#p!res:analytics

[puckpocketed]

Hello. Clips of the NHL sortable by year, team, player, shot type, and game state (even strength, shorthanded, power play) UPON YE

[puckpocketed]

excerpts from How Jared Bednar’s pregame meetings help fuel Avalanche success: ‘There is no detail missing’ by Corey Masisak for The Denver Post, published 3 April 2026 (archived)

Bednar gave the 300 people in attendance an in-depth look at how the Avs operate and how he coaches. One of the most fascinating topics was how he uses the data Colorado’s analytics department provides him, and what he passes on to his players. “(The analytics team) digs in on our opposition, so like we hadn’t seen Calgary yet this year,” Bednar said. “So I dig in, and I have an idea that they’re a quick breakout team — sixth-fastest in the league, and they’re fourth-fastest through the neutral zone — like everything’s up and out quick. So we have to manage our depth on our forecheck right away. We want to hunt their ‘D.’ They’re prone to some turnovers, but we need to manage our depth because we can’t let guys get in behind us. “That was an important point for us (Monday) night. I felt like if we got on top of them quickly on the forecheck, that we could create some turnovers. We did. Their (defensive) coverage is a similar coverage to what we’d seen before, that our team tends to have significant amount of success against. So you’re just hitting points again from some of the teams that you just played that had the same coverage. We wanted to make sure we’re shooting the puck and challenge them inside. We did that. And then on the defensive side of it, like what do you need to be aware of?”

One of the bedrocks of Bednar’s coaching philosophy is breaking the season into 10-game segments. That helps him and his coaching staff self-scout and identify any short- and long-term issues the team needs to address. [...] “I give them numbers, but they’re not getting decimal points,” Bednar said. “They’re getting ranked in the league. Like (Calgary) is fourth-fastest through the neutral zone. I don’t get too worried about 1-2 games a week or whatever, but when you’ve got through 10 games, we want to try and be top five in everything. “If there’s something like we haven’t scored enough goals in a segment, and I’m noticing that it doesn’t seem like we’re shooting the puck enough, we’re not getting inside enough — generally, when the numbers come back, all of a sudden we’ve slipped from fourth or fifth in the league in low-slot chances to 27th. Guaranteed, I’ve got a lot of video that can match that. … That’s not a recipe for success. So I sell it to our guys and where I’ll give our guys a lot of credit is, over the years, when I address something, it’s generally (fixed) the next night. … They fix things and implement them very quickly, which is why I think we’ve had so much success this season.”

[puckpocketed]

So you can't have conversations with me about this sort of thing because I'm happily in the weeds 4ever but like what does "grit" and "physicality" even mean. Your team GM and coach are always harping on about getting players who have it, every mediocre hockey podcaster and Twitter replyguy wants to blame lack of playoffs success on whether the team has it or not.

Try to define it right now, without using the umbrella words. Is it hitting? What kind? Is it more hitting, timely hitting, harder hitting? Is it about being meaner and more underhanded? But then, not all contact is hitting. So is it being able to take more contact? More efficient contact? Inviting more contact and drawing penalties? Utilising different kinds of contact at the correct moments? Where on the ice are they making the contact, and what state is the game in?

There are so many ways to contribute on the ice. I think we can loosely say that many players who we assign the "physical" attribute are good at a mix of the above, give or take some weaknesses, and in doing so they come through the contact with the puck in a better state than they found it (with possession for their team, with a scoring chance, successfully defending a scoring chance). But loose is not good enough for me, I want to KNOW!!!

Define modes of contact, where it happens and how it happens. Start tracking the data: who is where, and doing what, and who comes away with the puck in a winning state. And then we plot it on a league wide chart to figure out who is doing it best. Maybe then we can start talking about grit and physicality beyond the dawg-in-them -> no-dawg-in-them spectrum.

TO BE CLEAR!!! I think being good with physicality often does pair with passing the dawg-in-you eye-test; meanness, a willingness to make contact, fearlessness in taking contact, a visible ferociousness during contact, all likely contribute to a player's Grit factor. But, given all the ways to be useful while taking/making contact, there are probably players you wouldn't peg as "gritty" or "physical" at first glance for various reasons (they're ultra-skilled and hence they are seen to play a "soft" game, they aren't making crushing hits, they aren't outwardly mean/aggressive, whatever stereotypes players get stuck with) who when you actually measure/track their play, they're doing really well with some aspects of physicality.

