Corsi & Fenwick Introduction
Average Corsi per line
1st line = 52.1
2nd line = 50.7
3rd line = 49.3
4th = 47.1
For context, only the very best and worst teams in the league had shot differentials at or near five per game. Nine teams had shot differentials less than one per game.
Glossary
Corsi
Corsi /Shot Attempts (SAT): Corsi events consist of all shot attempts: blocked, missed, saves, goals.
CF: Shots produced by a player or their teammates when the focal player is on the ice.
CFoff: Shots produced scored by a player’s teammates when the player is off the ice.
New ones:
iCF: Individual shots.
BK: Shots that an individual attempted that were blocked.
AB: Attempts Blocked, shots that the individual themselves blocked.
CP60: Corsi Pace (per 60 minutes), equal to CF60 + CA60.
OCOn%: On-ice Corsi on-goal percentage: a player’s on-ice shots on net (SF) divided by their on-ice shots (CF).
OCAOn%: On-ice Corsi on-goal percentage against: a player’s on-ice shots on net against (SA) divided by their on-ice shots allowed (CA).
Fenwick
Fenwick (F)/Unblocked Shot Attempts (USAT): Take shots on goal and add shots that miss the net entirely.
FF: Unblocked shots produced by a player or their teammates when the focal player is on the ice. FFoff: Unblocked shots produced scored by a player’s teammates when the player is off the ice.
New ones:
iFF: Individual unblocked-shots-on-goal for.
MS: Individual shots that missed the net.
PFenSh%: Personal Fenwick shooting percentage: a player’s goals (G) divided by their individual unblocked shots (iFF).
OFenSh%: On-ice shooting percentage: a player’s on-ice goals for (GF) divided by their on-ice unblocked shots (FF).
OFenSv%: On-ice save percentage: a player’s on-ice goals against (GA) divided by their on-ice unblocked shots against (FA).
FenPDO: The sum of on-ice Fenwick save and shooting percentages (OFenSh% + OFenSv%). League average is 100 by construction. Not on the site at present.
FP60: Fenwick Pace (per 60 minutes), equal to FF60 + FA60.
OFOn%: On-ice Fenwick on-goal percentage: a player’s on-ice shots on net (SF) divided by their on-ice unblocked shots (FF).
OFAOn%: On-ice Fenwick on-goal percentage against: a player’s on-ice shots on net against (SA) divided by their on-ice unblocked shots allowed (FA).
IMPACT OF PUCK POSSESSION & LOCATION
FIRST SIGN OF DOMINANT TEAM
the statistic that most quickly accumulates that is well enough correlated with winning and losing that it is worth discussing is shots on goal. In the average NHL game a team will have twenty or thirty something shots on goal. After approximately ten games played, the differences between most teams in shots on goal are becoming statistically meaningful. The correlation between taking shots on goal and scoring and not allowing shots on goal and not being scored on and thus winning and losing from the shots on goal totals are meaningful enough that they give us a good picture of how well teams re performing. Of course, there are complicating factors such as shot quality and quality of goaltending on a given team which they do not capture, but nevertheless, they give us a first good look at how well teams are performing.
CORSI NUMBERS
(Oct 2007)
Unstoppably, player’s results will gravitate towards the Corsi numbers as the season goes on. A few good bounces here and there can make the early stats misleading. To pick on an old favourite, Lupul is EV+/- +3, and Corsi -41. The drop is inevitable, just is.
And as with all hockey stats, without applying the context (i.e. who these players were playing with and against) it will be of lesser value.
This Corsi expression seems like a keen way of eyeing exeptionally strong or extraordinarily ineffective forwards. I think there’s too many variables to make accurate assumptions on players who fall in the middle ground. In my opinion forwards would drive these results, and defenders, minus perhaps the rare exception (see: Lidstrom), would be passengers on the bus. By that I mean that I don’t often see defencemen forecheck, pressure for turnovers in the neutral zone, or cycle down low
To me it’s a pretty solid indication of zone time and scoring chances at evens. Of course if your team can’t bury their chances, that isn’t necessarily enough.
DRIVING POSSESSION - FENWICK
This is a chart of all shots directed at net while a player was on the ice at even strength, excluding blocked shots. And since that takes a long time to say, I’ll call these Fenwick Numbers from now on, since he was the lone lobbyist for this metric.
