corsi and Fenwick can and should be adjusted for score effects by only counting effects in certain situations. In addition to even strength, non-empty net situations, "close score" situations has been found to be a good in-between situation compared to score tied situations (minimizes score effects, a much smaller population) or any score situations (score effects plentiful, a much larger population). Close score situations take place when the game is within one goal in the first and second periods, or the score is tied in the third and overtime periods.

Behind the Net's team Fenwick% page defaults to close score situations, though numbers are calculated and available for other situations (up a goal, down a goal, etc.)

SF/60 - Shots For per 60 minutes

SA/60 - Shots Against per 60 minutes

GF/60 - Goals For per 60 minutes

GA/60 - Goals Against per 60 minutes

OZS - Offensive Zone Starts - A player starting a shift in the end of the ice closer to the opposing team's goaltender. These are easier starts because the opposition winning the draw is still at the opposite end of where they need to go.

DZS - Defensive Zone Starts - A player starting a shift in the end of the ice closer to the goaltender. These are more difficult starts because the opposition winning the draw forces them to get set on defense quickly - if it's even possible.

OZS% - Offensive Zone Start percentage is calculated by the number of offensive zone starts over the sum of offensive and defensive zone starts in even strength situations. Neutral zone starts are exactly that - neutral, so they're not included. A player with a percentage over 50% is better protected and/or the coaches want them as offensive players. A player below 50% is getting more difficult starts and/or the coaches prefer them as defensive players.

Corsi and Fenwick can be adjusted for zone starts as it's more difficult to generate Corsi and Fenwick events when the team is starting in their own end of the rink. JLikens figured out the adjustment for zone starts at Objective NHL which, in turn, can be adjusted for either value or rate.

GVT - Goals Versus Threshold. Tom Awad of Hockey Prospectus developed this stat as another way to measure player value. Basically, it measures a player's contributions against a replacement-level player and the difference provides the value for this stat. Tom Awad has a three part explanation going into the details of how GVT was formulated: Part 1, Part 2, Part 3

PDO - PDO doesn't stand for anything, but it's a quick check whether a player has been hurt or helped by good luck. It's the sum of the team's shooting percentage and the team's save percentage on the ice. 1000 (or 1, depending on the initial values) is considered average. Over time, a player will regress towards their true mean. So a player with a lower PDO than usual will eventually get better and a player with a higher PDO than usual will eventually go down. Whenever eventually is, who knows.

Scoring Bias - Ultimately, most (if not all) of these advanced stats are based on the official stats recorded during each NHL game. However, each arena has their own scorer and they have their own biases ranging from basic counting stats like shots on net and blocks to the real time super stats like hits and giveaways. For example, here's an analysis from a few years ago where I showed what the Devils' scoring bias has been.

There's not a lot that can be done about this. Using data from road games to minimize bias is the main way to account for it, but it does lead to a smaller pool of data.

Corsi Rel QoC - This is a stat that measures a player's Corsi relative to their quality of competition. Gabe at Behind the Net developed Quality of Competition as a means to measure who's playing against tough competition or not. This takes their Corsi value into account. A higher value at even strength means that they're playing against players who drive the play more, or simple, better players. A lower value (negative) at even strength means they're playing against weaker competition. It's a good way to identify who's playing against tough competition, especially while taking OZS% into account.

Comparing this value within players from varying teams doesn't work too well since players play different schedules and different teams of varying quality. A better means of comparison would be to compare the rank of those players on their respective teams.