I am not an eye-test only guy, nor I am a numbers only guy. I am a keen idiot and I want these things to make sweet sweet love and spawn beautiful analytics and write ups that I can use to MINDBLAST people who are wrong about players.

That said, here is an attempt at it: Defining Physicality and Skaters’ Ability To Play Through It (archived) by Louis Boulet - published 13 August 2025. I'm going to paste the categories of contact this project uses. If this interests you at all I highly encourage you to crack the full article.

In this project, we’ve isolated six different categories of plays that will help us. They fit the following criteria: A) Can be quantified using publicly available datasets B1) Implies direct contact with an opposing player OR B2) Occurs in an often highly contested area of the ice C) Encompasses successful plays that improve the team’s position in the game

1. Proactive Contact Peripherals

The first category falls more into the “when” rather than the “where” bucket. These are commonly-known peripherals (often found in most boxscores for example) where a skater proactively makes contact on-puck with an opposing carrier in some way to impede their progress. This can be directly applied to the carrier through a hit and stick-check, or indirectly by obstructing a shooting lane, thus resulting in a block. Penalties drawn are also included here as they are often the result of a player successfully inviting contact and forcing their opponents’ hand (whether purposefully or not).

2. Transition Contests

Up next are pressured plays in transition. When against the puck, this takes the form of entry targets where the defender rapidly closes space on the carrier to deny possession past the blue line. Meanwhile, examples with the puck are mirrored where the carrier is subject to pressure but successfully maneuvers through it for a successful breakout.

3. Forechecking Engagement

The following couple have been separated into their own sections but have a similar focus: the forecheck. This one encapsulates the initial engagement off of forecheck commitments, including puck recoveries, pressures onto opposing defensemen’s retrieval attempts, and disrupted exits. Prototypical checking forwards will especially thrive in these metrics, as their crash-&-bang style of play is well-represented with the common battles here.

4. Forecheck Offence

Once possession has been gained off those forechecks, we can shift to more offensive-leaning skills. This transition from applying pressure to creating from within the chaos is the main focus of this category, as we isolate for chance creation off forecheck involvement. Whether that’s with shots or assists, this prioritizes the blending of physical leverage, puck control, and processing to reward skaters with the ability to quickly shoulder between skills in a suffocating environment.

5. Down-Low Playmaking

Maneuvering contact is also rather important during offensive zone cycles. On the puck-movement side, the goal-line playmaker has become an important commodity in today’s game. Bumper plays relying on touches between the boards and net headline multiple powerplay strategies, cross-crease threads through the lower-slot are the endgame for a plethora of set plays. Those passes originating from low in the zone improve teams’ scoring states[1], and require comfort in one’s ability to wade through tight spaces consistently.

6. Net-Front Presence

Lastly is the type of offence most synonymous with grit: the net-front game. It’s at the core of Chris Kreider, Zach Hyman, & Brady Tkachuk’s playing styles, and that’s thanks to a powerful mix of body positioning, situational awareness, balance, and puck skills. Their volume-of and finishing-on rebounds, tips, and deflections are what’s being used to quantify this. Defencemen are not forgotten here, as a mirrored version of this score is used for them. Their performance is measured by how efficiently they can suppress chances close in front of their goalie. This can be achieved through establishing strong positioning, boxing out, or tying up sticks.

[puckpocketed]

MIT Sloan Sports Analytics Conference 2025 - Decisions on Ice: The Next Frontier of Hockey Analytics (YouTube video. Please be advised you WILL be hearing it entirely from your left ear.)

Arda Ocal is moderating, hear from: Philippe Desaulniers (Habs, makes the tools GMs use to fetch data, reports, anything), Meghan Chayka (Stathletes), and Jeremy Rogalski (Bruins - he does get asked about trade deadlines briefly due to the banger that they had this season LOL)

I liked the diversity of voices here, felt like I was getting an interesting cross-section of the kinds of people who work in analytics & their perspectives. Lots of fun things but the first section was what I found most interesting. They talked about how analytics has changed coaching rooms, players and their relationship to analytics, and give a bit of insight on how that interface all goes nowadays. Transcript of that part below <3

ROGALSKI: Echoing what Phil and Meghan said, like, just the way to describe the game. Y'know, working with a - I started in the coach's room doing video work for them, and watching how, even a pre-scout for a game evolves from, y'know, coaches previously would watch like, three games of video and kind of look for some common trends in that, maybe do a deep dive on the season. But now you can describe, like, how a team breaks out, or how their power play formations are in so much detail that wasn't available 10, 15 years ago to put that together.