It says a lot about which end of the rink the puck was at, and which direction the scoring chances were coming from.
POSSESSION IS EVERYTHING
(May 2009)
I think that Mike Babcock is right, possession is everything. Damn close to it at even strength, anyways. I also agree with his thinking that being a “puck possession” team has little to do with coaching style, and everything to do with how good your players are.
The idea is that the more games you’re looking at, the more luck washes out. And you can see that over a small batch of games the player’s scoring chance rates mesh reasonably well with results (goals) for these 20 Oiler players as a group. But as the sample of games grows larger, near the 77 total, the relationship becomes overwhelming and obvious. And if the season were 40 games longer there is every reason to think that the correlation would grow even stronger, even though there isn’t much room to grow at that point.
Bottom Line: Players that drive possession at EV drive results at EV. Maybe not the next game the Oilers play, or the next dozen, but eventually. It’s unstoppable, ability trumps luck eventually, you just have to be patient.
Corsi begets scoring chances, scoring chances beget goals, and even over the course of a full season the hockey gods have a huge say in a player’s scoring (and outscoring) results, and a big say in the same results for the team
WHAT IS CORSI AND WHY DOES IT EXIST
In a nutshell, the Corsi Number is the shot differential while a player was on the ice. This includes not just goals and shots on goal, but also shots that miss the net, and in some formulations, blocked shots. In other words, it’s the differential in the total number of shots directed at the net. This metric was presumably adopted by the Sabres because it’s a better indicator of a team’s play than goals for and against, which are highly-driven by factors outside of a team’s control.
Shot volume is much more a function of a team’s ability, and a much better predictor of future performance than goal-scoring metrics – in other words, there is basically no such thing as a team that shoots efficiently, just teams that get a lot of shots on goal…or not.
CORRELATION OF FENWICK WITH TEAM POINTS
(Sept 2009)
Both Fenwick and Corsi Numbers are attempts to find a replacement for +/- ratings that will include more events and thus have higher signal to noise.
Today I will address the question of whether the Fenwick or the Corsi Number better correlates with winning.
I have already shown the correlation between Corsi Number and team points is 0.626. Corsi shows a pretty strong relation with winning hockey games.
In general, teams that have higher Fenwick Numbers have more points. The correlation between team Fenwick Number and points is 0.602. This correlation is almost as strong as that of Corsi, but it is slightly worse. Omitting blocked shots that are included in Corsi makes this rating slightly less of an indicator of winning hockey.
Fenwick Number is a useful measure and in some circumstances may be better than Corsi, but in general Corsi is the preferable measure of the two.
CORRELATION BETWEEN SHOTS ON NET & WINNING
(Nov 2009)
How much of a correlation is there between shots on net and winning?
Using the Pearson Correlation Coefficient, the findings establish that there is a 0.48 correlation between shots on goal and accumulating points in the standings since the lockout. This, while not the strongest of correlations, is still fairly significant. Overall, the above numbers suggest that a higher number of shots on goal will generally lead to a higher ranking in the standings
Having said that, teams like the Leafs that lead the league in shots on goal per game should eventually find their way out of the basement. Talent, however, will dictate just how far out of the basement those teams rise.
SHOTS ON YOUR OWN NET AND STANDINGS
(Dec 2009)
it seems obvious to examine the correlation between shots on your own net and points in the standings.
The Pearson Correlation Coefficient between shots on goal and points in the standings was 0.48—significant but overly substantial. So, in doing last week’s piece, I wondered whether shots on goal would be more indicative of team success than shots allowed. Using the Pearson Correlation Coefficient, the correlation between shots against and total points in the standings was -0.53. Remember that fewer shots on goal is obviously the objective, so the -0.53 is actually very close to the 0.48 correlation demonstrated last week. In fact, shots against may be slightly more indicative of team success in the standings than shots on goal.
All in all, preventing shots on goal is an integral part of the game (maybe even more important than putting shots on the opposition’s net) but without strong play in other aspects of the game—like goaltending—preventing shots can only carry you so far.
EVEN STRENGTH EFFICIENCY AND POINTS ACCUMULATION
(Dec 2009)
Considering the majority of an NHL hockey game is played at five-on-five, you would guess that the correlation between even strength efficiency and points accumulated in the standings would be high.
Well, let’s take a look to determine if that’s the case.
With the naked eye, it looks as if there is quite a correlation between five-on-five ratio and points in the standings. Well, the naked eye is indeed correct. The correlation coefficient is a whopping 0.86.
So, we can talk about a lot of factors, but without even strength production, it appears your favorite team is doomed to fail.
We can talk all we want about power plays, penalty killing, discipline, etc. However, all you really have to do is look at the numbers. If you want sustained success in the NHL, you have to be a proficient team when it comes to even strength production.
THOUGHTS ON CORSI
(June 30, 2010)
In his role as the Sabre’s goalie coach, Corsi was attempting to evaluate the work load his goalies had in a game of play and found that simply shots against were not sufficient. The goalie can relax whenever the puck is in the oppositions end, but whenever the play is in his own end he can’t relax, regardless of whether a shot was taken or not. To get a better idea of his goalies workload he summed up shots, missed shots and blocked shots which should give a much better indication of a goalies overall work load
More recently others in the hockey community have extended Corsi numbers to evaluate a teams ability to control the play of a game (i.e. does a team play more in the oppositions zone vs their own) and evaluate individual players by looking at their Corsi numbers for and against while they are on the ice and comparing that to their teammates Corsi numbers
When used in this context Corsi and Fenwick numbers are calculated just as +/- is calculated which is to take the shots+missed shots+ blocked shots for his team and subtracting the shots+missed shots+ blocked shots numbers by the opposition while he is on the ice.
Using Corsi numbers as a way to evaluate a goalies workload is probably far more valuable than just using shots on goal. Beyond that, I am pretty sure that Corsi numbers will give a pretty solid indication of a teams control of the play,
One problem that Corsi hides has is that a team with a strong set of forwards but a weak defense and goalie may control the play more than a team with a strong defense and top tier goalie but is that team really any better at winning games?
I do believe that Corsi numbers have a use in evaluating a goalies work load and even in showing which teams are controlling the play, but in my opinion using it anywhere beyond that we are making too many assumptions about how important Corsi numbers are with respect to winning games.
RETRO NHL & ANGER AT CORSI
Goals for and goals against drive outcomes, and they are very highly driven by shooting percentage, which itself is not a sustainable talent. There are plenty of ways to figure this out – comparing first-half team shooting percentage to second-half; or even games to odd ones; or even shots to odd shots. None of them show a persistent relationship.
What does show a persistent relationship is shot differential, in particular shot differential with the score tied at even-strength. The best predictive performance comes from goals, saves and missed shots taken together – blocked shots are driven to a great extent by a team’s ability to block shots, so they’re not as useful. This is what’s commonly-referred to as “Fenwick”, while Corsi (usually) includes blocked shots.
Together, Fenwick/Corsi and Luck account for around ¾ of team winning percentage. What’s the remainder? Goaltending talent – which Tom Awad estimates at about 5% - and special teams, along with a very small sliver that’s due to shooting talent and the oft-mentioned “shot quality.”
Corsi is a proxy for something else – puck possession, territorial control, scoring chance differential, all of which are almost identical metrics – and does anyone really doubt that having the puck more than the other team does not lead to winning?
Goaltending and shooting talent have been studied by many different people in many different ways and have been found to drive only a small slice of the results – 1 or 2 wins per season.
WHAT IS CORSI AND HOW DO YOU USE IT
A player’s Corsi Number is just the sum of all shots directed by a team towards the opposition net (shots on goal, missed shots, and blocked shots), minus the sum of all shots directed by the opposition towards your team’s net. For an individual player the Corsi number is just the value determined while they are on the ice.
Corsi = Shot Attempts FOR – Shot Attempts AGAINST
We use Corsi as a proxy for puck possession. Logically, the team that has the puck more is going to make more shot attempts. Generally this has been shown to be the case in studies of actual puck possession time, scoring chances, and general shot attempts. The correlation is very high amongst these variables.
So basically, what this means is - players with high corsi numbers are on the ice for a higher number of shot attempts for their team than the other team. That is to say, their team has the puck more when they're on the ice.
While a certain player may not see a positive corsi result in an individual game, their cumulative corsi over time should indicate their overall possession. This long term seasonal corsi value is shown in a couple of different ways.