[...]

ROGALSKI: It's actually interesting you use the word 'storytelling', too, because I think Meghan's bang-on, I think, about a lot of inquiries coming our way now, which is great! But I think also in terms of, y'know, building a relationship with a coach, understanding what their schedule looks like, what their needs are, where they have windows of time to work through and educate on the data.

But also, I think packaging it in a digestible way, right? You're still, in some ways, doing the same thing with a coach where if there's a point you're trying to make, there is a time for a fact: "This guy's whatever on faceoff percentage," that's a fact, that needs to happen. But sometimes you gotta package it in a narrative that really helps them understand how they can then turn around and use that fact to improve the team or communicate to a player. So there is a component of that storytelling that's needed as well.

DESAULNIERS: To tack on what Jeremy was saying... So, I really see three things mainly that are important in order to have communications flowing between our group and the coaches.

The first thing is, really, it's really great to have a translator. You usually will have, like, someone on the team, or you want to have someone on the team who's able to speak the language with the coaches and who understands on the analytics side.

And so we'll be creating some reports, some setting up the information there so it can be consumed, and then you'll have, you, in our (the Montreal Canadiens') case it's Chris our Director of Analytics, who will be talking with the coaches and knowing what the coaches are expecting and presenting it in a way that they understand. So that's one thing.

The second thing that we found effective is, we've started creating some concepts that we try to keep as simple and as high-level as possible, but that then become part of the vocabulary that we have with the coaching staff.

So just to give an example, we'll be talking about players that are, that have a certain number of pillars, we call them pillars, which are at the bottom. They're metrics, very specific metrics, that we're tracking - you don't need to understand all the details to know that, okay, so this guy has a four pillar, or five pillar, and this way you know, like, where he stands at a very high level without having to dig into it.

The third thing that we do, that we've done, is get the coaches on board, so work on a way of organising the metrics based on what they're looking for, what their understanding of the game is, and organise the information that way. So then when they start digging into it they actually see what they were looking for.

And you get the buy-in from them, too, when you're doing that, because they've worked on it to build that structure.

ROGALSKI: I would add one thing, too, that I thought of. You raise a good point, too. I think visuals are huge here, too, right? Whether it's a heat map or just picking the right graphic or trend, or something that's -

DESAULNIERS: Colours!

ROGALSKI: I mean, reds and greens, like, it helps. It truly does help.

CHAYKA: Not to play your role Arda but I have a question for Jeremy, because you've been through some different coaches. Do you find, like, coaching staffs, as they change... From my work I find some are far more, just, ease of use and understanding and their ability to implement it in their game flow. Works a lot better for some people's style than others, have you seen that change with, like, different coaches?

ROGALSKI: Yes. It's definitely seen a change. Also, I think now you're starting to see coaches have worked with different teams, so. Before it's, you're educating on, "This is what we can offer you," now it's, you're coming in with a set of preferences, or, "This is what I worked with previously, " so.

Y'know, more recently it's been coming in with a set of expectations for, "I know this works for me, is there anything you want to add or change or do whatever?" So, it's definitely, there has been a change.

OCAL: That's a great question, I want to generalise that, like, y'know whether it's a first-time coach like Marty St Louis, whether it's a Jim Montgomery, or a long tenured coach... Let me start with you, Meghan, like, from different experience levels of coaches, in the NHL or otherwise, how have you found their acceptance and how they use that in their daily jobs?

CHAYKA: And I think it can be background, too. You know, you, you're a math grad, you like numbers. Some people come from engineering. People come to coaching in so many different ways, so they have such a different way of approach and understanding and ease with it.

And I think that they can just implement it and know what's assumptions and what's noise and what's actually important to them and their job that they need to do. So I think it's just being realistic about your personnel.

But I will say that I've noticed, doing the Draft a lot, and at ESPN and working with Draft people in the CHL and doing a lot of Junior work, that young people now expect data. They want it, they want to take control of their own, whether it's a coach or a player coming up. So I'm seeing, like, the 25-and-under crowd.