Corsi ON = (Total Corsi) x (60 mins) ÷ (The player’s total ice time)
Corsi % = the percentage of Corsi events a player’s team has while they are on the ice
Corsi REL is another statistic that is meaningful here, as it indicates how solid a player’s possession statistics are relative to those of their team-mates. Similarly to how we tabulate Corsi ON, we can also tabulate Corsi OFF, which is the team’s overall Corsi number while a player is OFF the ice. Corsi REL is just the difference between these two values.
Corsi REL = Corsi ON – Corsi OFF
Shots are an indication of which team is driving play. Whoever is shooting more, has the puck more, and the team with the puck more tends to score more often. This is how hockey games are won. It might not overcome bad luck every night, but over the course of a full season, it’s pretty important.
INTRO TO ADVANCED STATS - FENWICK
Fenwick is best used as a proxy differential in scoring chance opportunities. Players and teams with higher Fenwick numbers are typically getting more scoring chances than the opposition. In another form Fenwick Percentage (aka Fen%) is a solid representation of the percentage of scoring chances a team is getting while specific players are on the ice, or overall.
Fenwick is actually a slight modification of Corsi (which we described in the last posting). Basically Fenwick excludes blocked shots. So instead of ALL shot attempts, it only includes shots on net and missed shots. It is just the differential between the for and against for these values.
Fenwick correlates slightly more highly with scoring chances than Corsi tends to, though not necessarily by a wide margin and it doesn’t particularly do a better job of indicating possession.
The argument for using Fenwick is basically:
1. The whole (or perhaps best) use of Corsi is to have objective figures that can be used as a proxy for scoring chances (what else are you using it for?).
2. A shot that is blocked is either a) not a scoring chance at all, or b) on average from a worse scoring area than shots/posts/missed shots.
Blocking shots may in fact be a skill, so from a defensive standpoint, excluding them removes a penalty many defensive D men face as skilled shot blockers. Similarly, players that excel at getting a high proportion of their shots past shot blockers are rewarded in this metric
At the team level, Fenwick percentages are probably one the best predictors we have of future success. Teams that can sustain a high Fenwick percentage over the long term tend to be out playing their opposition at a high level. This also belies some teams that are succeeding due to unsustainable luck.
Generally speaking, teams that have the best Fenwick Percentages while close are the best teams in the NHL.
Individual Fenwick Numbers also provide a solid perspective on how a player does with regards to scoring chances.
Overall this is just another tool in our box of ways of examining the play of the players and teams we’re interested in.
A VISUAL GUIDE TO FENWICK
https://www.habseyesontheprize.com/2013/4/4/4178716/why-possession-matters-a-visual-guide-to-fenwick
Stats guys are always talking about Fenwick, which is a differential between shots plus missed shots directed at the opposing net, and directed at your net. Why is this statistic important?
You can see the magic number of success is +.500. If you manage to crack this number you have a greater than 75% chance to qualify for the playoffs. If you break the +.550 mark you have a 25% probability of winning the Cup.
FENWICK
When used in plus/minus format, it calculated by subtracting negative unblocked shot attempts from positive ones, usually expressed as a differential for every 60 minutes of play in a given situation.
While putting Fenwick in a percentage format gives a clearer picture of a player or team’s performance, the plus/minus format can add additional context if a player or team takes part in more events than another. New Jersey for example, limits shots both for and against, which is why their games are typically low scoring.
As with all possession metrics, Fenwick can also be broken up into for and against statistics, which is a better usage if you want to measure specifically offense or defense.
Fenwick has a great correlation with scoring chances, meaning a team that dominates unblocked shots is usually generating more scoring chances, and has a much greater chance of scoring, and therefore winning, the game.
The natural correlation of more shots meaning more chances to score a goal means winning the Fenwick battle carries with it predictive value.
Fenwick correlates with time of possession, but not as strongly as Corsi does. However Fenwick carries a much stronger correlation with future scoring than Corsi.
Fenwick has another limitation that applies less so to Corsi, and that’s sample size. Because Fenwick uses fewer data points than Corsi overall, its predictive power is weaker in small sample sizes than Corsi is. Once enough games have been played though, Fenwick is the better statistic.
CORSI PREDICTIVE VALUE AND SHOT QUALITY
Most analysts rely heavily on shot differential, or Corsi, to make predictions for teams. When they do that, they’re largely ignoring shooting percentages – a decision they make based on evidence that shooting percentages have a lot of random variance, and that shot differential is actually a better predictor than the number of goals scored or wins compiled.
But Phil Birnbaum had an interesting post Tuesday challenging that idea, proposing that shooting a lot might be correlated with taking lower-quality shots.
He pointed to some pretty well-established score effects, or changes teams make depending on the score of a game, wherein teams adopt a defensive shell with a lead, taking fewer shots but for a higher percentage. He additionally noted that on some plays a player might pass on a low-percentage shot for the chance at a higher-percentage one, and wondered if there weren’t strategies and/or random fluctuations that could result in both a low shot total and a high shooting percentage.
The key piece of supporting evidence he provides is this:
I looked again at the last six years of the NHL, 180 team-seasons. The correlation between shooting percentage and Corsi was -0.22. When I took only the 36 most extreme shooting percentages, the correlation was -0.37.
In other words, he’s saying that over the last six years, teams with a bad shot differential also had a modest tendency to have a good shooting percentage.
If true, that would mean that teams with bad shot differentials do a little bit better than you’d expect. It wouldn’t change our observation that shot differential is a strong predictor of future results, but it would make it a little easier to believe that system might play a role worth considering.
When teams trail or are tied, they all play essentially the same way. What difference exists is almost entirely in how teams play with a lead; it appears that some play more conservatively than others, which I find very interesting
Phil’s point that it’s possible for a team to deliberately adopt a low-shot high-percentage strategy in all situations is obviously true. But the only place where there is any real evidence for teams doing that is when they are protecting a lead, and even there the differences aren’t terribly large.
So if we’re going to make a case for a team doing better than shot differential would suggest, we should focus specifically on their performance with a lead, and expect that impact to be minor.
Of course, it’s possible that they’ve hit on a radical new strategy for tied and trailing situations. But since we’ve never seen that before, with each subsequent team that has a high shooting percentage in those situations, the safer bet is that they’ve simply run hot and will slide back to the pack.
POSSESSION SHOT DIFFERENTIAL
One of the most widespread tools for team analysis is shot attempt differential (Corsi) when the score is close. Looking specifically at those close situations (within a goal in the first two periods or tied in the third period) removes most of the impact of how teams tailor their strategy to the score, and gives a strong measure of which teams are best at controlling play and outshooting their opponents.
Since score effects show up most strongly in the third period, maybe just the shot differential from the first two periods would be a reasonable surrogate for score-close Corsi.
Shot differential in the first two periods correlated to winning in the mid-50’s, and at how score effects looked back then. He showed that his measure correlated strongly with zone time.
In the current era, the differences between teams in shooting talent turns out to be fairly modest. And if we don’t expect one team to have a much higher shooting percentage than another, then it becomes important to see who gets the most shots. As a result, we’ve observed that shot differential is a strong predictor of future results, which drives a lot of the current analytical thinking.
SCORE EFFECTS ON POSSESSION
(Jan 2013)
I took 450+ games from 1952-53 to 1954-55 and recorded the shot totals, scores, as well as the shot totals and scores through two periods. There definitely appears to be score effects at play here, and it seems pretty uniform at those tails.
This is good news for my approach, using Corsi% from the first two periods rather than the whole game to avoid the score effects
You’ll notice the easier slope of the 2-Periods Corsi%…two groups are being pulled out of the center, the legitimate strong-Corsi teams giving up Corsi ground in the third and the lower-level Corsi teams gaining ground in the third. So it appears, at least in the assessment of 1950s teams, that the 2-Period approach is doing what I want it to do.
The question of when score effects developed in NHL history has floated around for awhile, so I decided to see if I could pinpoint it. Sure enough, if you run the correlation of individual period shots-for percentages to final goals-for percentages, you see score effects really start to take hold after about 1977 or so. In other words, it’s been around long before the “loser point” (which was introduced in the 1999-2000 season).