It's not, "Do I, should I use analytics?" it's like, "Why aren't we using analytics? Why isn't my agent doing that for me? Why isn't my team doing that for me?"

And I think that, too, both empowers people working in tech but also empowers, like, the person who that data is being used to dictate their career, right? Giving them some control and oversight and insight into what they actually have to improve. So it's kind of, like, flipping it on the head a little bit.

OCAL: The paradigm is changing!

CHAYKA: Correct.

[puckpocketed]

Excerpts from Better Late Than Never: Playoffs Post-mortems - What style of hockey won in the playoffs? (archived) by Corey Sznajder for All Three Zones, published 30th July 2025 via Substack (emphasis mine, muddling through some thoughts below the cut)

[…] The main point that always gets beaten to death is how tight the playoffs are and how the games are won in the trenches, basically that you can’t win as a rush team. I’m kind of annoyed that is still a talking point because fans & analysts get too hung up on that you need to play one way or another to win. The split in goals scored on the forecheck vs. the rush has always been in the 55-45 range in both the regular season in the playoffs, with the scales tilting more towards the 60/40 range if we include goals scored off faceoffs. That was the case this post-season with 137 of the 329 5v5 goals scored coming off the rush compared to 146 being scored off the forecheck or the cycle, so a 54.8%/45.2% split.

What constitutes as forecheck offense is also kind of a Rorschach test, because it’s usually associated with playing physical in the trenches, while breaking the defense down when you have possession is the bigger part of the puzzle & one that makes or breaks you from becoming a playoff team. There’s a handful of teams that can play off the rush with some of the best in the league (Utah immediately comes to mind), but haven’t quite gotten the balance down of how to score when it becomes a half-rink game and the space is limited. On the flipside, you have teams like Anaheim and Vancouver who wanted to play the heavy territorial game with no rush component and had neutered their own offense. The good teams find the balance & it’s something the two-time champion Florida Panthers have mastered.

The rush might not be how the Panthers play most of their games, but it’s a major part in how they win and it’s connected to how they play in the other two zones. An interesting sidenote of how it relates to their roster construction is that 6 of their 22 goals off the rush were scored by Sam Bennett, who also ccounted for 12% of their shots on goal that were off the rush. Another three were scored by Brad Marchand and he assisted on another one while contributing to 24 shots overall. His center, Anton Lundell, didn’t score a goal off the rush but assisted on four and contributed to 26 shots. Right behind them was Eetu Luostarinen with 20 shot contributions off the rush, two of them being goals and four of them being assists. Absent from this list is Sam Reinhart and Sasha Barkov, who scored all 11 of their five-on-five goals off the cycle or the forecheck.

This is the balance you’re looking for when building a top team. Florida’s top guys can contribute off the rush, but they didn’t have to in this run & they build their other two lines around guys who can play with more speed & strike the other way when they get a turnover. […]

It’s been fascinating and somewhat frustrating to follow two NHL playoffs cycles and see how much hockey commentary/punditry/analysis on broadcasts pushes vibes-based (eye test) narratives about “winners” and what “winning hockey” looks like. What I want to read is analysis of tactics and how it evolves through a given series, what I often get is Big Player and Violence and Strong Forecheck Good.

If you put a gun to my head I’d say I think we (the broader/maintream culture, maybe not me and my good friends on tumblr dot com) overvalue grit/toughness and post-Cup discourse tends to support this narrative. I have no survey numbers, just something I’ve noted having listened to/read many disgruntled fans diagnosing their teams’ woes and what the Panthers did right in order to make it 2 years in a row. Feels like a lot of people were invoking nostalgia for the 2012 LA Kings committing real, actual murder on the ice and how the Panthers reminded them of that in their SCF breakdowns. If not the comparisons then heavy emphasis on their deadly forecheck and how they (surprise) did a lot of murder, and how their own losing team would be better served following the example.

I’d say it bleeds into Draft discourses, too. In following the Draft and all the media that surrounds it I’ve heard so much speculation about GMs seeing Stanley Cup winners and allowing their playing styles and team makeup to dictate the types of players they pick. And while I feel like this discounts the years-long process that amateur scouts (“amateur” as in scouts employed by teams who follow prospects as opposed to scouting other teams for strategic reasons) go through writing reports, watching tape, and analysing and categorising prospects to build draft boards, it’s not like I’m in the room with Mike Grier or Pat Verbeek! what do I know!!