POSSESSION ZONE TIME AND CORSI
(Jan 2013)
when I use 2-period zone time figures from 2001-02 (the NHL’s previous attempt at measuring possession, which they inexplicably stopped doing), zone time stabilizes more quickly over progressive seasonal measurement. I’d put that at roughly 20 games
2-period shot percentage (not to be confused with shooting percentage) catches up about 10-15 games later, hence why we’re able to predict with alarming accuracy regressions in the second half by the midway point in 82-game seasons
CORSI AND FENWICK
Corsi = shots on goal + missed shots + blocked shots
Fenwick = shots on goal + missed shots
Corsi and Fenwick are used as proxy’s for possession time
You don’t see teams with high shot quantity and low shot quality, because no team is going out there with the intention of throwing the puck away continuously.
Corsi and Fenwick, these are not the major causes for wins, but are the by-product of striving for what does create wins: scoring chances and puck possession. For this very reason they correlate to these things very, very strongly and therefore can be used as a proxy for the same thing.
Both Corsi and Fenwick are counted as “For” or “Against”. “For” is a shot or event that happens while the player is on the ice that is on behalf of his team. “Against” is the same but for the opposing team. They can be applied team wide or by player.
In general, Fenwick is usually regarded as a better indicator over a longer period of time. Corsi is a better indicator over a shorter period of time.
Corsi is best used as a performance metric to better guess future performance. Havoc was raised in the short season because it can take more than a hundred games for results to match performance. This isn’t a new concept in any way.
Players often talk about getting the bounces, but over a smaller sample of games, Corsi does a better job at separating the luck from the actual performance. Recover enough pucks, play with the puck in the right spots in the ice, and make the right decisions, and you’ll come out ahead.
To make this data easier to use, statisticians express a player or team’s numbers as a percentage. CF% (Corsi For Percentage) and FF% (Fenwick For Percentage) can then be easily compared among players, teams and games.
Corsi For (CF) varies greatly over these games as does his Corsi Against (CA). Focusing on these numbers alone could be misleading for the sake of comparing his performance with other players throughout the league, his teammates or even his own play from game to game.
Using Corsi For Percentage (CF%) allows us to see how the numbers work together and remove the game to game variables that would otherwise be misleading or confusing.
SHOT DIFFERENTIAL IMPORTANCE
It is a coach’s job to deploy his roster in a way that optimizes his team’s Corsi differential and consequently improves their chances of success.
Goals are what really matter, so we’d like to evaluate players on how often they score goals or set up goals or prevent goals. But there are always random fluctuations – players or teams can get hot for a little while and look better than they really are.
The issue is that shooting percentages fluctuate quite a bit. And with only a couple of goals per game, it turns out that it takes a lot longer than people naturally expect for those random fluctuations to even out.
Those shooting percentage fluctuations have a big impact on outcomes. Over tens of games, a team’s results will be heavily driven by whether they happened to shoot for a higher percentage than their opponents did. But in today’s NHL, nobody’s been able to sustain a big edge in shooting percentage in the long run.
Because there are more than ten times as many shots as goals, the randomness in shot differential evens out a lot faster. And since the long-run differences in shooting percentages turn out to be modest, shot differential ends up being a big component of success.
If you want to predict future outcomes and only get to look at one stat, teams’ current shot differential will help you a lot more than their current record or goal differential.
Even over a month or two, teams’ records can be quite deceiving, and a player’s goal differential (plus/minus) for a whole year doesn’t necessarily tell you much about his performance.
Goaltending is 5 percent of winning. JLikens came up with a nearly identical estimate for the spread of shooting talent, so we might infer that shooting is another 5 percent of winning, so the combined impact of shooting and goaltending is about 10 percent.
In that same article, Gabe said that Corsi and luck combine for ¾ of winning, and he has shown that luck alone is about 38 percent. So we can infer that he estimates Corsi to be about 37 percent of winning – about 3-4 times as important as shooting and goaltending combined.
Shot quality isn’t insignificant, but it’s clearly subordinate.
Even in the long run, when the randomness in shooting percentage has largely evened out, we still see shot differential being three times as important a factor as shot quality in today’s NHL.
Basing a projection on a team’s shooting and save percentages over the short run – with all the attendant issues of randomness – is a very shaky prospect.
FENWICK AND PLAYOFF SUCCESS
The Stanley Cup winning median FF% is 54.0, a mark that one-ninth of the population achieved over a full season. The arithmetic average FF% is 54.4.