The gun-to-my-head part of this is vital. I say all this because I know I’m throwing stones in glass houses when I get frustrated with narratives that go off eye-test; I understand the vibes impulse, much of what I’m saying is also vibes and I wouldn’t put it forward as a serious take until I had data to back it up. If someone with more resources and expertise wants to write an essay about this or just explain to me how I’m wrong, please CC me into the email chain !!

All that said, I enjoyed this piece for the narrative balance it provided. Reading Panthers Forecheck Good, while valuable on its own as an example of how to break down games, was getting kinda stale <3 So please have a read!

[puckpocketed]

Excerpted from Unmasked: Examining drop in NHL save percentage

Asked this week about League wide save percentage being at .901, which would match the lowest mark in 19 seasons if it stays there, Vasilevskiy’s first response was not about how much younger and better the shooters are, or how much more dynamic the offensive chances they create have become.

Vasilevskiy did reference each factor, but the first thing the 2019 Vezina Trophy winner pointed to was how much fewer easy shots goalies face now.

“I feel like guys don't waste shots anymore,” Vasilevskiy told NHL.com. “You know, back when I came into the League (in 2014-15), it used to be 30 or 35-plus shots each game. It was almost every night, you felt good about your game, you were into it all the time. Nowadays, guys don't waste shots. It's all about the quality. They're all looking for that perfect play, perfect pass, perfect shot.”

“In my opinion it's all those young players, they're super-skilled and they know if you shoot at net with no traffic the goalie will catch it and then what?” Vasilevskiy said. “Whistle, the play is dead, and nobody will pat you in the back for that.”

Entering Thursday, shots per game were down in the NHL this season to 57.4 from 60.6 last season and a five-season high of 63.3 in 2021-22. As for how much of that overall drop is a decline in those easier shots Vasilevskiy talked about, in the four seasons that NHL EDGE data has been tracking long-range shots, which has a .970 League average save percentage, the per-game average number of long-range shots has dropped from 14.9 in 2021-22 to 12.81 this season (entering play Thursday).

Former NHL goalie Steve Valiquette, who founded Clear Sight Analytics and works as a television analyst for New York Rangers broadcasts on MSG, said the number of unscreened shots -- he calls them clear-sighted shots because the goalie can see the release -- from long range is down 27.9 percent the past six seasons, and the amount of those shots from outside the slot has dropped 20.8 percent.

Some teams view those types of shots as a turnover, trading clear possession of the puck for the chance to win it back on the ensuing face-off.

“Vasilevskiy is spot on,” said John Healy, who is the chief operating officer at Clear Sight Analytics and also works one-on-one with NHL forwards and defensemen on how best to create offense. “We’ve stopped taking these clear-sighted shots.”

[puckpocketed]

Excerpts from Hockey Analytics - Building xG models in Python by Lars Skytte, 18 May 2025 (archived)

Much of the maths escapes me, but I liked picking through this person's rationale - defining xG, their considerations as they built their model... whole thing is a fun read but I pulled what I liked <3

What are Expected Goals

Let’s start with the “What” – What is xG really? Here’s my broad definition:

“Expected Goals is the estimated number of goals based on a set of variables.”

All xG models I’ve seen in hockey haven been shot based. Meaning it’s an estimated value of each shot or a goal probability for each shot. However, I don’t necessarily think xG models should be limited to only shots. That’s why I’m using the broad definition of expected goals.

With the public data available we’re forced to base our xG models on shots. Though, with proprietary data you could possibly take a different route. I know that even the private xG models are shot-based, but I believe you could be more innovative with your approach.

The most obvious alternative would be to build a possession-based xG model – to create an estimate of what each possession is worth. So, instead of looking at shots created you would look at possessions created in dangerous areas. This would allow you to estimate the value when a possession starts and compare it to the value when the possession ends. Did the player add or lose xG through his possession?

You could also better describe defensive value. A possession breaking defensive event would be the difference between the opponent having possession of the puck and your team having possession of the puck.

[...]

Why do you need an xG model?

This leads us to the third and final question. Why do we need xG models? What are the purposes of the model(s)?

Let me start by saying I think the name expected goals is somewhat misleading. It leads us to believe that the only purpose of xG is to predict future goals. I think estimated goals would have been a better name.

It’s very important to understand the “Why”. You need to build/choose the xG model based on the questions you want answers to.

One of the ongoing discussions is whether or not you should include shooter talent in the xG model. It makes sense that a shot by Alex Ovechkin has a higher goal probability than a shot by Adam Pelech… But if the purpose is to describe shooter performance, then we can’t include shooter talent in the model, because that’s the very thing we’re trying to determine.

Generally, it’s important to understand the difference between descriptive modelling and predictive modelling. Do we want to describe past events, or do we want to predict future events?

So, probably one xG model isn’t enough. However, expected goals is often considered a single metric like fenwick or corsi, when in fact it can be many different things based on the input variables.

Here are some of the things xG can be used for and some modelling considerations:

Describe shooter performance (goals scored above expected) – The model should be fenwick- or corsi-based because hitting the net and not being blocked is a shooter skill.

Describe goaltender performance (GSAx) – Should only include shots on net, as I don’t think we should credit goaltenders for shot misses.

Describe shot creation/prevention (Think “Deserve To Win o’Meter” or xGF%) – Should be a fenwick- or corsi-based model and you should exclude empty net shots, as they will skew the data (or look at 5v5 data only).

Predict future results/goals – Should be a fenwick- or corsi-based model, and some the variables might need to be weighed slightly differently. I once saw some research that removing rebound shots increases the predictive power of an xG model. Probably because rebound shots typically have very high xG values but are still somewhat random.

Possession value added – If we were able to build a possession-based model, then we could calculate possession value added. How much value a player provides through his puck possessions.

[puckpocketed]

As always I am trying to absorb as much as possible about the Beautiful Game + pushing my agenda against low % point shots. Please see full article if you have time, it's long but very interesting... also feels pertinent to some teams' seasons right now.

Excerpts from Passing Project Revisited - What does four years of passing data say about how NHL teams create offense? (archived) by Corey Sznajder for All Three Zones - published 13 January, 2021 via Substack

The NHL season starts in a few days and this is the time of the year where my brain goes into overdrive with tracking games & getting the data out in the open. You’ll get hit with a barrage of stats, charts and tidbits from me and other analysts. You’ll hear lots of terms like “shot contributions,” “low-to-high passes,” “transition plays” and “high danger passing plays.” What do they all mean, though? Are they important or are they just fun stats to look at during games?

[...] We’ve gotten better at estimating the individual impact of players using play-by-play data, but I feel like manual tracked stats will always have their place in help explaining the “why” behind macro-level stats. Unless a play led to a goal, you’re generally in the dark about most events, except for whoever shot the puck. Was it off a passing play or did he carry the puck into the zone & beat a defender to generate the chance? Which one is better?

Most of the legwork was done back in 2015 by Ryan Stimson with his passing project and his initial observations revealed some cool things about offense in hockey. Some of it is very intuitive, like how a shot preceded by one pass alone has a small, but positive impact on a team’s shooting percentage (7.2% on unassisted shots to 8% with one assist). That number jumps to 9.6% when another pass is added to the sequence, which bridges the gap a little between expected goals & actual goals. Through his work, we also know that passes from behind the net lead to goals at a higher rate than other passing plays, most notably shots from the point (which are the lowest of low percentage shots). Things to help bridge the gap between transition stats like entries & exits have also been looked at by him, so there has been a lot of neat discoveries made through manually tracked data. [...]

The Basics of Passing Data

Even with over 180,000 shots added to the database over four seasons, most of Ryan’s initial findings on the impact of a pass hold true. When factoring in rebound chances it’s especially true. Over the four seasons tracked, teams scored on 8% of their 5v5 shots on goal without a passing play. This is including rebound chances. If we exclude rebounds from this sample, that shooting percentage drops to only 4.1% out of all shots on goal. A rebound is recorded as an unassisted shot, but they are usually reactionary, second chance opportunities that are created off a shot from another teammate. They also come closer to the net & are a little different than your standard unassisted shots. Rebounds are also closer to the net and are somewhat difficult to track with 100% accuracy if you’re dealing with a situation where two or three players are going for the puck at the same time & are just jamming it into the goalie’s pads. All other unassisted shots had a shooting percentage of 5.6% or less, which goes to show the impact of just one pass. Yes, it’s only 3% but think of how many shots teams take in a year. Incremental improvements over time add up & that’s especially true when you consider that teams shot at 12.6% when they completed three passes before a shot on goal. [...]