Only one team had a 5-on-5 shooting percentage less than 10% in their playoff runs, and having a hot goalie certainly didn’t hurt either.
Poor goaltending can undermine strong possession, and that theme is a constant.
There’s more than one way to build a successful playoff team. Building a team capable of possessing the puck at full-strength seems to be the most common one, however.
WHAT MAKES A STAT GOOD
Corsi was never intended to be a holistic number or a WAR-type statistic. It was merely brought in to try and measure an area of the game not yet being accounted for with goals, assists, and other statistics then available. It was shown to predict future success better than goals and comparably to scoring chances (but without the need to be manually tracked, like scoring chances).
We did not use Corsi over other statistics like plus/minus for arbitrary reasons. We cared because we tested them, and Corsi performed better.
We then improved upon this statistic. We adjusted Corsi to account for confounding variables like score and venue effects (adjusted Corsi) We adjusted shots for the shot quality variables we can account for, like shot location, player situation, and shooter history (expected goals).
The next step after that was creating player value models based off these, and other, statistics. GAR translated different statistics into one currency (goals) in order to compare the relative importance of different areas of the game: scoring, shot differentials, shot quality, penalties, etc.
Goals have a lot of natural variance. They are rare enough and have enough confounding variables to look and act random, even in a sample as large as a full season.
Most people who follow hockey analytics understand the consequences of goal scoring’s randomness, at least up to a point. They understand that this is why goal metrics tend to be inferior to shot metrics (like Corsi, Fenwick, and expected goals).
As we noted, goals are the object we generally try to predict in out of sample correlations. But as we also noted, goals have a lot of “randomness” mixed in with the talent portion.
When you build a model that is supposed to measure player talent or performance level, you want it to predict future outscoring well, because good players are more likely to do well in outscoring.
However, when you are trying to predict something that is in part “random” (ie: not controlled by that individual player), you are not wanting to actually predict the success that is not controlled by that individual.
SHOT ATTEMPTS DISTRIBUTION CHANGE
As ideas like shot attempts become more engrained into NHL front offices, and players who tilt the ice in their team’s favour become appropriately valued, that the spread in shot attempt ratio across the league is becoming tighter and tighter. In other words, the gap between the best and worst Corsi teams is getting smaller. Certainly, there is some evidence for this claim.
It’s also important to look a little more deeply and examine why this measure of spread is getting smaller. The way we (or at least, I) tend to think of convergence is that everyone is uniformly moving closer to 50% shot attempt ratio, which represents true parity.
We can do something simple and heuristic to examine where the convergence is coming from. This is not fully robust, for a couple of reasons that I won’t bore you with, but I think it’s sufficient to get the point across. Below is a table that tracks how the top ranked, 5th, 10th, 15th, 20th, 25th, and last ranked team in shot attempt ratio has performed in that metric in each year.
First off, note the scales. We’re talking about a pretty small shift, all things considered.
The most obvious thing here is how much better the ‘worst’ teams are.
The two or three worst teams from 2013 to 2015 were far worse than they are now, at least by shot share. This is one area where the spread has definitely converged dramatically, whether it’s due to the reduced incentive for tanking (which took effect in the 2015/2016 season) or other reasons.
On the other hand, the convergence of the top ranked team is not quite as dramatic.
So, what’s the takeaway on the claim I started this off with. Well, shot attempt ratios have converged over the last five seasons. However, we’ve seen no obvious pattern in the last three seasons. As such, it’s not clear to me that this convergence is continuing.
Teams will be tight in terms of shot share.
More and more, what I think teams will invest in, and where they can find huge returns is in the ability to manufacture higher efficiency shots (i.e. shot quality). A sustainable edge in shooting percentage can go a very long way. While it’s generally accepted that most players do not exert significant control over their own (or linemates’) shooting percentage, there are some that do. I think they’ll become even more valuable in a post-2016 world where the spread of shot share is smaller than it was during the Enlightenment Era of 2007 – 2015. This is where we are thankful for Auston Matthews.
More significantly, I think coaching will turn towards systems that generate more high quality chances, in a sustainable way. We’ve already seen teams like the Leafs completely remove point shots from their diet, and focus on working the puck into the slot more.
Comments