Royal Road & Behind Net Passes Still King

Going strictly from a percentage standpoint, the easiest way to boost your team’s goal total is to generate more passes from behind the net or create a cross-seam play to get the goaltender moving. The same rules in 2016 hold true today. The interesting part is how rare those plays are, as they account for only 13.2% of all shots (excluding unassisted ones). Let’s say a team averages 44-45 shots a game during 5v5 play, this means that only five of those will come of what are categorized as “high danger passing plays.” Ideally, you want to find players who generate these types of passes, as the payoff is huge. The problem is most teams can only complete 2-3 of these plays per game unless they’re playing Doug Weight’s Islanders or last year’s New York Rangers. [...]

Do teams get burned more frequently if they miss on cross-seam plays or behind the net passes? Analysts will often gripe about players trying to be “too fancy” by looking for the cross-ice pass instead of taking the simple play. Sometimes forwards can get caught deep if two guys are behind the net, leading to a rush the other way, but is it any riskier than a shot from the point getting blocked? Ever since Ryan’s study, it felt like more teams were attacking from behind the net, but the overall frequency of these plays has stayed roughly the same throughout all four years. Maybe this suggests that more innovation can be done here.  It would be interesting to see if there is a drawback to chasing these more dangerous plays while sacrificing shot volume. [...]

Cycle vs. Transition

If there’s one area that I’m still working to iron out with this type of tracking, it’s bridging the gap between zone entries/exits and the passing data. Looking at the table above, it looks like plays that happen in transition (stretch passes & passes from defensive or neutral zones) are low-percentage plays. A shot off a stretch pass is about as effective as a normal give-and-go pass and plays from the neutral zone don’t result in many goals either. This isn’t totally cut-and-dry, though. [...]

More passing plays in a sequence yields more goals but the reward from deeper passing plays isn’t as great as you’d think. At least compared to other sequences where 2-3 guys touch the puck before getting a shot on goal. This changes a little bit if it’s a stretch pass,* but the shooting percentage boost you get from these plays is in line with the other passing sequences listed. The one caveat here is that plays that originate from outside of the offensive zone produced a higher percentage of scoring chances, which could come down to an Expected Goals vs. Actual Goals debate & which you prefer as a coaching staff. [...]

Risks & Rewards of “Safe” Offense                                                           

If there is anything that stood out to me, it’s how often teams still revert to point shots for offense. Over 22% of all primary shot assists were low-to-high plays, which are essentially point shots that have roughly a 2% chance of resulting in a goal. These shots even getting on goal is a crapshoot because unless you’re shooting from 40 feet away on an unscreened goaltender, the puck has to travel through layers of bodies, traffic and other layers of defense. It’s the definition of a low-percentage play & an uncreative way to run your offense.

So why do teams revert to this? It goes back to a few things we discussed earlier. Hockey is a read/react type of game and most of the open space in the offensive zone is going to be up high where the defensemen are. You have more time & space to operate from up high and sometimes the only play available is to take the shot. It doesn’t mean you should do it, but in the heat of a moment during a game, reverting to the safe option is pretty normal. Scoring directly off the point shot also isn’t the primary objective, as you are usually hoping to score off a deflection or a rebound. How does that compare to, say, trying to setup from behind the net? From a shooting percentage standpoint it’s an interesting tradeoff.

Deflections & rebounds have the highest conversion rate out of all 5v5 shots tracked, shots that are deflected having a 29.9 shooting percentage and rebounds having a 23.4 shooting percentage respectively. [...] You get a decent scoring chance that about a 50-50 chance of getting on goal, which doesn’t sound that bad compared to some of the other options listed.

The issue with this is that we’re only looking at successful tip plays. If a team is attempting this setup & misses, what are you left with? A simple point shot that has less than 2% chance of finding the back of the net. Out of all 5v5 shots tracked that involved a low-to-high sequence, only 8% of them resulted in a deflection. So, while a deflection is a high-danger chance, generating them is tough and most of the time you’re just settling for a low-percentage shot. Combine that with shots off low-to-high passes only reaching the net 35% of the time, and it’s a tough way to win if that’s the only way you can create offense. The scarcity of goal-scoring in the NHL might encourage this type of hockey, but I still think more can be done with how skilled players are now.